0:11
Hey, hey, hey. >> [music] >> Heat. Hey. Hey. Hey. Hey power.
Search this video
1:59
>> [music] >> Hey, hey, hey.
3:13
We came to this world to reach the stars.
3:23
We came to this world to shape our future.
3:34
You came to this world [music] to shape a future.
4:03
If you feel [music] UAV online.
4:39
[music] >> [music] >> You're watching TVPN.
5:00
>> Today is Thursday, November 20th, 2025.
5:02
We are live from the TVPN Ultradome, the temple of technology, the fortress of finance, the capital of capital. Ramp. com. Time is money. Save both.
5:07
Easy to use corporate cards, bill pay, accounting, and a whole lot more all in one place.
5:12
Um, Nvidia beat earnings and job uh the job numbers came back very positive.
5:19
Um, 119,000 new jobs and Nvidia beat earnings.
5:23
The revenue came in at uh 57 billion for the quarter, up 62% from this quarter last year.
5:34
Uh, fantastic result for Nvidia.
5:38
Of course, that's why the stock's selling off and the market's melting down and Bitcoin's down 10%.
5:44
>> That was >> Bitcoin's down.
5:45
>> That was my prediction.
5:45
That was my prediction after yesterday.
5:48
>> One of your predictions, >> but I was but I was uh I was wrong on the timeline.
5:52
It was an interesting >> Yeah, it pumped shortly and now now everything's selling off.
5:57
Uh very unclear where we go.
5:59
unclear where we go. Um, but I did think it was just interesting that the uh [snorts] I didn't really piece this together into like some grand thesis or actually write a piece about it, but uh I did think it was funny that we are in
6:10
a world where uh demand for robots is surging and also demand for human labor appears to be surging like uh Nvidia you know the chips that they make sell artificial intelligence that should be uh replacing human labor and yet uh the job demand is is surging as well. Um
6:26
Um >> uh and it's notable they they said they have visibility for a half a trillion dollars >> in revenue. Yeah.
6:34
Through 2026, which >> I mean it seems crazy, but >> it's not enough anymore.
6:40
>> They're but I mean they're making 57 billion a quarter just for the next quarter.
6:44
Guidance is at 65 billion.
6:46
Analysts had predicted that uh revenue uh guidance would be 62 billion.
6:49
So uh everything is trending up.
6:52
Uh Jensen said, "We've entered the virtuous cycle of artificial intelligence.
6:57
AI is going everywhere doing everything all at once." What a great quote.
7:02
Tyler's very happy about Jensen.
7:05
Uh and I'm also happy about Reream one live stream 30 plus destinations.
7:11
If you want to multiream, go to reream. com.
7:12
Um so we talked about it a little bit yesterday on the show.
7:18
There's a new product from Travis Kalanick, the founder of Uber, of course. Uh it's called Picnic.
7:22
We discussed it on the show yesterday and we got a reply from none other than Travis Kalanick himself.
7:30
>> Uh why don't I read his uh his reply and then you could kind of take us through uh the uh what you wrote in the newsletter kind of sort of >> why don't I start why don't I start with a little bit of context and he can add to it.
7:43
So, >> uh, I wrote in the newsletter this morning, uh, of course, uh, the subject of the newsletter was Daddy's Home.
7:50
[laughter] Uh, and that is, of course, Travis Kalanick.
7:52
Uh, Travis is back on the timeline.
7:55
>> It's so good to see him back on the timeline.
7:56
Like, he's he's never he's never been not been doing business, but he's been so quiet.
8:00
he's been so quiet. dropping like deep alpha on the market that he's operating in that is like non like in my view like I never it makes sense what he what he shared which you can get into but I never looked at it exactly like that >> uh but he's obviously back on the
8:16
timeline uh with picnic picnic is a new business under city storage systems okay >> so you don't know the name city storage systems but cloud kitchens is actually a subsidiary of city storage >> I thought cloud kitchens was the top >> that's what I thought too but it's actually the opposite city storage systems. Great, great name if you want
8:31
Great, great name if you want kind of an under under the radar holding company to uh you know verticalize food delivery.
8:39
Uh so Picnic is kind of a front-facing platform focused on meal delivery.
8:44
The offer sounds too good to be true.
8:47
Uh meals delivered from 50 plus restaurants with no tipping and no fees.
8:52
>> Uh they also bundle orders so a company can order from >> 10 or so different restaurants, get it all ordered at the same time. Mhm.
9:00
>> Uh he's got a bunch of customers already.
9:01
Wells Fargo, Live Nation, EY, KPMG, uh PWC, and a bunch more.
9:07
>> And so we were talking yesterday about how like broken the tipping experience is when you're tipping directly, it's a way to encourage great service by like tipping.
9:17
If you're checking into a hotel >> and you're tipping uh you know, somebody on the way in, >> they're incentivized to make your stay great.
9:25
Uh same thing, you know, valet tipping on the way in.
9:27
They're going to park your car right up at the front.
9:30
>> I saw a viral uh uh maybe Instagram reel or something about a guy who says that whenever he checks into a hotel, he says, "You know, uh we always tip the valet when we're here.
9:40
We tip the bellman and the person that cleans, but you know, you folks at the front desk just don't get enough love."
9:46
And so here's a nice crisp $100 bill.
9:48
And he says, "Right desk folks really get >> they never get tipped."
9:52
And he said every time he does it, he gets upgraded to an insane suite.
9:55
And so he was just like sharing this alpha.
9:58
I think it's >> at the hotel I worked at Michael Jordan would stay and he just actually would carry around like 10 grand and just any he was just handing handed out like candy on the property >> and he would have a very nice stay >> as you might imagine.
10:13
Um so anyways we were talking about that Travis responded and you can get into it and kind of give your reaction. >> Yeah.
10:20
So Travis said uh delivery app tipping isn't about feedback feedback mechanisms.
10:25
It's a tool for maximizing the price paid by consumers.
10:28
Eaters are economically irrational with tip.
10:31
For every $1 in tip, they economically behave as if it were 80 cents.
10:35
Uh this is just a hypothetical figure, but it's directionally true because you feel emotionally good about tipping.
10:46
Mentally, it you you give it less it feels less painful to part with those >> purchase. Yeah. Gas buying gasoline. Exactly.
10:55
So if you the way you look at if you if there's $10 in taxes and $10 in tip, you'll be like, "Oh, I feel good about the $10 in in tip.
11:01
That feels like $8 and the taxes that feels bad, right?"
11:07
Um and and and and it happens on the other side.
11:09
This means that uh less price elasticity for the same price.
11:12
So couriers are also economically irrational with tip.
11:17
For every $1 in tip, they economically be they economically behave as if it were a$120. Again, directional.
11:23
And so you feel good when you're tipped and so you treat those dollars more as more valuable.
11:27
And so this is a hack on the human psyche which apps must implement and maximize or miss out on economic surplus that their competitor will use to defeat them.
11:40
And so even if you have, you know, your whole brand is built around our app doesn't tip.
11:45
Remember this happened with Uber.
11:46
Um, if your competitor is is using tips, if they implement tips, they will just be making more money than you because of this economic inefficiency that arises from the nature of the human psyche.
11:59
I thought it was very, very interesting.
12:00
The app that decides to pay the same net amount to the courier, but as a square deal via a drop fee plus tip, will lose market share every day to an equal marketplace player that implements and maximizes tip.
12:13
implements and maximizes tip. now equal marketplace payer that's doing a lot of lifting because it's hard to just spin up like you know I can't just start an Uber right now it's it's hard um but uh he makes a very good point here and so
12:28
what's interesting is that >> I read this as >> adding tipping is inevitable adding tipping is inevitable we're not doing it right now but eventually someone will come to the market do it we will have to in order to compete is that not the read here. >> So the the diff the difference here is
12:46
>> So the the diff the difference here is that I I think that one picnic is like is already counterpositioned, right? So it's pricing thing.
12:56
It's a flexibility standpoint.
12:57
It's also counterpositioning on like focusing on one key >> buyer.
13:01
Obviously, you know, the Door Dashes, the Uber Eats have their kind of like corporate offerings.
13:05
Um, but I think like just creating a a creating a different and more transparent model makes a lot of sense.
13:14
He's also like I think you have to factor in there's a lot like TK's been kind of like secretive about cloud kitchens secretive about Otter which is like the the Toast or Square competitor that he has.
13:26
And so, uh, un, you know, when I when I hear like no fees, uh, no tips, like it just screams like there's been so many attempts at food food delivery, >> uh, and just like new restaurant concepts that have been venture-backed and a lot of them haven't worked out, right?
13:44
Because it's just like becomes unsustainable.
13:46
Uh, and so I think that I think Travis is basically by focusing on a key customer type, trying to make it up with uh, volume uh, and then having this like vertical approach. >> Mhm.
14:01
>> I I I just like I want to I want to believe that uh I believe that he's somewhat of a massochist and that like going and trying to win in food delivery is just like >> uh the hardest arena just like food in general.
14:12
We have David Chang coming on >> uh at at uh noon which I'm excited to talk with him about.
14:17
But uh it's just like the most competitive space.
14:20
It's low margin all the way down.
14:23
Uh but I think that he I believe just given the domain expertise, I believe that he's uh he has a real play here and a real strategy.
14:32
And I think that I think that already uh we were talking with um uh uh the person on our team that handles like food ordering.
14:41
He got on the phone with Picnic yesterday and he was like, "This offering is way better than what we're seeing uh with the with the other delivery apps and wants to switch to it immediately."
14:50
So again, if if if TK can make the model sustainable, I think it'll be quite competitive.
14:57
>> Yeah, I mean, you would you would imagine that vertical integration should allow low true lower prices, like true cost competitiveness.
15:04
That's like, you know, an age-old business adage.
15:07
If you vertically integrate, you can undercut your competitors and just offer lower prices.
15:12
Uh almost like, you know, buying Kirkland brand at Costco is typically like sort of like the canonical example of like heavy verticalization.
15:19
of like heavy verticalization. Um, and there's there's a ton of other examples, but uh I wonder I still wonder is is this like the I remember in the early days of Uber like it was amazing because you didn't need to think about the tip and so there that mental load wasn't
15:37
there and there was the star rating system and it felt like there actually like the VCs might have been subsidizing it a little bit but it felt felt affordable on the on the on the rider side and on the driver side it felt like people were getting paid pretty well and everyone was sort happy, but maybe the VCs weren't. But they wound up getting,
15:52
But they wound up getting, you know, a stake in a $200 billion company.
15:56
So, uh, you know, I think it all worked out for everyone involved.
15:59
But, um, it seems like Travis is reflecting on this idea that, uh, it was tipping was inevitable to come to the Uber ecosystem.
16:09
Is tipping going to come to the Whimo ecosystem?
16:12
Is tipping going to come to this uh, picnic ecosystem, the picnic product eventually?
16:18
I just >> Do you think Picnic will have tipping in 10 years?
16:23
>> I just view this more as like a as a as a corporate service >> in its current positioning. >> True.
16:29
>> Than a consumer service.
16:29
And when a consumer is buying food, yes, >> it is if you're ordering food delivery, it is it is not a uh uh it is like it is a luxury, right?
16:39
Like food delivery has been extremely normalized.
16:42
But if you, you know, re, you know, rewind to 40 years ago and ask like, oh, how often do you get food delivery?
16:50
Most people will be like, I never get food delivery.
16:52
I just go pick it up myself, right?
16:54
So, it is a luxury, but this is being positioned like as a corporate offering.
16:59
And I think that if Picnic can get just like deep relationships with a bunch of these different companies that have, you know, I listed off some of the logos before.
17:08
if they can um if they can just become embedded in these companies and part of their workflows, I think they'll they'll uh it's possible to like make it up in volume.
17:17
>> Yeah, this is so funny the way Shield puts it. Uh truth bomb from TK.
17:19
Tipping is a hack to maximize price. It's psychology.
17:24
Consumers are willing to pay more in tips than they are willing to pay in fees for or menu price.
17:28
So, a $16 burrito plus a $4 tip feels far cheaper to people than a $20 burrito that has no a no tip option.
17:36
Uh, >> but but again, from a from a from a from a business standpoint, I don't I don't know.
17:43
I don't know if it's exactly the same thing.
17:45
I feel like businesses like want to have like more predictable more predictable >> costs, not have that like variability and like, okay, sometimes the fees are like this, sometimes the fees are like that.
17:56
I think this will be a better consumer experience.
17:58
A lot of companies you know, will give like credits to their employees, which is like you get $20 of credits. >> Yeah. >> Every day.
18:06
And then if whatever you're kind of like spending on top of that, you have to eat. >> Yeah.
18:10
>> And so I think consumers will could very likely like picnic more. So >> we'll see. >> Yeah.
18:16
The uh uh this was one of the original like DTOC uh evolutions that happened um with a lot of like Shopify merchants.
18:24
I remember looking at I think it was like Kylie Cosmetics and uh there was a trend for a while that was like uh consumers want transparent pricing.
18:31
Uh don't do all the crazy psychological hacks.
18:36
So you'd be like yeah I'm just going to put it's 30 bucks and that's what it is and has free shipping and that tax and shipping is included and it's just like what we say up front is feels really good.
18:46
Feels really good to say that.
18:48
And then you'd go to like the high performing stores and all across the board it would be like $9.
18:51
99 and then you go in and there's like $642 added in taxes and then you add shipping and it >> like a popup that's lading you up and just like keeping you on the reel reeling you in like a fish adding adding fees adding fees until you're like okay well now I'm like entered all my information and I'm ready to click the button and so okay you added two more bucks whatever I'll just deal with it.
19:17
Uh so these psychological hacks are just like somewhat inevitable.
19:19
Um but uh avoiding them I think in the short term is a great go to market.
19:24
I just wonder if there's something um if there's something like truly uh like counterposition that that will be durable like Costco Kirkland Costco like has not has been like the lowcost affordable option and that model has held for like >> decades right >> the one thing the one thing that we learned from uh having somebody on the team call picnic is that they are focused on higher volume orders so like teams of like 25 and up. Yeah.
19:49
And so I think that I think that they're just betting like, hey, there's we can we can get a lot of volume.
19:55
We we will be able to like handle having it's one person that goes around and picks up every order from all the different restaurants, right?
20:04
>> And so uh it's quite a bit less of one individual's time.
20:08
>> Much higher order volume than when a company is like, "Hey, we're giving credits to people."
20:11
And then each employee is making uh individual orders and then there's like ends up being like 20 drivers on the road to deliver one lunch which like makes no sense.
20:21
>> Uh I have one more take on this delivery question but first I need to tell you about cognition uh the makers of Devon.
20:28
Devon is the AI software engineer crush your backlog with your personal AI engineering team.
20:32
My question is uh what is TK's uh drone strategy?
20:38
What's his autonomous delivery strategy?
20:40
because he's vertically integrated at the kitchen level.
20:44
He has the point of sales system.
20:47
He has the the sort of ordering front end.
20:49
You can interact directly with him.
20:51
He's going he's cutting out several of the middlemen.
20:53
Um but is he going to be a logical partner for Zipline?
20:56
Is he going to be a logical partner for Coco and Starship and these robotics companies that are delivering uh food?
21:04
Uh Ryan Oscenhorn here says, "People aren't ready for how much better food tastes when it arrives 5x faster."
21:12
That's a hilarious take because like [laughter] in fact I have tasted food right when it's made.
21:19
Like it's not it's not like an entirely novel thing.
21:21
But what he's pointing at here is that Zipline is is getting food delivered in four minutes as opposed to cars that take 20 minutes.
21:32
And so hot food arrives hot.
21:32
Um, which is certainly a benefit, but uh it just does create more of like a, you know, benchmark to the uh to the actual restaurant restaurants.
21:43
>> I tried I'm trying to find if Travis is an investor in Zipline.
21:45
The Google AI overview says >> yes, Travis Kalanick is an investor in Uber.
21:51
Gemini says Uber >> I could not find definitive evidence that >> Okay.
21:57
>> So, so the the AI overview says yes.
22:00
Gemini says there's no evidence. >> Maybe we can ask him.
22:02
I I wonder I wonder if that's like a logical partner.
22:06
Um I mean on the self-driving side, his original vision at Uber, it felt very much like he needed to own that technology.
22:14
He wanted to be uh not just a like a buyer of it from a different company.
22:18
It seemed like while he was at Uber, he considered self-driving technology as critical path as something that should be owned by Uber.
22:24
And then once he was out, the company spun down ATG, their advanced autonomy group.
22:29
Um, >> I mean, but think think about it.
22:32
So, I mean, right now, if you look at city storage systems, you have cloud kitchens, which is making the food.
22:38
You have Otter, which is like the payments and ordering infrastructure.
22:44
>> And then now you have Picnic, which is like the front end.
22:47
>> And any type of delivery method actually like fits into that system, right?
22:50
So, I think he's being I I I would imagine he'll either add a strategy, but potentially more likely he'll just integrate with a variety of drone delivery uh and then autonomous vehicle delivery and and and continue to use uh traditional labor. So, we'll see.
23:10
>> Uh well, let me tell you about Gemini 3 Pro.
23:11
You've probably heard about it, but we're telling you about it anyway.
23:15
Google's most intelligent model yet with state-of-the-art reasoning, next level vibe coding, and deep multimodal understanding.
23:21
I took it for a spin in AI Studio this morning.
23:23
Had it build a scrollable like, you know, as you scroll, the the the bubbles move around and tries to visualize how deep mind and Google brain merged.
23:34
And these these sort of like generative UI around uh deep research reports, I think, are going to be really really fun.
23:41
Uh, I need to continue iterating on this particular one.
23:44
But >> Gabe says, "How could it be a logical partner if this product is for larger teams?
23:48
Like Jord just said, teams of 25 plus.
23:50
Don't think a zipline can fit 25 different orders in the Great. >> It's a good point.
23:54
Uh, Keller said they can fit two full grocery bags worth of food in their drones. >> 25. >> Yeah.
24:01
But but I like what what it'll come down to is the actual cost.
24:06
>> Your team does a group order and instantly gets swarmed by drones. Yeah, that would happen. >> I mean Yeah.
24:09
I mean, right now you order on picnic, one delivery driver is like driving around to a bunch of different restaurants and getting all that food and bringing it to the off office office.
24:17
You can imagine like six different drones end up carrying out >> No, I I mean I think I think for for this particular uh >> for for this particular uh like use it just feels like uh they will be much much more a buyer of like a Whimo type uh autonomy solution as opposed to a zipline autonomy solution. I I would assume. >> Yeah.
24:39
I just think it I I think the most notable thing about Picnic is >> Travis doesn't want to just sit at the infrastructure layer of food delivery, right?
24:47
He wants to own the end customer experience in the brand. So, >> yep.
24:50
Uh well, uh we have a beautiful picture of Alex Karps watch the PC Philippe Aquinaut with the orange band. We love to see it.
25:01
>> We clocked this >> TJ the wheel Jensen, one of Jensen's leather jackets.
25:06
Jensen, of course, has many leather jackets.
25:08
Uh, I've only seen Karp in one single Aquinaut.
25:11
There is a fascinating story of how Karp wound up with this particular watch.
25:17
We'll have to get him to tell it on the show though, uh, at some point.
25:20
Um, we'll also have to tell you about Cognition, the makers of Devon, which I already did. I already did Cognition. I'm out of it today.
25:30
Adio is the ad CRM >> build scales and grows your company to the next level.
25:39
>> We our routine our routine is so dialed in that if we go to bed like 2 hours later than normal, it just throws everything off.
25:47
We had a veryaa very chaotic morning. >> Yeah. >> But we're back. >> We're back.
25:51
Uh Nano Banana is remarkable.
25:53
Look at this Golden Gate Bridge image.
25:55
It generates the image and also all of the diagrams around it.
25:57
Uh, >> this is how Tyler sees the world, by the way.
26:03
>> Sundar says, "You went bananas for Nano Banana.
26:05
Now meet Nano Banana Pro.
26:05
It's state-of-the-art for image generation, editing with more advanced world knowledge, text rendering, precision, plus controls. Built on Gemini 3.
26:13
It's really good at complex infographics, which is awesome.
26:17
Uh, much like how engineers see the world. That's very fun. >> Even see that thing.
26:21
We were playing around with it this morning.
26:23
It is It is absolutely wild.
26:26
>> It's really, really good. The text is flawless.
26:27
There's there's just truly it doesn't make any mistakes with text anymore.
26:32
Uh it, you know, we were we were in the era of like the text looked good, but you would still see a double S every once in a while.
26:39
One thing would go wrong.
26:40
Um and uh and now we're in a much better uh spot.
26:44
There's still one test that it fails.
26:46
That's the where's Waldo test.
26:49
If you have it go generate a Where's Waldo?
26:51
It will not be uh it will like you you you will clearly be able to real this morning. Did you generate that?
27:00
>> Uh I had Tyler generate that one. I generated another one.
27:03
>> This this is like the most funny uh where's Waldo ever.
27:06
Uh John like says like where's Waldo?
27:09
And I'm like are are you are you messing with me? Is this a joke?
27:13
And it's like this massive crowd of people and then Waldo is standing on a stage going like this. [laughter] >> Yeah.
27:19
For some reason it did not hide the Waldo at all.
27:20
the Waldo at all. like Waldo was jab just perfectly in the center very obviously >> most novice where's Waldo >> it was very no and then also there were actually two Waldos uh and and as you dig in normally when you when you're hunting around a where's Waldo there's different little sub stories that are
27:35
happening and this was more just like a generic crowd I mean still remarkably impressive but uh but that is currently my go-to uh evaluation and uh and we got a lot closer but uh it's not it's not super human it's not super Waldo yet um anyway we have Doug Olaflin from semi analysis in the re waiting room. Let's
27:52
Let's bring him into the TVPN Ultra Dome. Doug, how are you doing? Welcome to the show.
28:00
>> How did you decide to take a vacation in the fall of of 2025?
28:03
You were off for >> No days off.
28:05
We're in the midst of the biggest.
28:07
>> You were like, "Oh, it's going to be it's going to be mellow.
28:09
No, there's not going to be any news. >> No days off."
28:12
>> So, dude, every single time I take a vacation, stocks always drop.
28:15
But, uh, dude, I did I proposed to my girlfriend in Japan.
28:18
That's the reason why I went. So yeah, big deal.
28:24
>> Oh, we're going to hit the golf. >> Massive news. That's massive news.
28:28
Anybody can pull together a $200 billion LOI, but to to find true love is beyond special. So congratulations.
28:34
That's >> what's 100 billion between friends today.
28:38
I saw the 100 billion uh the 100 billion Brookfield thing.
28:40
I was like, dude, don't even It's just another day, man. Another day.
28:43
Did you did you have somebody uh did like a semi analysis intern come up during your vacation and and go on your ear like sir Sarah Frier has requested a federal backs stop.
28:55
[laughter] >> So okay I mostly consumed it on a 12-hour lag and the 12-hour lag I was like holy And it's just like very funny to to get it in like slow motion where I'm like okay that was a bad that was a bad interview Sam.
29:07
I'm going to be honest with you.
29:08
I was like oh Sarah Brier.
29:10
And then I'm just like ooh ooh.
29:13
And then I'm like, but I do think um and the Fed, the Fed is what I think people are freaking out on the stocks wise, but it's just like this weird thing to witness in like slow motion half of on the other side of the world when everyone's asleep and It was just really weird. >> Yeah, totally. Well, welcome back.
29:29
>> Well, uh yeah, what's going on with Nvidia?
29:31
Uh take us through how you're processing the news.
29:33
Uh we we've been batting around two takes.
29:36
One was uh we're we're we're extra analytical over here.
29:40
The first take we had was uh Jensen was was seen drinking a beer and therefore he will beat Drinking a beer.
29:48
He was linked linking arms in South Korea. Just absolutely. >> This is our rigor. Our rigor.
29:53
>> So I I was confident uh that they were going to do quite well >> and then we just seem to be in the era where uh things beat on earnings and then immediately sell off for some reason because expectations are so high.
30:04
And maybe we're in that era now.
30:04
But how are you processing it?
30:07
>> Um I think it's like almost a perfect beat. Uh it's very clean.
30:09
You have like almost nothing to complain about.
30:14
Margins, which was like a story last year, doesn't matter.
30:16
Like they did a great job.
30:17
They had a pretty solid like a meaningfully above buy side consensus.
30:22
It's it's like a perfect quarter.
30:22
You have no problems with it.
30:24
Um but the thing is you're the biggest most profitable company or not most profitable, but like you're one of the biggest companies of all time.
30:29
Uh perfection is expected every single time you report.
30:33
So I think it's totally fine, dude. I'm I'm being serious. It's just totally fine. Like stocks do go down. >> Yeah.
30:38
>> People forgot about that. People forgot. >> They do go down.
30:41
They go to down sometimes, man. It's crazy.
30:46
>> What is the uh interpretation of or what should the read be on? Uh Gemini 3, the TPU.
30:52
Uh it feels like that's like uh um you know, is Nvidia still a monopoly?
30:56
if you can train the best model on all the benchmarks and without a single Nvidia chip, that seems like maybe a crack in the narrative, but does it matter at all or is it uh just uh irrelevant?
31:10
>> Um, I think it matters a little bit.
31:10
I think Google being really aggressive is really nice because like do they have power and they're like waking up and TPUv7 is going to be awesome and [snorts] anthropic and we're doing it we're doing like actual deals.
31:21
So like that's good I think.
31:23
that's good I think. uh because you need like like Gemini has been like I don't know asleep at the wheel despite inventing all this stuff and so it's really nice for to see them be back but I don't think it's um I mean I think it's a big deal clearly TPU is number two and it deserves number two I think
31:39
Nvidia being number one um what I really want to see is like why isn't there a new pre-training run from OpenAI like I got to I got to ask that question out loud again we've seen so many um RL scaled up versions but we know there's no new based pre-train And we know that there's Gemini 3 cooks because uh it's a new base pre-trained model. So like
31:57
So like where is that happening?
31:58
Like is it because the GB200's aren't stable enough?
32:02
Is it because like they're just totally not like dialed in?
32:04
I think that that's like the question to be answered.
32:08
And um OpenAI is just I don't know. They're not cooking. I want to see them cook.
32:13
So right now >> uh 3 >> on the on the base pre-train.
32:17
Is it still is it still fair to kind of set up that storyline with uh GPT4 now called 4.
32:26
5 now sunset used to be GPT5 potentially didn't really pan out.
32:32
There's been a lot of uh debate over what went wrong with that pre-train.
32:35
Is it fair to say that that was like an order of magnitude more compute spend like cost went into it?
32:42
Is there anything real about like it was expensive to serve it?
32:45
I've heard that bandied about as like why a lot of people said it actually was a better pre-train. It was a better model.
32:51
It would have been amazing but we just messed up something about the economics and so once we tried to deploy it it wasn't very economical and so it was slow and that's why we pulled back.
33:02
Not that it wasn't a good pre-train.
33:04
Not that it wasn't a good model.
33:07
>> I still think it's a failed run, dude. Okay.
33:09
>> I still think it's a failed run.
33:09
I don't think it got quite to where it should have been given its size and something was wrong with that. 4.
33:15
5 was decent and like a really good creative writer.
33:17
You talk about the economic side.
33:19
This is where I have to to uh to pump inference max right which got shouted out three times.
33:26
Uh so yeah, look man, I think the economics work now >> or could work with GB200 because it's like you know 10x better performance.
33:32
So you can >> in theory you probably could serve it but for some reason they still don't want to.
33:38
And >> it's probably something on the RL like it's just a bad base model that is not able to scale with higher chain of chain of thought.
33:45
Like maybe because of how much compute it takes.
33:47
It's you know the still distilled version wasn't doing as well.
33:51
All this stuff um matters and for whatever reason 4. 5 isn't it.
33:54
Um we know there's been we know there's been failed training runs and so it's like dude Open AI I want to see it and I I think we'll I think we'll get it right.
34:03
I think we'll get it right. like nothing makes them excited >> but like competition will there >> so so open AAI has a consumer business they have a front end for AI it's it's the brand that people think of when they think of AI
34:21
at some point you could imagine them not doing another scaled pre-training run because they're they're just like you know it's not really worth it to take it from you know this IQ to this IQ it's like the our our average users is just not really going to care. Meanwhile, if
34:34
Meanwhile, if you have a company like Anthropic, which is like an API business that like relies on kind of like raw horsepower capability intelligence and and maybe is like easier for end platforms to switch in and out of.
34:48
Like I don't know that they can afford to not keep doing the bigger and bigger training run, but do you expect OpenAI to at some point just say like, "Yeah, we're kind of good on the core product.
34:57
Maybe we don't even need to to to do the next run."
34:59
Dude, I think I think at the same time you say all that if they're being a consumer business.
35:06
U but you know, OpenAI has massive like they have coding FOMO, man.
35:11
They're really really really concerned about the coding models, right?
35:13
The I'm I'm sure you saw like the I can't remember what it's called, but I want say it's like project 2027 or no not project, but it's like the 2027 scenario or something like that.
35:23
>> Project was the right way agenda.
35:23
That is [laughter] Yeah, this isn't this isn't >> you're talking about uh AI 2027, the fast takeoff scenario scenario where OpenAI buys Ford Motor Company to make uh to make humanoid robots. >> Yeah.
35:40
To make more widgets, bro. No, no.
35:42
So, I think I think that um while like, hey, look, it doesn't seem like it's on track anymore.
35:46
I do think the thing about the fast takeoff that people feel very strongly about is better coding models means better AI agent uh you know you know AI agents and those AI research agents will make better models and that's there is a recursive loop there.
36:01
I think that that's where dude that's what like you know they were they had so much codeex fobbo and and despite all this man like Gemini still doesn't have the anthropic lead today. So yeah. Yeah.
36:13
I mean I think everyone wants that sweet bench.
36:15
Um, and >> yeah, I don't know.
36:17
I I I just think >> it's it's just like this weird I I think it's just like a perfect vibes time on the like the finance side, dude.
36:24
People are freaking out about the market uh Fed cut.
36:28
>> Um, it's not going to happen.
36:28
And and so it's like >> uh we I didn't I learned this yesterday, but this is like the second longest run above the 50 DMA, which is like, you know, stock chart men males astrology vibe.
36:40
Um the this the second longest run since like 1997.
36:43
And so it's just like we've we've been we stocks have been going up for quite some time and sometimes they can go down or even uh sideways.
36:51
And so I think uh people are freaking the out and >> it's just kind of a long long long powerful run.
36:57
And and and also I think that people are freaking out because like stocks go down, people's vibes get bad and then they're like, "Bro, maybe it's actually over.
37:06
>> Maybe it's actually over."
37:06
Like nothing changes sentiment like price, man. >> Totally.
37:10
Uh what did you think about the Financial Times published an article that uh was pretty I felt pretty misleading.
37:18
They said Oracle is already underwater on its astonishing 300 billion OpenAI deal.
37:22
And they said that because the stock >> this was Alpha Bill. It's their blog.
37:27
They're having fun, but they're rage baiting. They're rage baiting. >> They rage baited me. What What do you think?
37:32
>> They they rage bait pretty hard, bro.
37:34
Like let's be let's be clear.
37:34
Like I don't think uh I don't think you know making $400 billion of revenue is uh being underwater.
37:40
But I I if you if you're if you're betting on just the stock then sure yes they are underwater.
37:46
I think they're going to be their headline today from Alphaville.
37:47
It was like it was like who is open to auditor?
37:52
>> They're really taking shots at everybody. >> Um it's funny.
37:54
Um I I do well yeah sorry >> I I was going to say man but Alphaville like I don't know. Alpha doesn't cook man.
38:01
Their alpha is so is it's kind of >> it's kind of mid.
38:04
I don't know what to tell you, man. >> Petition to read it.
38:06
It's our Alpha Midville to beta to Betav. >> Betav beta boys.
38:12
I I've enjoyed Alpha from time to time.
38:15
I I want to know about uh this Gemini 3 pre-training run.
38:17
Is there any way for us to understand the rough order of magnitude of compute or dollars that went into it?
38:27
I I from what I understand, Google has more of a distributed training system.
38:33
They train across data centers. Is that might be right?
38:35
And so before with like the GPT3 training run, the GPT4 training run, it was like they raise a bunch of money, they go build a data center or they acquire a whole bunch of GPUs and then you kind of see like there was this much energy that went into it.
38:49
This many GPUs were marshaled for it.
38:51
But it feels like with these Google training runs, they're harder to understand the actual scale of the investment.
38:56
But do you have a do you have a more solid understanding of how how big the Gemini 3 project was like from a capex perspective?
39:07
>> I have no I can't tell you because I don't know how big the model is like on a parameters basis but I do I I have a pretty good vibe that is multi-data center.
39:14
They were first to do that.
39:14
Uh Pathways has always been first in terms of like the OCS the distributed scaleout network like they've always [clears throat] or sorry scale across like they've always been first in that. Um, yeah. I don't know.
39:25
I I don't have an actual number, but uh I don't think the actual pre-training of of like the final run probably wasn't that much money. Um, right.
39:34
Uh, but the thing is all the experiments to get up to there, all the other things that uh that goes into training a really big model costs a lot in R&D.
39:43
And so I think uh the sing the final shot or whatever in terms of compute is probably poultry compared to like the actual total spent, right?
39:49
you probably have a multiplier of like 10x on top of it of what the final um number is. But I don't know.
39:55
It's probably Dude, it's probably a billion bucks if I had to guess. Um there >> Yeah. I don't know.
40:00
I'm just going to throw out a number. You heard it here first. That sounds right. >> Billion.
40:06
>> Uh because I mean we we we've heard about uh uh training runs that were, you know, like a couple years ago they were in the hundred million range and and that the the billion dollar training run was kind of rumored.
40:15
Uh, I wonder I mean I wonder if they if we get another 10x next year or the year after and we're seeing $10 billion flow into a single training run like from an SEC perspective does that need to be disclosed at some point?
40:27
Does this wind up going into the filings uh into earnings kind of uh just as an number?
40:34
>> Yes, that' be pretty sick honestly.
40:36
>> Yeah, I I would just imagine that at a certain point investors would want to know uh I mean it's like a mega acquisition.
40:43
It's a It's a significant slug of >> I guess >> I guess it goes into cost of goods sold.
40:48
Like I don't like AI accounting is like completely made up today.
40:50
So no, who who the hell knows?
40:52
Um but it's probably a cost.
40:54
>> I feel like it should be capex.
40:54
I I I I liked and I don't know what your take is, but I I liked Daario's uh framing of each model is individually a profitable company when you you spend a billion dollars and then you make a hundred million a month for you know a whole bunch of time hopefully.
41:09
But okay, the in order for it to be capex like to be an accounting brain is it has to have a multi-year lifetime.
41:14
And so if you if you train a model every year, it's R&D.
41:18
>> So that's that's the issue, right? You can't capitalize it. So I don't know, man.
41:21
Um >> well I mean I think >> well well I mean I have a big question about this.
41:25
This is something that I've been I've been going back and forth with with is like um we have seen that there is demand for 40 chat bots from like a group of Redditors potentially forever because those those people are like 40 is my friend.
41:40
I don't care about Gemini 3. I don't care about GPT5.
41:41
I don't care about 035 thinking max deep reasoning.
41:43
I want 40 and I'm willing to pay for 40 maybe forever. We don't know.
41:50
May maybe the turn rate will be very low for a long time.
41:52
And so you wind up with this weird thing where you can actually amortize 40 over years with that cohort.
41:59
Now we don't know how big that cohort is and what the churn will be, but it could be 10 million people for 50 years.
42:02
It's just like their buddy.
42:05
And and I'm wondering if the same thing will happen in businesses where you have some company that's like we have a model that is 40 level intelligence or Gemini 2.
42:14
5 and we have no reason to update to Gemini 3 because this this model just sits there and it looks at papers, scans them, summarizes them and it does that a million times a day and we're happy with that and we don't need it we don't need it to be more intelligent ever.
42:29
So we're just going to keep that workload going in perpetuity and we'll leave it on A100s if we need to.
42:34
like we we don't need to go to the the latest and greatest.
42:38
Do you think that's going to happen in the enterprise?
42:41
>> I mean, I I don't know if it'll happen.
42:43
I mean, enterprise just like let's get like more enterprise bullshitty.
42:45
Um people have to have price raises and you have to be like, well, why did you raise my price?
42:51
And they the single best way to do this is say uh we spent more compute, we have a better model, we do something like that.
42:57
But I also want to say like in the consumer side something that's like a good example is like dude Runescape Classic is probably like a perfect case study of this.
43:03
People want to play Runescape Classic.
43:06
to play Runescape Classic. They don't give a about like the and like obviously Runescape Classic is like kind of become a fork universe and there's like a lot of other stuff but it's run by like 10 people bro and there's like
43:15
millions of probably like you know tens of thousands hundreds of thousands of people who play it and like yeah I I wouldn't be surprised that we see like these long lived little projects that are really stable and they're like dude no notes do not change it. I don't care. I don't care. I want to play this one.
43:28
I want to use this model forever.
43:30
Um, that's probably like a really good example of like a but I feel like that's like a niche and it's very hard for you to underwrite everything becoming these like weird cohorts.
43:39
That's like a massive fragmentation of the internet and everything like everyone just has their one little like, you know, you know, freeze all my my memory at this exact point. No new information.
43:48
You're my favorite version of Gemini 3. 5 or whatever. I don't know.
43:52
whatever. I don't know. I I think that that's it's kind of hard for us to be like like and also dude that that's not AGI like I if you're talking about like vibes that's extremely depressing like if we're talking like last year to now that's like so depressing like that's why I think this is like >> I think this is why why markets are sad
44:08
people are sad they're like dude you're telling me 40 is all I want then what are we buying why are we spending 100 gigawatts so I think >> I think we're just in a weird time dude it's in it's you know the market didn't go down at all and now the market's going down and everyone's It's getting sad, >> but but the the the the leverage is still coming in. >> The the the circular deals haven't
44:29
>> The the the circular deals haven't actually hit the books yet.
44:31
They've just been announced.
44:34
Like that's we got >> Didn't Didn't Jordan from semi analysis say that uh uh there was potentially going to be like an H100 index that retail investors could like buy into or something like that?
44:47
Like I'm just interested in like how many different pools of money haven't like come online to the AI trade yet.
44:51
Uh obviously private credit is coming online now.
44:56
There's obviously corporate debt.
44:58
There's just sucking down all the big tech earnings.
45:00
There's also potentially like retail traders getting in on the action one way or another.
45:04
Uh if some of these foundation model companies go out and go public.
45:09
>> So the last note I had my team who is sitting in the office right next to me and behind me uh do before >> you're in an office right now.
45:15
It looks like you're in a forest.
45:17
>> Yeah, I love the water. >> We are in a forest. Thank you.
45:18
To be clear, uh, dude, we don't we we are squatting.
45:23
Thank you to our squatting overlords who let us work here. We're sick.
45:27
We're super happy about that.
45:27
But, okay, if you just do the math, man, because uh here's the thing.
45:31
I think uh all the hyperscalers could raise like $2 trillion.
45:36
>> Like, I really think the number is [clears throat] so large.
45:38
In fact, I am trying I'm trying.
45:41
>> I was just looking at the free cash flow and then you multiply it by 10.
45:42
If you if they were paying 10% interest, like 10% interest, and it's trillions of dollars because they produce so much so hundreds of billions of dollars. No. >> Okay.
45:53
So, I'm I'm going to I'm going to I'm going to give you the maxed out version of how I think about what we could do.
45:58
So, the >> leverage max the new report from analysis leverage max.
46:05
>> Also, also I've been told by I've been told by my corporate overlords, you have to star the inference max. That's super important.
46:10
I I you have to star inference max on GitHub. Sorry, before. Okay.
46:14
So, how how Max on GitHub, please. >> Thank you. >> Cool. >> That'll help. Okay.
46:20
So, I think I think they could probably raise something like $6 trillion by9.
46:28
>> And And is that like a 5% interest rate you're assuming on like corporate debt basically?
46:31
And that's >> we um Yeah.
46:33
So, we just essentially we're we're doing the current corporate interest.
46:38
We're just saying like, okay, the current market rate.
46:39
There are actual problems with how this is done, but like let's use Meta.
46:42
Meta is the most like the most aggressive version of this.
46:45
You completely do all your data center capex off the balance sheet.
46:48
You have blue come in pay for all that.
46:50
You do a sale lease back and then you spend the rest of the money just buying GPUs and you can probably do like >> you could and then they can >> issue debt in the market that's like 50 bips above the government. Yeah.
47:03
>> And also the rating agencies are like okay as long as you don't have more than one turn of debt by 2029. Yeah.
47:10
>> Um >> you're good to go.
47:11
So that so our we did that number for all all of them for all the hyperscalers.
47:15
Exoracle Oracle's pretty tapped out. $6 trillion.
47:17
That's like the that's the the big number. That's great.
47:22
>> Um >> so but here's the thing.
47:22
So if if Oracle's tapped out already and they're about to spend four years where free cash flow is going to be negative, how does that actually work?
47:34
>> So here's here's the thing about this though is free cash flow doesn't like free cash flow goes negative if you assume there's no revenue growth.
47:39
But um this is like kind of like a shale well.
47:44
Okay, you get a lot of your money on a GPU cluster up front.
47:45
Uh let's say fiveyear economic life.
47:48
People are gonna fight me about this, but whatever.
47:50
You you will have your payback for a brand new cluster in something like 18 months.
47:57
And so after that on the like let's say on the 18 to 24 to 36 months, which is like the 2 to three year, you're just going to start now you're going to start to gather in cash.
48:06
And that cash you can go turn around and borrow more against or respspend again.
48:10
And so that's where this like you know the shale one of the reasons why shale went so insane in terms of supply is like 12 month payback period which is way more insane than what this is but like um 12 month pay so you get all your money back and you can just do it again do it again. Do it again.
48:26
So I think next year Oracle will make a lot more money and they're going to be able to raise against a lot more money.
48:31
So yeah but they're they're tapped out this year. >> Yeah. Yeah. That makes sense. Uh okay.
48:36
Okay, I have kind of a lightning round because I know you have a heart out in a few minutes.
48:40
>> What is going on with Coree and Core Scientific?
48:43
Like they coreweave is is reliant on core scientific.
48:47
They they've had it seems like some kind of frustrations around getting >> capability delivered from Core Scientific.
48:56
They tried to buy Core Scientific.
48:57
Core Scientific rejected it.
48:59
Core Scientific has now traded down 20% or so since the acquisition uh was attempted.
49:05
Uh but can you explain that dynamic and why the core scientific shareholders are kind of still holding out at this point?
49:16
Okay, so Kors uh got offered to be bought in all equity and this was before all this all the other Bitcoin miner like energy names ripped and then I think a firm called 2C's wrote this thing be like hey look at everyone else's results and how much they've ripped and you're telling me you're selling out at this price and so rightfully if you do the math you're
49:35
like maybe we should just not get sold or we should uh deal break and we should ask for a higher price and so most of the investors went for a deal break and asking for a higher But uh you know people who are like the stock does go down and you have a shareholder turnover when when you reject the deal because like a lot of people are in it for the deal and then they have to sell. They're like no more
49:53
They're like no more deal I'm selling.
49:55
And so that's like uh that's pressure.
49:56
Um but at the same time Core Scientific is not delivering their Denton facility on time and that delay is like kind of a big deal for Coreave. Yeah.
50:06
So I mean that's that's like the the the Sparkos version of it.
50:07
Um, I think the problem is like you look at iron or something like that and you're like, "Wait, wait.
50:12
Why isn't Kors getting the iron multiple yolo, this is a 2x and that's how the deal broke." >> Got it. That makes sense.
50:18
Uh, product idea for you guys.
50:21
Maybe it's something you're thinking about.
50:23
Maybe it's something that doesn't make sense, but uh I was pitching John last week on an idea for something called a semi analysis product called diffusion max.
50:34
uh which would be like how what I want to understand is like how is AI actually diffusing across a bunch of different key industries.
50:41
So legal accounting you can just go on and on marketing on and on and on actually understanding like I would I would want you guys to have phone calls with like thousands of business owners and employees in each of these different like categories and then give us a read on okay are they actually laying people off because of AI?
51:00
Are they hiring more people because of AI?
51:04
What tools are they actually using?
51:04
Are they getting a lot of leverage?
51:05
Are they increasing earning >> margins? >> Yeah. Is it affecting margins?
51:09
Um, and I don't feel like that I I I don't know that there's like a definitive data source that I trust on that.
51:15
And that's something that like >> I only trust semi analysis for everything. >> Thank you. Thank you.
51:20
It's we're the only source of truth, bro. Um, thank you. Thank you.
51:22
I I want to I appreciate it.
51:25
>> Bitcoin max for semi analysis. [laughter] >> Same dude.
51:29
Um, so, so, uh, that sounds like, uh, I need an AI agent to call, you know, like 100,000 people, but, uh, dude, honestly, that kind of survey work is stuff that we're really interested in, but I don't think we're, we're Dude, we, Dude, there isn't something we haven't thought about. Yeah.
51:42
But we have a lot on our plate.
51:44
It's like a throughput problem, man.
51:45
Like, it's another important, too. Yeah. >> Yeah. Yeah. Yeah.
51:49
We're doing We're really really interested in the energy side.
51:52
Like, we're going to do a grid by grid breakdown.
51:54
like we're very focused on all the I mean we we like we're putting a lot of effort and and energy energy into it.
51:59
Um and I'm really excited about that but even that is still probably a little bit far out.
52:03
still probably a little bit far out. So we each of these like bets take a little bit of time and you have to reinvest in them and give them time to work out and so yeah man I would love to do like I don't know diffusion but the problem is
52:14
diffusion max is a bad name uh you know diff it sounds like diffusion right but like >> AI penetration max who the hell knows don't name it don't don't name it that sounds >> yeah another question another question uh XAI and Nvidia announced like a new data center project in Saudi yesterday. Uh I don't know if you caught that. Do
52:35
Uh I don't know if you caught that.
52:35
Do you see >> uh do you see XAI just getting into the cloud >> into the AI cloud business and and helping power uh and helping basically deliver infrastructure for other other companies?
52:52
>> I you know, no, but that not until this conversation. But they're the best.
52:57
They're the best at being quick.
52:57
So, if you don't if you have like infinite capital and like the zoning laws can be whatever the hell they want them to be, I would sign up XAI to put up a cluster as fast as possible.
53:06
They're the quickest with Colossus like I think they literally have the speedrun record and uh Saudis want it, dude. They want it so badly.
53:14
And I think um with this new uh we're like allowing to export it and I mean yeah, if they can buy it, bro, Xi is going to be like give me money in my pocket.
53:23
I'm going to make a Saudi Gro and then I'll make a Grock 5.
53:25
So, I think that X, dude, X is down for business and I think they like Tesla's always been um supported by a lot of different funders and I bet you some of the people who took X uh private probably were Saudis. So, yeah, might as well. >> Makes sense. >> That's true.
53:42
They Yeah, they actually were. I remember.
53:44
>> Uh quick take on the on the Brookfield deal.
53:46
You mentioned it early uh in in uh after you joined. >> Mhm.
53:51
100 uh 100 billion, but you should look at the actual number.
53:54
It's 5 billion committed.
53:55
Um, so just just [laughter] so ju >> the press release economy allegations. >> Yeah. So so so yeah, dude.
54:02
This is the press release economy, man.
54:05
I mean I mean you can dude you saw the 10Q, right?
54:09
It's like >> we got to do a deal.
54:09
We got to do a press release.
54:11
We'll pay you a hundred billion if you pay us hundred billion.
54:16
>> Dude, I think we can make that work with our accountants.
54:18
And then >> I think we'll write we'll write checks to each other.
54:23
We'll hand them off right at the same moment. Exactly. Yes. Right there. Right there. Right there. Yeah.
54:29
Like >> the economy will just circulate, bro.
54:32
And then we can use that money >> to raise capital. >> Yeah.
54:36
>> It would be beautiful. Holy. >> No. No. No. No.
54:38
We We don't want to We don't want to take advantage of it and raise capital.
54:41
We just want to We just want to make fun of the Yeah.
54:46
We want to aura farm everyone who's doing it unironically. That's what >> I mean. Yeah. Well, okay.
54:50
>> I mean. Yeah. Well, okay. My most Galaxy brain take from Nvidia actually that I feel pretty strongly about is um >> and this is you know in our call with them they're like you know this $10 billion check is like kind of pennies compared to what we're trying to do the real game we're trying to play and if you think about it uh being closer to
55:07
their customers and understanding what they're doing is probably the number one thing that they have to do as Nvidia to understand where the technology is going right and so I think if you if you think about it it's actually just R&D opex bro it's just a check that you pay in order to make sure that you're closer to open AI and anthropic and you know exactly what's going on in their data centers. So that's the that's the most like
55:28
So that's the that's the most like >> bullish take I can think.
55:30
But realistically, man, the Fed said something and everyone's freaking the out. [laughter] >> Yeah. Yeah, that makes sense.
55:36
Um yeah, I mean on that uh are are are you referring to that that line that's getting shared around from the Nvidia earnings around the quality of the deal with uh with OpenAI versus the strength of the deal with Antropic? Did you read into that?
55:51
Did you read the same thing that everyone else read into it?
55:53
>> Yes, we we did read into the same thing that everyone else said.
55:55
Um, it's actually kind of funny though because in the same language they're like the opportunity to invest and it's like yeah, it's just like they glazed Samma like they glazed him a little bit, but then they also were like, "Yeah, this also couldn't happen at all."
56:06
also couldn't happen at all." So I think um I mean dude a lot of the press release and the press release economy as you know you just say the biggest number and you're like dude until 2030 you have let's say I'm going to invest $600 billion in the United States I'm going
56:21
to be meta right uh I'll invest 200 billion from here to 2028 and then uh 400 billion from 2028 to 2030 right you just push it into the back half and then like if it comes it comes right that's that's how you do it >> I feel like people could go further here too I mean Ray Kerszswhile famously said, "Singularity 2045." You should be
56:39
You should be doing RPO all the way out to 2045. >> Why not?
56:45
[laughter] >> Yeah, I I'll just have my children.
56:48
Actually, uh they'll uh we'll do a deal with our children.
56:50
Hey, the my big my favorite big number of the week was uh NBS >> was hanging with Trump and he was like I said $600 billion yesterday.
56:58
But actuallyion let's actually and and Trump literally like hit went like like this and like hit him on the knee like he was so he was so happy to just hear that one trillion.
57:12
>> He's like finally finally someone said the trillion, bro.
57:14
Everybody, everybody gets around Everybody gets around Trump and they just like detach from reality a little bit and they just start saying like Zuck Zach had this at the at the AI dinner, remember?
57:24
Like he just threw out he threw out a number and had to had to correct it the next day.
57:29
>> He's like uh what what number are we going with? Oh, 600 600 600 billion. Yeah, dude.
57:34
I mean it's like uh it's like his uh his his warp field is that uh everyone just says the biggest number around him. I love him.
57:40
Like it's kind of it's like kind of ridiculous, dude. >> It's powerful.
57:43
[laughter] It's a it's it's uh it's real stimulus.
57:49
>> It's it's stimulating. That's for sure. >> It's stimulating.
57:50
I don't know if it's real stimulating.
57:53
>> You need a semi- analysis plan like like a a subscription where it's like $5 a month for the first 20 years and then 20 billion dollars in 2045.
58:02
I will sign up for that and then you can book it and diffuse it back. >> Defer the revenue.
58:09
It would beact cancel anytime.
58:14
>> We need to bring massive RPOS and press release economy to the to the newsletter analysis.
58:19
We need we need a Substack feature baked in for this.
58:22
>> We can do we can do we can do a deal, bro. It would be great.
58:25
We'll do Yeah, we'll we'll have a preliminary $1 billion advertising deal.
58:28
How >> we'll do some circular [laughter] economy.
58:32
>> Well, thank you so much for taking the time to hop on.
58:34
Uh we I know we got Congratulations again. So so happy. Thank you.
58:39
I appreciate the announce. >> Inference max. Inference max. Inference max. >> Yeah. Inference max. Inference max. Inference max. Inference max. Inference max. >> Go star it on GitHub. >> Okay.
58:50
Did you know did you Sorry.
58:50
Did you are you familiar with We actually made a video of Dylan at our offsite just screaming analysts.
58:55
And I was like that's literally it sounded exactly like that.
58:59
I was like what the How do you know about that?
59:00
Um [laughter] like it's >> spies and we have spies eyes and ears everywhere.
59:07
[laughter] >> Your information is incredible, bro. Holy crap.
59:11
>> I think it's just we're just having fun.
59:13
Thank you so much for stopping by the show. Busy day. We'll talk to you soon. Have a good one. >> Yeah. Nice to see you. Take care. >> Goodbye.
59:19
Let me tell you about Linear.
59:21
Meet the system for modern software development.
59:22
Linear is a purpose-built tool for planning and building products.
59:26
Our next guest is David Chang.
59:26
He is an American chef, restaurant tour, author, and TV personality.
59:31
I believe he is in the reream waiting room.
59:33
He's the founder of Mafuku restaurant group.
59:38
>> Not sure we have him quite yet.
59:40
>> Well, we can also uh go all over the world.
59:42
We have lots of content today.
59:42
We will work on getting him in the studio.
59:47
Uh in the meantime, I will also tell you about fall to build and deploy AI video and image models trusted by millions to power generative media at scale.
59:54
Um, speaking of generative media, Gemini 3 Pro image only has an 8% error rate when generating text.
1:00:04
OpenAI's model is at a 38% uh rate.
1:00:08
And so there's been a very significant quantification of the improvement in Nanobanana Pro, I believe, or Gemini 3 Pro image.
1:00:16
I don't know exactly the difference in the naming conventions. Do you know?
1:00:20
>> Yeah, I mean I I think so.
1:00:20
Originally, Nano Banana was like the uh the insider like that that was the code word and then then they're like, "Oh, let's just bring >> Interesting.
1:00:29
I was wondering how that happened."
1:00:31
Yeah, >> because uh it's very same thing.
1:00:33
>> It's cool and it's quirky and it's actually very on brand for Google in my opinion to run with a keyword like this.
1:00:40
But um but at the same time, it's added a lot of confusion because they've done so much work just to establish the Gemini brand and now they're also have a nano banana brand and it's a little bit confusing.
1:00:49
But uh we can talk about that more after our next guest joins.
1:00:51
We have David Chang uh in the studio. Welcome to the show. How are you? What's up, guys? >> Good to meet you.
1:00:59
Thanks so much for taking the time to talk to us. >> Great. Great fit, too. >> Looking great. Oh, >> you're ready. Ready for anything? >> Yes.
1:01:06
>> Well, I'm uh prepping out for a bunch of things right now.
1:01:10
And then again, I'm planning and we'll be in Vegas uh by dinner >> for F1, right? >> Yes, sir. >> Fantastic.
1:01:17
>> We'll be there Saturday.
1:01:17
Can you help uh everyone in uh in the audience understand just the shape of your business between >> the shape of your empire?
1:01:25
>> The shape of your empire.
1:01:25
Empire is the correct term.
1:01:27
Sorry, not just business.
1:01:29
>> Um [laughter] >> well um it's it it's it's restaurants.
1:01:34
We have some quick service, fast casual. Yeah, >> we have casual.
1:01:37
We have um fine dining and uh we have couple spots in Vegas.
1:01:40
We have a couple places here in Los Angeles.
1:01:44
Um I think it allowed us the pandemic allowed us to sort of refocus exactly our growth strategy. >> Yeah.
1:01:51
>> Instead of trying to open up all around the world um which we had been doing until 2020.
1:01:55
I think um we we we uh >> had a lot of sort of plans in place for doing CPG.
1:02:02
So, >> uh, that like many other things in the world at that time sort of expedited the plans and and we, uh, went head first into, um, you know, we had dabbled in making some sauces here or there, but we had always wanted to go into making noodles and and, um, that's sort of a good part of our business, too.
1:02:20
So, it's it it's equal parts restaurants, even though they're now split out into two completely separate entities.
1:02:26
Um, they take up a lot of my time.
1:02:29
And then me, it's it's mostly media stuff these days.
1:02:33
>> So, I'm I'm I'm dressed up right now because we're >> doing um practice runs for our Netflix show on uh uh which we air at 400 p. m.
1:02:42
Pacific Standard Time, dinner time.
1:02:44
>> Have you Have you ever tried live streaming?
1:02:45
We've we've we've wanted somebody to do the version of our show that's just like a live like a daily live cooking show, and I feel like you've got you've certainly got the personality for it.
1:02:55
It's it's basically a full-time job, but I could see that being a hit.
1:03:01
>> Um, we we we tried to do that.
1:03:01
You'd be surprised how adverse I think a lot of people that are running networks still because uh I think if anything, it might have to go on um you know, one of the free streaming platform services, but that's not out of the question.
1:03:14
It's certainly on the long-term projects list for major do media is to do uh just all day, you know, totally transparency.
1:03:23
What you see is what you get.
1:03:23
And you you do see some people doing it on Twitch.
1:03:27
>> Yeah, that's what I'm saying.
1:03:27
You don't need the network.
1:03:28
Just >> I was wondering >> create a Twitch account.
1:03:30
Create an X account, >> but uh I'm currently pretty preoccupied with Netflix and Amazon >> and and Spotify these days, especially since our podcast is moving over to Netflix. >> Oh, really?
1:03:42
>> You're in that you're in that package. >> Yeah.
1:03:43
Talk talk about I mean, obviously there's a bunch of uh particulars that are are probably under the under the hood, but uh what excited you about taking a podcast to Netflix?
1:03:52
It's an interesting strategy.
1:03:54
It wasn't on my list of predictions for what Netflix would do.
1:03:58
Uh what have they shared with you about how they'll surface that, what the audience might be like, what the Netflix viewer is looking for in podcast content?
1:04:06
I find that whole strategy fascinating.
1:04:09
>> I would love to answer all of those questions, but I don't think I'm the >> Okay.
1:04:12
Anyway, >> I I I I would I think if I was the person to answer that, it would be funny to both Spotify and Netflix because there are other people that are, >> you know, certainly designed to answer that. We'll have some.
1:04:23
>> I will tell you that it's it's it's been something we you know, we have what 600 plus episodes of our podcast and it certainly changed over the years.
1:04:30
When I first did the podcast, it was much more of an insidider take on the restaurant industry.
1:04:34
Um and and sort of a sneak peek in terms of the thought process of opening restaurants or, you know, pre-opening of a friend's restaurant like my buddy, you know, opened up Angler in San Francisco.
1:04:45
We we sort of gave everybody a sneak peek of the, you know, the the the the philosophy behind it.
1:04:51
And then the pandemic happened and you couldn't travel.
1:04:53
So it just sort of shifted and we've been waiting for this moment quite frankly where instead of talking heads cuz food is the one thing which is sort of dumb, right?
1:05:02
It's the one kind of podcast that we can't really react to culturally.
1:05:07
Uh I'm very close with Bill Simmons.
1:05:08
I'm part of the Ringer podcast network >> and um you know we can't watch some movie and react to it and we certainly can't watch a Monday night football game and then react to that either, right?
1:05:17
So food is so ephemeral and in the moment and also more importantly not necessarily scalable.
1:05:24
>> So um >> and now with video right and more people um you know with all the data and like more people are watching podcast than actually listening to it.
1:05:35
>> You know there's certainly a lot of people still listening to it, don't get me wrong, but it's it's certainly in the near future going to outpace it if it hasn't already.
1:05:41
And now it gives us the opportunity to evolve again and to offer a podcast that is somewhere between a TV show and a podcast. >> Yeah.
1:05:52
>> And I can do that cuz cooking is something I can do that you know other people that are maybe doing interviews or such as yourselves like you know would you be cooking and doing this interview right now?
1:06:01
I don't know if a lot of people would sign up for that. It' >> be very hard.
1:06:04
I [laughter] suspect >> it's hard to eat and and uh do a podcast eating mic possible.
1:06:08
But even cooking >> uh we were we were talking earlier uh on the show about kind of the delivery app experience like the the dynamics of like tipping in and delivery apps.
1:06:20
Travis Kalanick >> uh uh uh was commenting on on on a on a post of ours yesterday.
1:06:27
He's got a new product called Picnic, which is like a front-end delivery platform uh just focused on corporate meals.
1:06:34
And like the key value prop is uh no tipping and no fees.
1:06:39
So they're focused on like higher higher volume orders.
1:06:41
So like teams of like 25 people plus.
1:06:44
I'm curious like how like getting I wanted an updated take on from you on like what it's like working with the delivery platforms today, where you think there's opportunity.
1:06:56
uh right >> and all that stuff.
1:07:00
>> Well, um I don't know if many people know or remember in 2016 we created the very first to my knowledge um there might have been something called a ghost kitchen but no one called it a ghost kitchen.
1:07:12
We we teamed up with um uh Thrive and Will Gabri and Josh and um >> uh Caleb etc.
1:07:19
Great team and we opened up Maple and we were doing like 10,000 meals a day out of New York City as a full stack app. >> Yeah.
1:07:26
Yeah, we were delivering. >> I had no idea.
1:07:27
I I didn't I didn't realize you were behind that.
1:07:29
I remember I remember Maple, but uh that's cool.
1:07:33
>> And um yeah, like it's something that I've been wanting to do and had done for a long time because I saw that's where food was going particularly because of the the it's it's been a 30% cut for for some time on delivery fees and that's just not a sustainable model for the the delivery companies and the restaurants.
1:07:52
So uh it was a real uh opportunity to just sort of bridge that and to to do everything oneself.
1:07:58
Um and uh yeah and and so much so I believe in that we started another one with um called Ono which was more of a fast food.
1:08:05
So like Maple you might get a nice kale salad and a butternut squash soup. and Onondo.
1:08:13
We started with U Garrett Camp's um uh Fun Expo and we did which was like a cheese steak and fried chicken. So I was all in. I was all in.
1:08:22
We were probably just 6 years too early.
1:08:24
just 6 years too early. Yeah, >> it worked but not enough where it is today because >> but is it so is that is that model is the model like the hard thing is like cloud kitchens is the dominant like like company in that space but they're just incredibly like secretive right and so
1:08:44
it's hard for me to get a I haven't done you know much digging but it's hard for me to get a read on like is the model durable is is is there going to be a lot of value creation there or does it ultimately does that ultimately kind of fragment like a lot of the the restaurant industry has as well. >> Listen, I I I think it's like anything
1:09:01
>> Listen, I I I think it's like anything else in tech, right? Remember in the.
1:09:02
com bubble you had like tubeox. com, right?
1:09:07
Like >> yeah, >> there's only like three or four companies that really >> came out of that.
1:09:13
>> Uh same stuff what I imagine with all this AI right?
1:09:15
this AI right? um [laughter] you know like I can't even tell you in the food space how many companies are a food logistics company but now it's AI right like [laughter] >> it reminds me again in 99 when at least in New York City every company and I
1:09:32
won't say every company but a lot of lot of places were now changing their name to sort of pizza 2000 dry cleaning 2000 right it's just a sign of the times and [laughter] and uh I I just I think that in food delivery, you're you're going to have about three to four winners, if that. And certainly, I would never bet against
1:09:50
And certainly, I would never bet against Travis and and the team there at Cloud Kitchen.
1:09:54
Uh definitely not I I think what Tony and the team at Door Dash is doing is just unbelievable.
1:09:58
And clearly you have Uber and the Postmates guys.
1:10:03
>> So, um yeah, to me that's pretty much going to be the space.
1:10:06
And I think whence I'm not sure when, but when they are able to be more open and transparent about everything, people can be like, "Wow, that's a that's a pretty goddamn huge business." Yeah, >> makes sense. >> Interesting.
1:10:20
>> Uh, >> drinking culture.
1:10:21
And >> any restaurants been caught using AI generated imagery for their for their menus or any of the food delivery apps yet?
1:10:29
>> That's a good question.
1:10:29
No, but you know, like I I think it's been it's the same that's happening, right?
1:10:34
And that and like >> AI to me is like the getting the lowest common denominator things and sort of like crowdsourcing and just getting something that is not necessarily perfect but just good enough.
1:10:46
>> And I I just think that in restaurants that's basically been like consultants, right? Yeah.
1:10:52
>> Um they've just been that just like I can go I feel like I've been going to AI generated restaurants for some time now.
1:10:59
You know, it just hasn't been called AI.
1:11:01
[laughter] >> That's funny. >> It's a good take.
1:11:03
Uh are you excited about drone drone delivery?
1:11:07
>> You know, that's one thing where I thought it was going to be a total zero.
1:11:10
I'm dead wrong about that one.
1:11:10
I think it's definitely going to be a thing.
1:11:13
And um >> Well, isn't it is exciting as a chef to to know like if I make this, it will arrive hot?
1:11:19
No, it's not going to arrive.
1:11:22
Well, that's a whole other thing with this whole the one thing I will tell you under the food delivery space and I've talked to just about everyone under the sun over the past 10 years um that's tried to start up a food delivery company because they're like, "Oh, this guy's done it a few times.
1:11:33
Let me just sort of steal all the ideas."
1:11:35
And I'll tell them every time I'm like, "Unless you've created some kind of new technology to cook the food, it's going to be hard to really um make the food hot ultimately, right?"
1:11:48
like um how should I say this?
1:11:48
There's no new technology to make the food better. None.
1:11:55
So um the delivery drone unless it's cooking the food as it flies, it's always going to be limited.
1:12:03
>> An oven oven in the air basically. >> Yeah.
1:12:06
I mean like that's just the truth, right?
1:12:07
Like and also a lot of these places just have a bottleneck because everybody wants to eat at 6:30, 7:00.
1:12:13
So there's just there's not much you can do to make the food go out faster. >> Yeah. >> Or hotter.
1:12:20
And more importantly than not, everything can be delivered well.
1:12:23
Like French fries will never be able to be delivered well. Right.
1:12:25
The next step is going to be whoever makes the food literally right outside the house >> or apartment.
1:12:32
>> Have you heard any pitches on that? >> I have.
1:12:34
>> I have. any any [laughter] >> No, you don't you don't have to you don't have to give names, but like I'm like it gets to the point where it's like a street we just have like like a massive proliferation of like street carts and it's and maybe maybe like >> well I mean I I'll tell you this a lot
1:12:50
of these p I haven't been to it in a couple years because I just don't want to do it anymore but um a lot of times there'll be a very success like a a chef that's worked 20 years at a three mission star restaurant making the food, you know, and it comes out and I'm always like, is this person going to be making the food? You're going to get
1:13:08
You're going to get this quality talent making the food at every single sort of satellite location.
1:13:16
And the answer is they haven't even thought that far.
1:13:18
And the other answer is that's not a reality.
1:13:20
You know, you might as well be pitching me a unicorn, literally a unicorn with a horse and a a horn on it because it's not going to ever work.
1:13:28
Because that's the hard part about this business, right?
1:13:31
uh cooking is still a physical endeavor and for all the the the VC money and tech money, it can't sort of solve that riddle of how do you make physical labor go away >> or done better?
1:13:50
>> So, you don't you don't think we're going to you're you're very bearish on the humanoids up. >> No, no, no.
1:13:56
I'm not bearish on that either.
1:13:57
Um I spoke to somebody pre- pandemic.
1:14:00
We did a show and we did some research and and uh uh a robotics expert and we talked to some people at Caltech and a few other experts and if somebody was like ah maybe 40 50 years away and uh I talked to [laughter] I talked to someone recently and they're like yeah we're probably 15 years away um from getting somebody that has a robot that has the dexterity of a high-end best-in-class sort of chef.
1:14:26
Um, so no, I I I it's going to happen.
1:14:30
If anything, I think you're going to see um the next 5 10 years, you're going to see robots.
1:14:36
You already see it in um I [clears throat] mean I again like I I don't think it's been a sudden, oh my god, there's robots.
1:14:42
There's robots in the kitchens all the time.
1:14:44
Like if you go to a good restaurant, there's dishwashers pretty much a transformer robot. >> It's amazing.
1:14:51
And that does the work of like 20 people.
1:14:54
>> Under pretty underrated >> underrated.
1:14:55
And and if anything, you're going to see machines that take the binary movements out.
1:14:59
So a fryer that goes up and down, a bathroom cleaner, things like that.
1:15:04
And already dishwashers are pretty advanced and that can handle very expensive stemware.
1:15:12
>> So finding somebody that can polish stemware, you know, you might get a wine glass that could be $250 per glass and if you have a hundred of those, that's quite an expensive inventory for a restaurant.
1:15:22
You need someone specifically trained to do so. >> Yeah.
1:15:25
And that's a hard position to find.
1:15:28
So yeah, that that kind of position will be a robot. No questions about it. >> That makes sense.
1:15:33
Uh what's the most overrated trend in food right now? >> Oh, man.
1:15:41
I'm trying to stay positive these days, guys.
1:15:43
[laughter] I I uh uh I I think I think there's a way to answer the question by just saying like there's things that can be popular now that are not dur like durable trends.
1:15:54
So, >> well, I I would say the the most annoying trend is that everything has to be the best.
1:16:03
>> This hyper hyperbole, right?
1:16:06
>> I have to have the best X. >> Mhm.
1:16:09
>> This restaurant has to be, you know, world class number one.
1:16:12
And um they hate to tell it to you guys, but I think most people don't wouldn't even know what best is if they ate it. >> Yeah.
1:16:20
Um, and I think for the most part, I'm just now on this mantra personally of does it bring you joy?
1:16:24
Does it bring you happiness?
1:16:26
And that's really all that should matter.
1:16:27
Um, it's such a relative subjective thing.
1:16:29
But more importantly, I'm just trying to tell people like good is hard to do. >> Mhm.
1:16:36
>> Like just good is hard.
1:16:38
>> And I think we need more people to sort of appreciate just good or like even boring good than the world's best.
1:16:43
Oh my god, this is the greatest thing I've ever had.
1:16:47
That to me is the worst trend in the world is, you know, and the media and sh we're all part of the problem too, right?
1:16:53
You know, all these lists and it's all stupid ultimately. >> Yeah.
1:16:58
>> But >> do you think do you think we'll ever get to a point where like in in tech in the technology industry, many many products get better with scale for like a variety of reasons.
1:17:10
food has felt almost always the opposite of that where you take an amazing concept in a in a tiny tiny little restaurant, right?
1:17:17
Like, you know, a thousand square feet and the second you add the second restaurant and the third it just kind of tends to get it tends to get worse and worse and worse and worse over time.
1:17:26
And that's just like there's one there's one, you know, maybe it's one chef who had a had an amazing idea and it and it and it just is very difficult to scale quality.
1:17:37
Are you optimistic that there could be any new technology introduced that would kind of change that dynamic or is that just kind of an iron law?
1:17:47
>> No, I I think it's not an iron law.
1:17:47
It's not like a law of thermodynamics.
1:17:49
I'm sure somebody could figure it out.
1:17:51
But you know the the you just mentioned something the fact that something that's great is not scalable.
1:17:58
Um and for years, you know, I've certainly tried to do it and scale these things.
1:18:02
I I just sort of spoke about this at Reed Hoffman's uh Masters of Scale conference, right?
1:18:05
Like I sort of everybody's at that conference because they want to scale an idea.
1:18:08
And I said, you know, the the easy ideas are to scale an idea in food that is um you say cheap, but affordable and mass-produced, >> right?
1:18:20
The other end is um high-end experiential dining, >> which weirdly has become very scalable because of its inaccessibility.
1:18:30
Um it would it's the equivalent of like getting front row tickets at the NYX or you know Chase Center or something like that because you're now eating at say the French Laundry.
1:18:39
No one else can get there.
1:18:41
You may not even appreciate the food but it's now a social flex.
1:18:43
It's cultural currency that you can sort of have and it's ephemeral and because no one can have it.
1:18:49
Weirdly, now that experience is weirdly I mean it's scalable because that actually is crazy marketing for the French Laundry, right?
1:19:01
And the demand for that kind of restaurant is through the roof.
1:19:06
And what I mean by that is restaurants, it doesn't have to be super high-end in in Napa Valley.
1:19:09
It has to be anything that can't be copied immediately. >> Right?
1:19:14
You can't watch a YouTube video and decide I want to open up a restaurant like this.
1:19:17
I can't just you make an easy faximile.
1:19:19
So it could be barbecue, it could be sushi, anything that is bestin-class that people have a hard time copying.
1:19:24
That's like the barb value.
1:19:27
So you have really affordable make mass-produced stuff on food on the one end.
1:19:33
The other end you have things that very few people are going to be able to experience or eat.
1:19:36
And you know that's been sort of uh elevated because of technology right in different ways.
1:19:46
But I chose sort of challenge the audience that you know because every I mean I mean you guys know I I I talked to a lot of people in tech and probably a lot of your peers and they're always wanting to know the next big thing in food and I'm going to tell them like the hardest thing that the answer that needs to be solved is how do you scale the middle >> right? Does that make sense?
1:20:07
like the mom and pop restaurants, the diner, the the the restaurants that are just like good again.
1:20:13
Uh how do you make it so they can survive?
1:20:17
Cuz they're like cultural banks.
1:20:18
They're great, but they're not they don't have the sizzle.
1:20:21
They don't have the >> maybe the bottom line that makes it sort of cool for investment.
1:20:25
Um, and again, it's not about creating a company that's like the pickaxes and shovels for that middle market restaurant, but there's got to be something else that can like be something that's gamechanging.
1:20:38
I don't know exactly what it is, but uh I never talk to people that are trying to make food concepts or invest in food concepts that are actually concerned about the [clears throat] middle.
1:20:48
And I'm not talking about credit card processing and like that.
1:20:50
I'm just saying like in general, >> um, there's a lot there.
1:20:54
That's the meatiest part of the food industry right now, but it's just too damn hard and nobody really wants to touch it. >> Interesting.
1:21:01
>> Have you seen any interesting experiments on like the capital side of of uh like new restaurant creation?
1:21:05
Has anybody tried to make like a Y combinator for for restaurants where there's, you know, a talented, you know, uh owner operator chefs can get some seed capital and kind of support uh to go from zero to one.
1:21:25
Um >> because I feel like you would have tried that by now.
1:21:28
>> We we have definitely tried it.
1:21:28
I I I won't say all there there. We've tried it.
1:21:32
I know a lot of restaurant groups have tried it.
1:21:33
I know there's spuns out there um that try to do this.
1:21:35
But um I would say um you know uh Ron Parker created something called hospitality NX and that's a website that is a little bit like a job board legal zoom but also a place for people that want to raise funds.
1:21:54
So that that's something.
1:21:56
Um but I think for the most part um it's not as organized as other you know at the end of the day it's because um it's hard to create an idea that uh um has a high barrier of entry in [clears throat] food. >> Yeah.
1:22:14
>> What and there's no moat to really create. Right. >> Yeah. >> Yeah.
1:22:18
on on the on the topic of like you know starting up and go to market strategies are there uh are there risks to like going too viral early?
1:22:27
Uh we've experienced a bunch of uh like rage bait in tech recently where people have sort of designed products that are that are in designed to the pro the whole product is just designed to enrage and go viral and then get some attention.
1:22:43
>> What do you mean by enrage? I I'm not familiar.
1:22:45
So, so, so a company a company made a product that is like a developer tooling.
1:22:49
So, it's like software to help you make software and they use AI in it.
1:22:54
So, there's minute there's time periods where you have a little twominute break and so they added the ability to gamble with stake while you're while [laughter] you're making software and that made a lot of people mad for for obvious reasons.
1:23:07
So, >> or just deliberately picking >> rage bait in food would be like a product that had like a single meal that has 1,00 g of protein.
1:23:15
Like I could see a restaurant doing that just for the just just to try to get people to make Tik Toks about it. >> I mean, yeah.
1:23:24
I mean, I don't know rage bait, but like again like this has been happening on for for forever anyway.
1:23:29
you know, doing something that is probably going to, you know, I've opened up restaurants that I guess have been like that, too.
1:23:38
You know, it just [laughter] you're not, you know, I think if anything, it's just taken to another level because I would say that a lot of chefs now when they're talking about dish, is it is it something that the younger generation will find uh appealing to to to to uh record?
1:23:56
Um, >> and that's what I mean.
1:23:59
It's like it's this it's vaguely experiential, but it's very ephemeral at the same time. So, um I don't know.
1:24:08
I want to be optimistic again. Uh I'm usually Mr.
1:24:10
You are over here about this, but I do think that with all of this aside, with all of this access, with all this democratization of knowledge, um because culinary knowledge with a younger generation is higher than it's ever been.
1:24:22
I mean, it's never been better to eat in America.
1:24:24
better to eat in America. uh it may not have the sort of the titans of the industry as it used to because things have sort of leveled out but eating today like I talk about this with people a lot in the industry that travel um you can find a great restaurant in every
1:24:42
city in America for the most part now >> it's pretty remarkable >> if you just look at that right so maybe New York or San Francisco or other metropolitan cities are not as great um they're still great but it's really broadened out and flattened out across the country. So, Oklahoma City and you
1:24:57
So, Oklahoma City and you know, places that are tertiary cities to most people are actually might have some of the best restaurants in the in the country.
1:25:06
And and I think that sort of pattern is what you're going to see throughout food.
1:25:09
[clears throat] And this is a long-winded way of sort of answering this sort of rage bait.
1:25:12
answering this sort of rage bait. I think because of that need to sort of find something that is going to create kind of some kind of spark um in food that is the catalyst that's going to cause people in food to get better at their craft >> because at some point all of that is just going to wash away and you're going to be left naked with
1:25:39
something and if you want to be able to have the real goods to show for it and I think that I really feel strongly that food is about to go into this very specific point of like a little bit like Japan where you can open up one specific kind of bakery that that makes one specific type of thing and you do it better than anybody else and you're going to see that here in America. I I feel very strongly about
1:26:01
I I feel very strongly about that.
1:26:03
>> Yeah, I I love that approach.
1:26:03
What uh what advice do you give to uh kind of emerging chefs on media strategy?
1:26:10
I think in in tech there's like like we tend to see kind of a high low strategy where you want to be like super online getting a lot of attention or you want to be kind of the the mysterious dark horse that's kind of going over the radar and there's like a messy middle that's probably a disaster. >> Yeah.
1:26:32
I I don't think that that pattern is any different than than what you see in food.
1:26:36
Um but at the same time, I think uh you know, I don't know if apathy is the right word, but I don't care about it as much anymore either.
1:26:45
>> Um because it just I know I'm not the only chef that feels this way.
1:26:51
It's just some people are doing it more than ever and getting better at it, but others I think are just sort of getting exhausted by the whole thing because um I just don't know what that best long-term strategy is.
1:27:04
>> Um and now you have an older generation of, you know, I'm I'm 48 years old.
1:27:07
I know chefs that I won't say who that are like clearly gotten a social media strategist or somebody because their content is really good right now.
1:27:19
And I we we I I still don't know which one works, right?
1:27:23
Um because once you feed that beast, you have to do it all the time. Oh, yeah.
1:27:29
>> And that's a lot of time.
1:27:29
So I I I don't know if the better thing is to just be word of mouth because ultimately all of this is is word of mouth. >> Yeah. >> Right.
1:27:38
And >> and do you build a relationship and that repeatability and and like there's my favorite restaurants in LA like they don't have to do marketing to me, you know?
1:27:47
I don't need to get an email.
1:27:47
I don't need to see them on Instagram.
1:27:48
I'm just going to go there like when I have the time, right?
1:27:52
Um and so I I think >> I mean, yeah, I think I think that's the zag, right?
1:27:57
Um >> but you can't do that unless you actually have a point of view that resonates with somebody. >> Yeah.
1:28:04
>> And if you are constantly sort of pandering and figuring out like how to execute other people's dreams, wishes, and visions, and what the hell are you actually making? >> Yeah.
1:28:13
And you don't want to get to a place where your content is better than better than the product.
1:28:16
And I'm sure that's like, you know, a lot of the more the more you time you spend on content, like the the more greater there is a likelihood that that it could get to that point, I think. >> Yeah.
1:28:27
I mean, but like do you guys care about what you see on social media still?
1:28:31
Like I I actually think there's a bifurcation that's happening with what people see versus what actually people are going to eat.
1:28:36
are going to eat. I do think that there's I don't know maybe the the steelman argument for the the viral over-the-top uh you know Tik Tok that gets me to go to a restaurant is that it can in some ways create like a shelling point and like a a coming together like a if
1:28:55
there's something that's trendy and and I and it's an excuse for me to pull my extended family my friends different people and it just gets us a an opportunity to kind of come together there and experience that like even silly, trendy, over-the-top thing. Uh I
1:29:08
Uh I think that there's something that can be good about that.
1:29:13
But um but it's certainly not like the primary uh reason why I go to a particular restaurant.
1:29:22
No, I mean that's the thing is like I actually [laughter] I I we're working on a show and I can't say which or where, but um you know sort of the thesis is we're going to take these lists that people find or things that are viral and actually >> go out of our way to avoid it. >> Okay.
1:29:39
[laughter] >> You know, go next door to the restaurant that you're supposed to go eat at. >> Okay. Oh, that's cool. That's very cool.
1:29:49
I mean sort of that that in principle, right? >> Yeah.
1:29:51
No, I like that as a philosophy.
1:29:53
>> It's just like the other thing is I I I I sort of mentioned it earlier in this the conversation about if somebody was tasting something that was truly good and remarkable, would they actually know what's good and remarkable?
1:30:02
And I I think currently we we again have a a knowledge uh uh that is greater than it's ever been in terms of food.
1:30:10
But, and maybe this the same way in fashion and architecture and film and other arts, but does your audience actually know what good is anymore?
1:30:21
>> Because I don't I don't know, right?
1:30:21
And and I'm not it's not trying to be snoody or an artist.
1:30:26
I'm just saying like let's just talk about wine right now.
1:30:28
If if I'm giving somebody like a like 1998 Ravino, you know, from B white [clears throat] Burgundy to somebody that has never tasted it before, I know that it might taste good to them, but will they appreciate it? >> Mhm.
1:30:46
>> Because this person might be more into natural wines than, you know, oxidization, etc. , etc.
1:30:51
So, it's like I'm not saying that they're not right, but I always joke like you can't, you know, my friend used to say um you can't um you can say that you can never say that Salaryi was better than Mozart. >> Mhm. >> Right. >> Yeah.
1:31:08
>> You he was good, but he was not better.
1:31:10
And that's just sort of unequivocal.
1:31:10
And you can appreciate Salary, but you can never say that he's better than Mozart.
1:31:15
My concern is people don't even know who Mozart is right now. [laughter] >> Yeah.
1:31:21
And and I and that's sort of my concern when it comes to um sort of social media and food is who's deciding what is actually good.
1:31:32
>> Just because something looks good doesn't mean it actually is good.
1:31:36
>> And I know this is getting into a meta sort of philosophical conversation, but this is the I think about. >> Oh, I love it. >> That makes sense.
1:31:43
>> Uh last question on my side.
1:31:43
I'm curious how uh restaurant operators are planning around America just drinking less than ever. >> Yeah.
1:31:55
Well, that is the, you know, I I feel like the boy who cried wolf.
1:31:59
I've been sort of screaming this flag for a long time.
1:32:01
Um this has been this is the real existential threat.
1:32:06
>> Um like for example, LA, the biggest thing that happened in LA over the past 10 years in food was really ride sharing because people were getting drunk.
1:32:15
And you saw that in revenues.
1:32:17
Restaurants are going through the roof.
1:32:19
And if anything, restaurants was a bubble, right?
1:32:21
>> Um too many restaurants.
1:32:21
And I think we're still sort of in this bubble.
1:32:23
Um that's a whole another conversation.
1:32:27
But um I I think that you can see now, at least in LA, people are drinking much less.
1:32:37
>> I think you see a younger generation maybe taking some edibles.
1:32:39
They're just not, you know, the crazy thing is I Kids just don't drink anymore.
1:32:44
Like kids start when they start a tab, which is crazy to me.
1:32:48
They close it out every time. >> Yeah.
1:32:51
[laughter] >> What is going on?
1:32:53
Like they're never going to know what it's like to wake up at 3:00 in the afternoon being like, "Shit, I left my credit card at that bar.
1:33:01
I got to go back and get >> They're two responsib There's a responsibility.
1:33:06
>> It's hurting small businesses.
1:33:06
>> It's hurting small businesses. it is hurting small businesses and and um but I think that there the if you look at the sort of the only look at the blended numbers for most restaurants and beverage sales I think that it might look flat or down but it's actually I think way worse because once you split
1:33:21
out the 1% of the 1% that are drinking like these huge bottles of expensive wine right and that is through the roof right now again talking about the barbell experiential thing like totally >> people that are drinking things that no one else can really afford >> that's gone like 3x 4x of the past 5 years. It really has. And you know, It really has.
1:33:38
And you know, younger people are not drinking cocktails and they don't want mocktails because mocktails are actually way more difficult to make than a regular cocktail with alcohol in it.
1:33:47
But nobody wants to drink it for the same or more.
1:33:50
>> Why is it Why is it more Why is it more difficult just to actually deliver something?
1:33:55
>> Imagine if we were making the alcohol too. >> Yeah. >> From scratch. That's hard to do.
1:33:58
That's And that, you know, a normal restaurant ratio was 70 to 30%.
1:34:04
For the most part, you want 70% food.
1:34:06
I mean, this is not the I like >> roughly >> roughly 70% food to 30% bev sales.
1:34:11
And I think that is completely shifted.
1:34:13
And for a good restaurant >> Yeah.
1:34:18
I mean, like if you want 10% of your, you know, profit, for example, right?
1:34:22
Like >> something's going to give when you're down like 18% on dev sales, >> you know?
1:34:29
I think that's the average right now or something like that. 15 18%.
1:34:31
Um, so I don't have an answer.
1:34:31
Uh, food needs to get more expensive.
1:34:35
I've been saying that for a long time, >> but that comes across as terrible when people read that as a pull quote. >> Um, >> yeah.
1:34:45
>> Um, because it's already expensive.
1:34:45
So, I don't know what the answers are.
1:34:47
I will tell you that like, you know, it's one of the reasons why I invested in Athletic Brewing >> uh in in uh 2019 >> u because I saw the data within our own restaurants.
1:35:00
it was slowly going down year after year just a little bit like half a percent 1%.
1:35:05
>> But you know I and I'm I think that's what we can do is sort of figure out what the alternatives are. I don't have the answer.
1:35:12
>> But isn't one of isn't one of the challenges is like these non-alc products like somebody's not like there's not the incentive to have the second or third.
1:35:20
Like I I feel like a lot of this stuff people just have one they they get a little bit of the taste but they're not getting like a real they're not like getting they're not >> they're they're just like not getting drunk, right?
1:35:33
So they're not going Are you guys drinking as much as they used to? >> Absolutely not. [laughter] >> No.
1:35:40
>> You know, I feel like the way I used to was like Don Draper and Madman the amount I used to drink. >> Yes.
1:35:45
You know, and I, you know, part of that is just a generational shift, but I I can assure you if you talk to people under a certain age group, the younger Gen Z, they think of drinking like it's smoking cigarettes. >> Oh, yeah.
1:35:57
>> It's just not something they want.
1:35:57
I I've seen this in kitchens.
1:35:58
Like, >> you finished your 12, 14 hour day.
1:36:00
All you wanted was that cold beer at the end of your shift.
1:36:06
>> And now they don't want that.
1:36:08
>> And I just don't I'm just like, what is happening?
1:36:09
You know, and I'm not saying they're wrong.
1:36:11
It's just so that we're sort of dinosaurs, but interpretation like the data didn't change but it was contextualized through podcasts and there's a lot of health data out there.
1:36:21
Uh I I you could maybe call a little bit of the Huberman effect but there's a whole bunch of there's a long lineage of folks who have been like actually ringing the alarm bells on the health consequences of of drinking alcohol even in small amounts.
1:36:33
And so that's I feel like that's what's really cascaded. >> Yeah.
1:36:37
Maybe what we should do restaurants should start a lobbyist and just muzzle hover and [laughter] we'll be okay. Yeah. Yeah. Live life. Uh >> yeah.
1:36:46
I mean like the the dual pressure right now from from uh like just labor costs on one side and then and then uh just like declining alcohol sales.
1:36:56
Like it's just creating I mean I've seen some uh the place we go for breakfast adds like 4% on top of every bill for for benefits.
1:37:06
I'm I'm sure that that's helpful.
1:37:08
But like it's a very real cost, right?
1:37:10
It's now 25% between effectively for or 24% for service.
1:37:17
>> Uh >> at the end of the day, food needs to be more expensive and and I'm not it it just sort of has to and it can't be sort of be passed down.
1:37:24
I think I I've been talking about this for many many years.
1:37:29
I don't know why, but people have a real allergic reaction when it talks what to raising prices.
1:37:36
For example, I think you know it's good.
1:37:38
I I I'm pro when a restaurant jacks up their prices to like I'm hoping we see a restaurant where the the the the ability to eat there is basically like going to a Taylor Swift conference a secondary secondary market. >> Yeah.
1:37:56
>> You know, >> like that's sort of what has to happen.
1:37:59
And I I do believe there's going to be innovation.
1:38:01
And again, the problem with the restaurant industry as a whole to sort of um mitigating this decline in beverage sales is that we are too slow and proddding to to try new things out to um embrace new technologies.
1:38:14
And as my sort of spiel and joke about this as a whole, we're so goddamn allergic and slow to changing things, we still are using the metric uh imperial system instead of the metric system. >> Mhm.
1:38:29
>> I mean, that's so dumb.
1:38:29
The metric system is scientifically proven to be more accurate and more effective.
1:38:34
Why are we still using ounces, pounds? It's so dumb. >> America, baby.
1:38:39
It's because we're Americans.
1:38:40
We do things the dumb way sometimes.
1:38:43
>> Americans [laughter] can still do it.
1:38:44
But as an industry, >> Yeah, I know.
1:38:45
As restaurant leaders, we can just use the metric system. >> You just use metric.
1:38:49
But >> and again, you know that it's bad when drug dealers use the metric system.
1:38:55
>> But drug dealers use the metric system. That's [laughter] right.
1:38:58
>> What the hell are we doing here?
1:39:00
>> So, if we can't adopt the metric system as an industry, what what are we doing here? >> Yeah. Yeah. Yeah. What a mess. What a mess. >> Last question.
1:39:06
We've got a bunch of people in the chat have asked, who do you think is going to win the AI race? Hot take. >> What? [laughter] Really? Okay. We had to ask. This is a tech. We This is a tech show.
1:39:18
Just give me your gut gut answer. First reaction. One word. One word. >> Google Anthropic. Open AI. Who you got? >> Commodore computers. >> There we go.
1:39:28
Commodore [laughter] >> coming at you.
1:39:32
>> Dark horse in the race. Love it. Uh thank you, sir. >> Great, great hanging.
1:39:35
Um >> well, we'll see you guys at F1.
1:39:40
>> Yeah, we'll be very excited to see you there. Can't wait. Have a great one. We'll talk. Be good. You're the man.
1:39:45
>> Have a great rest of your day.
1:39:45
Uh, let me tell you about graphite. dev.
1:39:46
Code review for the age of AI.
1:39:48
Graphite helps teams on GitHub ship higher quality software faster.
1:39:52
We have been keeping our next guest waiting for far too long. Uh, Lauren from Figma.
1:39:57
Thank you so much from for holding. >> Look at this. Look at this background.
1:40:02
I couldn't even tell if it was I could just noticed that it was a TV, but it took >> This is how they do it.
1:40:08
A lot of the a lot of the professional TV hits on uh CNN and CNBC, people will be sitting right in front of a TV and they'll put uh some sort of fake background. It works very well.
1:40:18
>> Uh so great to have you on the show. Thanks for having me.
1:40:22
>> We we've been reacting to uh Nana Banana Pro this morning.
1:40:26
Very very impressed on a bunch of different dimensions.
1:40:28
But before we get into that, would love uh an introduction on yourself for the audience and and your background. >> Yeah, of course.
1:40:36
So I joined Figma two months ago as their chief design officer and before then I spent close to a decade at Meta primarily working on messaging.
1:40:46
So I led the messenger and Instagram DM teams and more recently leading consumer AI on the product side.
1:40:53
And um you know before you ask me I'll tell you why I joined Figma.
1:40:56
Um, I did so because in my seat watching all of the um AI improvements that that we're seeing with these frontier models, it became very very clear that product development as a process is going to change drastically.
1:41:11
And um I truly believe and I saw that Figma has the opportunity and I think the responsibility from my um point of view to really build the creative environment that helps people like me um people that really love to live at the intersection.
1:41:28
Um I used to be a musician.
1:41:28
I became a designer then a you know product leader.
1:41:33
I really believe in the thing we're making more than like how different disciplines kind of like line up to to get the product done.
1:41:39
And so I believe that a creative environment that helps you get that idea from your head into a finished product is what we need right now.
1:41:47
And I'm excited to help Figma um build this. >> Amazing.
1:41:50
So so many different ways that you can integrate AI into Figma.
1:41:53
You guys have been doing Figma make.
1:41:57
There's also like the core product.
1:41:58
What what's been your priorities kind of in the first couple months? >> Yeah.
1:42:04
So for everyone who doesn't know, Figma is the place in Figma where you could take your ideas or designs and prompt them into working software.
1:42:11
And that's really important because it takes your design and like helps you understand what it looks like, what it feels like in motion.
1:42:17
Um, and what we've been um looking into with this um is an aspect of AI that um I think gets overlooked sometimes.
1:42:28
As a creative tool, it is important for AI not to box you in.
1:42:31
Um, so you want to be able to take your design from Figma Figma design and um, generate it, but then you want to take those generations back to canvas and be able to like manipulate them as well.
1:42:44
So, and we're moving pretty fast.
1:42:46
In the last two months, I think we've shipped over 20 major features and a lot of them have to do with like putting the designer Yes.
1:42:53
designer Yes. [laughter] putting the designer in the driver's seat and enabling the designer to take these AI tools but really wield them as tools that are precise and that go in their direction versus the number one most frustrating thing with generative AI right now is you generate an asset that's like 98% amazing and then there's
1:43:14
like one tiny element and you try to like reprompt it and you try to say like could you remove that could you like try again on that you know with text yeah and then it's all changing and So just like make yeah making it making it easier to like go back go back and forth I think is like probably some of the most important work on the on the creative side. >> Do you have a personal evaluation that
1:43:32
>> Do you have a personal evaluation that you run when a new generative image uh model drops.
1:43:39
I have this uh the where's Waldo test.
1:43:42
I try and get it to generate a full Where's Waldo because that's there's a lot of detail in there.
1:43:46
It's this whole laded up image.
1:43:48
Um, do you have a favorite image that you go to as like your ground truth just to kind of get the flavor? >> I don't necessarily.
1:43:54
I have a ton of styles that I put it through the ringer with and a number of like creative tasks that I want to see if it does.
1:44:04
What's really important with these images and um hasn't happened uh necessarily predictably so far is that they take that certain first scene that they generate or the photo that you give them and then they're dependable in recreating the style and telling the second part of the story.
1:44:20
Otherwise, they're not helpful.
1:44:21
An image that doesn't tell a story is not helpful.
1:44:25
>> This is why I like actually um Nano Banana Pro uh because it's dependable.
1:44:30
Um the the way we um one of our companies um said today it's like it's it's a model that behaves.
1:44:35
You should actually watch the the video that they've put out.
1:44:38
It's hilarious and it's made in Weebi with with Gemini um 3. >> Yeah. >> Yeah. Yeah.
1:44:44
Apparently I mean uh Prince here on on X is saying Nano Banana Pro is a reasoning image model and shares a quote.
1:44:50
This enables enhanced image quality, better rendering of long text passages in many languages, improved factuality, which is something like we didn't like I was never thinking about the factualness of an image generator, but that's actually extremely important like you don't want errors.
1:45:05
>> Of course, is it realistic?
1:45:05
>> Of course, is it realistic? Is it something that that really connects like our our eyes right like will pick up on details even before you understand what's going on and you'll understand that this is an AI generated image and the truth is that people prefer um to look at things that feel human that a
1:45:24
human has put out there in the world and what's really cool with products like um Nano Banana Pro is that you're able to manipulate that and because it's your creative tool you could layer all of the different elements like as an example you know Dylan loves to post um videos with his like Figma quilt behind him. >> Yeah. Yeah. >> Yeah. Yeah.
1:45:42
>> And so I took that I then generated the quilt um directly.
1:45:46
Then I um turned it into a sweater and then I put it on, you know, one of Dylan's photos.
1:45:52
And all in all of these steps, it kept um each square of the quilt exact.
1:45:57
It did not distort Dylan's face.
1:46:00
I could do what I what was in my head.
1:46:02
versus in in other types of tools like this, this has not been possible yet.
1:46:08
>> What how how much do you care about like leveraging some of these models to help people generate new ideas?
1:46:15
Because in in my in in uh my creative process is just like creativity often times is just like taking two different kind of like random disconnected ideas and bringing them together and sometimes it just hits.
1:46:27
Uh and I feel like >> creativity is messy.
1:46:30
Yeah, >> it's like bringing a lot of like disperate things into into the canvas in one way or another and letting them inspire you and like take the taking the next step with those.
1:46:41
That's probably the biggest role of AI in the creative process right now is like how much can you explore because these tools exist.
1:46:51
Um because in the end what you're trying to create is still the thing in your head. >> Yeah. Yeah, that makes sense.
1:46:55
Uh, have there been any internal uh memes that have been floating around in within Figma?
1:47:04
Like uh I'm thinking of the Studio Giblly moment that was really big on the internet broadly, but have there been any uh like like just fun prompts?
1:47:12
I I'm seeing people use Nanabanana Pro to make RPG style maps.
1:47:14
Uh people are are using there's always like a new uh like fun prompt that kind of goes viral on the internet.
1:47:23
I'm wondering if you have any glimmers of uh what might be the fun prompt from Nano Banana Pro based on what you've seen in uh in the internal team >> chat.
1:47:33
>> I haven't seen any in the team chat outside of like just broad variations.
1:47:38
So like none of them um really came up to the repeating like pattern so far, but insane variations like you know taking things that are just sketches and like filling them in with like complete 3D and like as designers really what we love to do is explore.
1:47:51
So, we pushed this thing pretty hard. >> Yeah. Yeah.
1:47:54
I'm I'm excited to get uh deeper into it.
1:47:56
It's uh it's such a fun tool.
1:47:59
>> What's your >> What's your updated read on just like general designer sentiment around AI?
1:48:05
Because I feel like it it fluctuates from from fear to excitement and you have pockets where people are super excited and you have pockets where people are are kind of not excited about it or or calling it slop.
1:48:15
But uh what is like the most upto-date read from your view specifically with like designers >> relate to all of those points of view in some way right because if AI is just about speed and mass production of software and design like that is very anti what I'm here to put in the world.
1:48:35
>> Um but at the same time if design becomes a tool that you could actually control and it starts to inspire you as you you were saying that's [snorts] a very different thing.
1:48:43
It really widens the canvas and this is why we're so interested in all of the new models and we put them through the ringer because we want to see how in the hands of designers these become clay that they could mold.
1:48:53
And so I think it the the different opinions are just really at which point which part of it do you look do you look at the potential and what's you know kind of what's what's coming up and how it could work or do you look at exactly what it produced yesterday in which case a lot of times it is not great.
1:49:09
>> Yeah, makes a lot of sense. >> Makes a lot of sense.
1:49:10
I think David Chang was saying something about like you know good enough and you know producing just something good like we have this thing like at least at Figma we believe good enough is not good enough if [clears throat] all you know we're able to do in the future is create the same software a million times um that is just humanity losing. >> Yeah. Yeah. It's interesting.
1:49:29
I I I process those two things very very differently.
1:49:34
Um but yeah, I I I understand where you're coming from on that. Um >> well, say more.
1:49:38
I I I just I process uh just this idea of like I guess my question is like what is the mom and like what he was getting at was like what is the mom and pop restaurant that's not going to make an awards list that doesn't have the most viral turd duckan where it's >> oh reliable. Yeah. Yeah. >> Yeah. Yeah.
1:49:59
It's it's not it's not superlative.
1:50:01
It's not the the world's heaviest donut, the world's like most you know gold flakes on a steak possible. Like it's not viral. It's not the best.
1:50:10
Even just in terms of fine dining, it's not, oh, it has the 10 Michelin stars.
1:50:15
It's the best, the best, the best.
1:50:16
There's this demand for the superlative in the restaurant industry.
1:50:20
And then there's also the demand for just the cheapest, fast, casual, just get in, get out.
1:50:23
It's a complete commodity.
1:50:25
And I understand what both of those are in the design world a little bit.
1:50:30
bit. I I mean I I feel like we've we've seen design trends uh you know like from Apple and you know where we've all been like wow like that is truly like the best UI possible >> emotional connection >> the absolute top and then we've also seen just like okay like that's just like the bootstrap design library that
1:50:50
everyone uses for everything and that's like the the fast food of design and what's interesting is to think about that messy middle of design like what is the mom and pop shop for of design that's been there for decades that's reliable that's not you know it's not going viral and winning awards but it's good and you love it. I don't know it's
1:51:08
I don't know it's a hard it's a hard I I don't know enough about enough design to like draw an analogy but maybe you can I don't know.
1:51:15
Yeah, I think it's it's really use case dependent, right?
1:51:17
In some way, like um you want a you know Tuesday night restaurant that is not all the bells and whistles and you want it to just deliver in some case and maybe that's your to-do app or where you keep your tasks for development, etc.
1:51:30
Like there is no reason for design to kind of get in the way like in in um in those use cases.
1:51:37
But then there's moments even in those flows where you want to feel something.
1:51:40
You want to feel like that developer that thought about like the app had you in mind and those are really the the surprise moments, the delight moments that that make people be loyal to an app.
1:51:51
[clears throat] Y >> yeah I something something I've been thinking about is like what will be the product design equivalent of the mdash or or like when you when you read let's say somebody like publishes an essay and then you start reading it and you get to the second paragraph and you just like immediately close it because you realize like they just fully generated all the text.
1:52:09
I feel like we're gonna start to get that with software more and more where you you'll go to a website or or an app and and from afar or at least when you first land on it, it looks like cool, this looks like a nice product and then you start using it and you realize like okay, like they generated a bunch of like nice animations and it like looks okay.
1:52:28
But then the second that you actually start using it, you realize there was no real human thought put into the product.
1:52:34
like it is now you can now make a product that looks like linear >> in one prompt.
1:52:41
>> You cannot make a product that's going to feel like using linear. Uh >> exactly.
1:52:45
>> And and so that's where that's where the human element is just going to continue to be super super powerful and that and that and and taking user feedback and like having that empathy with the user and being super thoughtful and using the using the products for yourself.
1:52:58
Uh and not just uh because yeah, it's never been easier to create any type of application.
1:53:03
Uh it still feels like just as hard in many ways to create like a product that's truly magical to daily drive or rely on.
1:53:12
>> Yeah, that makes sense. >> Yeah.
1:53:13
And I think AI will play a role into that.
1:53:16
But actually to go back, I am so pissed about M dashes.
1:53:18
[laughter] >> Such a good tool.
1:53:23
And every time I write now, I use them and I'm like whatever.
1:53:28
People [laughter] are going to accuse me of of using AI.
1:53:33
>> Yeah, I think you just have to use the minus sign.
1:53:34
>> I do think >> just kind of be recognizable >> um when uh you know a website is just like kind of vi prompted vibe coded and put out there in the world and you're going to want to feel that um the developers spend more time considering that. >> Yeah. >> Yeah. >> Uh amazing.
1:53:54
Well, thank you so much for for joining.
1:53:56
Congratulations on the new role as uh as a Figma Figma DAU for going on >> a decade now.
1:54:04
I'm very very happy that you're on board.
1:54:08
>> Try out all the all the new toys. >> Yeah, we will. Thanks for stopping by.
1:54:12
>> We'll talk to you soon.
1:54:12
Have a great rest of your day. >> Bye.
1:54:15
>> Um if you want AI to handle your customer support, go to Finn.
1:54:18
ai, the number one AI agent for customer service.
1:54:22
Let's react to some of these nano banana prompts. They look fantastic.
1:54:26
Uh here's one where someone took an uh a map uh Google map screenshot and just turned it into uh an RPG style map, an a San Francisco monster map.
1:54:37
And uh it's it's really is that reasoning model like you can see the Golden Gate Bridge is there and what would be logical to have attacking the Golden Gate Bridge a giant octopus.
1:54:46
And then Alcatraz Island is there and there's this sea monster next to it and everything kind of like fits like you didn't get the dragon is up at Twin Peaks.
1:54:56
You don't get the dragon in the water.
1:54:59
You get the sea monster in the water and so all these things are like pretty logical.
1:55:02
There's of course some things that are a little bit repet like repetitive and >> like ogres in in Golden Gate. >> Yeah. Yeah.
1:55:09
It's just very very cool.
1:55:09
I think this is going to be a lot of fun.
1:55:12
Um then there's someone else with a with a benchmark here.
1:55:14
Uh, Angel says, "Nana Banana nailed the burger test.
1:55:18
Uh, it's the first model to truly do this perfectly."
1:55:22
And so the prompt is remove the ingredients, leave just the top bun and the bottom bun in the exact same place and render the rest of the image uh just with a prompt.
1:55:31
And uh previously uh this would sort of confuse models a little bit here and there um because uh it would be um uh they it would it would sort of shift the colors or shift the the the sections and kind of not not uh >> this next one the make it Lego.
1:55:53
>> The Lego prompt is crazy.
1:55:53
This could be the next like Giblly moment for sure.
1:55:55
Uh if you can just take a whole bunch of photos and pipe them through.
1:55:59
We should take some of the Take a take a picture of us and put it through Nana Banana Pro and make it Lego. >> Yeah, I did.
1:56:06
I I I tried doing this earlier. It it works pretty well.
1:56:09
Sometimes with the people it it doesn't use like the the minifigure like >> Oh, it doesn't.
1:56:14
So, this is particularly good because it's already uh because so this dog one is remarkable, but it's a cartoon character and so yeah, is it going to make us a mini fig?
1:56:22
I mean, maybe that could be worked into the prompt.
1:56:25
Do you want to take some of the iconic uh photos that we've used through the press images, through the Wall Street Journal photo?
1:56:32
Um there was the New York Times photos.
1:56:35
Let's take some of those photos that we have.
1:56:36
Um and let's put those through Nano Bonanana and ask them to uh render us as minifigures in this Lego world.
1:56:43
I want to see the Ultradome in 4K Lego.
1:56:45
Uh I uh while we're doing that, Pietro Chiron uh Chirano uh shares that Nano Banana is wild banana pro.
1:56:56
That is here's my favorite use case so far.
1:56:59
Take papers or really long articles and turn them into a detailed whiteboard photo.
1:57:04
It's basically the greatest compression algorithm in human history.
1:57:07
This is a very cool video where attention is all you need gets turned into this uh image.
1:57:14
I I can already imagine, you know, when we got Chachi PT, it was like, "Oh, wow.
1:57:18
You can take uh you can take bullet points, you can expand it into into an essay, and you can take an essay and expand it down to bullet points."
1:57:24
And I imagine that people are going to be sending these and then they're not going to be reading them, [laughter] and then they're going to be like, "Actually, like, >> okay, turn this diagram into an essay and then summarize it >> and then summarize it.
1:57:34
[laughter] Turn it into two words for me.
1:57:35
Just one just turn it into just one word."
1:57:37
Um, but it is very cool and I'm I'm excited where people will will play with this. It is.
1:57:44
It does look really good.
1:57:44
I think we're going to play with this tomorrow.
1:57:47
We we have a diagram uh a market map of our own coming tomorrow.
1:57:50
We're going to break down the state of AI uh from the TBPN perspective.
1:57:54
Um Dee shares he that he literally fed Nano Banana Pro raw graph viz of AI compute commits generated by Gemini 3 and it oneshotted rendering it with logos perfectly.
1:58:05
What in God's name is in this model? That is very very cool.
1:58:11
Um I've seen uh so that's not quite at the level of uh the elegance that I've been seeing from the Wall Street Journal's uh visualization of all the circularity in the we've seen that circular graphic from the Wall Street Journal >> but it's like 70% of the way there. >> Yeah. Yeah. Yeah. I would say it's 70%.
1:58:32
I'm I'm just excited that it actually puts it gets logos correctly because with the right direction um it can definitely do some also I imagine that sketching a little bit of the ground truth of like how you want this laid out um would probably give it a lot of like scaffolding to build off of. That would be very cool. >> Look at this. >> Okay.
1:58:50
Uh let's zoom in on this.
1:58:50
So we are us >> we're not minifigures.
1:58:53
We're just Legoed Lego people. Okay.
1:58:56
Color temperature is a little off. I'm not into it.
1:58:59
Let's move on to the next one.
1:59:01
The gong looks cool in the background. I like the Lego gong.
1:59:04
Uh, have we have we done any others?
1:59:06
Well, >> this is the only one I made so far. It's pretty slow. >> It's pretty slow.
1:59:10
>> Well, um, >> this is funny.
1:59:12
So, on this on this next one, >> yeah, Gemini 3 Pro image versus GPT.
1:59:16
>> Okay, so I didn't read the caption.
1:59:16
I just looked straight at the image.
1:59:18
I just assumed that the image on the left was an actual image and then this was the output on the right because it looks terrible, but it's actually the No, this is a real image.
1:59:28
same prompt, two different uh two different results.
1:59:33
That's pretty pretty remarkable.
1:59:33
Um yeah, V4 is going to be a big big moment.
1:59:38
I'm very excited because V3 I mean such a huge leap over uh the original Sora was it uh it was pre Sora.
1:59:46
What was what was chatbt's video model?
1:59:50
Not Dolly but pre Sora app. Was it called Sora?
1:59:55
>> Yeah, it was always called Sora.
1:59:55
>> It was always called Sora. Okay.
1:59:55
Yeah, that cuz Sora one or whatever the precursor was really kind of hallucinatory and and and and and crazy.
2:00:05
V3 got a ton of the physics down, but it still has this sort of like plasticky look that you can just clock.
2:00:09
Um, but uh whatever they did with Gemini 3 Pro image uh is really pushing the photo realism much farther. Very exciting.
2:00:20
What else is going on here? Uh, Nano Banana Pro.
2:00:22
Edit this image and face swap it with Sam Alman. Slow show thinking. Nano Banana Pro.
2:00:27
Does that is that a good face swap for Sam?
2:00:31
>> I think it's just okay. >> It's okay.
2:00:32
That one's a That one's That's That one's a five out of 10, I think. Uh, let's see.
2:00:36
Uh, someone is dropping strategy and saying, "I don't want to play with you anymore."
2:00:40
Um, wait, but this is back.
2:00:44
>> No, they're dropping Gemini.
2:00:45
>> They're dropping Gemini.
2:00:45
They're going back to the model wars are really really uh heating up constantly.
2:00:49
Uh ROA is saying that they're all in on GPT 5.
2:00:52
1 Pro because it's rolling out to all pro users.
2:00:58
It delivers clearer, more capable answers for complex work with strong gains in writing help, data science, and business apps. >> What is this? >> That's exciting.
2:01:06
>> You just made this one, but it only turned you into a Lego. Jord.
2:01:08
[laughter] >> What did it do to me? >> Look at John. >> What did it do to me? >> Look at John. >> What did it do to me?
2:01:14
I'm just a [laughter] human. It missed me entirely. What's going on here? >> You got Lego Jordy. >> Lego Jordy. And what is it?
2:01:20
It doesn't know what to do with the turbo puffer because the turbo puffer is like already Lego.
2:01:24
>> Wait, I know what you are. You're already a Lego. >> That's funny. That's funny.
2:01:27
Well, speaking of turbo puffer, uh, sign up today.
2:01:31
Serverless vector and full text search built from first principles and object storage.
2:01:34
Fast, 10x cheaper, and extremely scalable. Um, uh, so GPT 5.
2:01:39
1, have you had a chance to take it first spin, Tyler?
2:01:42
What's the latest with GPT 5. 1?
2:01:45
>> Uh well, so so the main new thing is the new model is uh it's not 5.
2:01:48
1 because that came out like what two weeks ago. It's 5. 1 CEX. >> 5. 1 Pro.
2:01:53
Well, so there's Codex, but then there's also 5.
2:01:56
1 Pro for like research tasks, I I believe. >> Okay.
2:01:59
>> But anyway, Codeex, how are we doing on the benchmarks? Oh.
2:02:02
Oh, and you have you have a take. You have a take. Give me your take.
2:02:05
>> Yeah, I mean basically uh so I I think the main graph or the main kind of benchmark that everyone is is now kind of watching Yes.
2:02:12
is uh this one from meter. >> Yeah.
2:02:14
Show us where the goalposts have been moved to most recently.
2:02:15
[laughter] Where where do we move the goalpost most recently?
2:02:20
>> We still we still need to get goalpost. >> We do need goalposts. Okay.
2:02:21
So, we moved the goalpost from uh from like you know uh just surprise me with something that's uh remarkably huge.
2:02:31
>> That's totally totally qualitative.
2:02:31
You can't measure that at all.
2:02:33
I think the reason that people are using this benchmark is because like you can't saturate it.
2:02:39
It's not like a it's you just can keep measuring.
2:02:40
It's not like MMLU like there's math questions and then at some point you just answer them all correctly.
2:02:45
So it's not like interesting. Okay.
2:02:47
Where where this is like >> it this is a benchmark that you could keep doing in in 30 years, right?
2:02:53
Because it the time just goes up and up. >> Yes.
2:02:56
>> Um so if we can pull up this graph >> um it's the time horizon one. Mhm.
2:03:00
>> Um and there's basically what you've seen uh for the past like 5 years is every 8 months the um the the the time that a model can do and this is just on coding task but it's kind of generally applicable it doubles. >> Yep.
2:03:13
>> Um >> and so I think this is kind of the main thing that we should be looking at like are models stagnating? Are they decelerating?
2:03:20
>> Um and what you see is it's basically a straight line.
2:03:22
If you put on a log it's exponential but if you put on a log uh scale it's a straight line. Yep.
2:03:26
And the new model is like perfectly basically on that line. >> Yeah.
2:03:32
>> Um so I think it's like this is just a great sign like the model >> so uh how long like what time what task duration measured in time would you would you say qualifies as AGI? >> Yeah.
2:03:49
I mean, I I don't know if it's exactly I don't know if that if that's my definition of AGI because >> I think there are a lot of tasks that um take a long time but don't really require general intelligence.
2:04:01
>> Um but I I do think if you're getting into like weeks or months that's like a big kind of project that would take a person.
2:04:07
It's like a big part of their life. >> Yes.
2:04:09
>> I think if if we get up to there and I guess you I mean you can just chart it out to see if you follow this path how long would that take.
2:04:16
Um, but I mean you you said it's like basically human lifetime.
2:04:21
Isn't that your >> that's my that I think that's my correct benchmark.
2:04:25
>> But I I think that's wrong because if you think of like build a build a company that's not your entire lifetime.
2:04:31
That's like for some people that's only like >> 30 years. >> 30 years. Okay. >> Four years maybe. Yeah. I don't know.
2:04:35
Um the the the the initiation prompt is the genesis prompt.
2:04:42
It's be fruitful and multiply.
2:04:44
Like that is the AGI initialization prompt.
2:04:47
Just just replicate, you know, that's your goal, AGI.
2:04:51
Just just go create value. Just go exist.
2:04:54
Uh and then it goes and does whatever it needs to.
2:04:56
Um that's when it's like truly like, you know, uh embodied, I suppose. I don't know.
2:05:01
Um all I do know is that you can go to profound, try profound.
2:05:06
com, get your brand mentioned in Chetchup, reach millions of consumers who are using AI to discover new products and brands.
2:05:11
I guess the question is um this this task duration thing is so odd because um like does time move slower or faster in AI world?
2:05:25
You would assume manipulating time?
2:05:29
>> You should be able to manipulate time if you're in the computer, right?
2:05:30
So >> you know what I mean, right? >> Uh >> okay.
2:05:36
So, so what I'm saying is that like is that like uh if if GPT5 can do two hours of if it can work for two hours without losing consistency and still complete long tasks.
2:05:47
Um if you get a new chip that speeds that up, you do the same amount of work in half the time.
2:05:56
Like if you just actually speed up the inference, you're you're bringing this curve down.
2:05:59
And so you have this weird counterveiling force where um like I would expect a computer to be able to do problems faster than humans, right? >> Uh yeah.
2:06:13
I mean so so at least at least over time compared to uh like how long it takes a human to do it, right?
2:06:19
It's like 30 hour project answer a question like five minutes is count words in a passage >> find fact on web.
2:06:26
Uh yeah, it's like compared to how long it takes a human. >> Interesting.
2:06:31
Okay, so they have to >> because obviously you could just like if you >> How are they going to How are they going to benchmark?
2:06:36
I I I always thought this was like come up with a prompt like do they even have a prompt that can that that theoretically could take months to do and and and it wouldn't be sleeping.
2:06:49
>> Build a co massive company that takes months. Years. >> Yeah. Years.
2:06:52
Is that where we're going to be with this meter chart in in like what six more doublings or something like that?
2:06:59
>> Yes, >> that's what they're Yeah. >> I mean, you think so?
2:07:01
That seems >> that would what it would be comparable to the the time scales. >> Yeah.
2:07:06
I just I just I wonder how they're >> how they're mapping that.
2:07:11
>> Just 10 more doublings, sir. >> Yeah.
2:07:14
I mean, it certainly does seem like like good progress.
2:07:16
And I mean everyone I I I feel it very much in the sense of like um just the the amount of work that a single prompt can kick off just feels like it's doubling for sure. >> Yeah.
2:07:28
I think this is just a good benchmark.
2:07:29
A lot of people it's getting harder and harder to find good prompts that show a model is like actually better. >> Yeah.
2:07:36
>> Um and this is like a very kind of like objective thing that there's a >> there's a you know what we expect it should be where it actually is and it actually is where we expect it should be.
2:07:44
So this is like a good model like this is we're on track.
2:07:48
>> Well, if you're looking for sales tax AGI, head over to numeral. com.
2:07:49
Let numero worry about sales tax and VAT compliance for you. >> Well said, John.
2:07:58
>> Meter says, "What is my purpose?
2:07:58
All you put new AMI models on the graph." Meter. Oh my god. Uh guys, please. I need to see sonnet 4. 5 on this.
2:08:07
So sonnet's not on there and this is update the graph.
2:08:12
There seems to be there seems to be a mistake.
2:08:13
I planned on assessing risks from automated AI R&D. That's funny. Uh they're having fun.
2:08:19
What else is going on in in in AI world?
2:08:21
These uh things are looking smoothly exponential for AI over the past several years.
2:08:27
And I continue to think this is the best default assumption until the AI R&D automation feedback loop eventually speeds everything up.
2:08:34
I we we got to have the meter folks back on the show and and and understand this a little bit further.
2:08:40
Um, I I really I really wonder how they're actually de developing.
2:08:46
>> I really want to get their take on uh protein uh the amount of protein in in fast casual concepts and and potentially get them to chart that out too. Okay.
2:08:58
And and then someone someone took this chart and put it next to the AI 2027 graph. Is this correct?
2:09:02
So there's meters data GPT5 codeex max.
2:09:03
Uh it looks exponential but not super exponential.
2:09:11
Is that what this read is? >> Uh, yes. >> Yeah.
2:09:13
So, this is still in the log graph.
2:09:14
You see the blue line is the meter. >> Okay.
2:09:16
>> And then the green is the the AI.
2:09:19
>> So, AI 2027 was was um was expecting like even more of an exponential. >> Yeah.
2:09:24
And I think that's mostly because um uh they thought agents that would help develop the next AI would come a little bit sooner. >> Okay.
2:09:30
Um, but I I think I I think Gemini 3 seems to um do very well in the kind of computer use stuff, which you should imagine uh should like greatly help out uh kind of agents. >> Yeah.
2:09:43
>> So maybe they're just maybe there's a month or two, you know, ahead.
2:09:47
>> Oh, so you think we're going back to the green dots there? You're optimistic.
2:09:51
>> You think I mean it's reasonable.
2:09:53
>> You think we might jump from one line to the other from the linear to the super linear uh or super exponential?
2:09:56
from the exponential to the super exponential. Daniel says, "Yep.
2:10:00
Uh, things are going somewhat slower than AI 2027 scenario.
2:10:04
Our timelines were longer than 2027 when we published and now they are still a bit longer still around 2030.
2:10:08
Lots of uncertainty though is what I say these days."
2:10:12
Um, meter of course is uh evaluating GBT5 uh.
2:10:16
1 Codeex Max uh triggered drastic AI acceleration or automate autonomously replicate.
2:10:22
They concluded this was unlikely. survey said unlikely.
2:10:27
Um but uh obviously big uh big growth in the capabilities.
2:10:32
Uh SweetBench, I wanted your reaction to this from Val. ai.
2:10:38
Uh different evaluation, different eval but with this company, they say Gemini 3 is number one on the independent SWEBench leaderboard.
2:10:46
So >> yes, this is their own SWEBench.
2:10:48
It's also um they did not test the the actually the newest um OpenAI model. Oh, >> it's not Codex. >> Codex Max. Okay. >> X high or whatever. Like the maxed out. >> Yeah. Yeah. >> I don't know.
2:11:00
Yeah, they named it kind of poorly, but [laughter] >> for the 10th time. >> Yeah.
2:11:05
Uh but yeah, I mean I I'm curious where where that'll end up.
2:11:08
And also um people are saying uh there's like rumors of of uh Gemini 3 Flash, the small model, and there's also rumors of Anthropic releasing a model soon.
2:11:17
And I assume it would be uh Opus 45, >> right?
2:11:22
Because that that's like they have like three tiers.
2:11:23
They have the haiku, sonnet, and opus. >> Yeah.
2:11:26
>> Um, so I'm very curious to see where all all those end up.
2:11:30
>> Um, very curious to see where uh Doug landed with his >> 100 g protein.
2:11:37
He he got >> he tried it >> the 100 gram max protein bowls from Sweet Green. Yes. >> He got three of them.
2:11:46
>> That's the core research.
2:11:46
I don't think No, no, no. He didn't do three.
2:11:50
>> But wow, look at this. It's really on there. Chicken. Chicken. Chicken. Chicken. Chicken. Chicken chicken. It's so insane.
2:11:54
So if things go badly, >> it's listed out.
2:11:57
>> If things go if things go well, we can big bulk on Nvidia. Little update.
2:11:59
I ate over a little over half and my tummy hurts. What is this? It's not good. Protein max. Is tummy hurting?
2:12:06
Protein Max is the uh is the semi analysis of >> uh Phil Arstein said yesterday, "You're telling me the CEO of Sweet Green is on TVPN that he's a Chad with slick back hair, a golden tan, and a sick leather jacket, >> and his handle is Johnny Nemo.
2:12:22
Annie's a vibes guy disregarding surveys.
2:12:24
Annie added seed oil free 106 gram protein bowl, sweet green one meat. So hilarious." >> Uh I love it.
2:12:33
Uh Doug spoke a little too soon yesterday.
2:12:36
He said, "I survived the great bare market of October 29 to November 19."
2:12:41
>> Um, >> uh, let me tell you about public.
2:12:43
com investing for those who take it seriously.
2:12:46
Multiasset investing trusted by millions.
2:12:48
Um, Nvidia emerges successful.
2:12:51
Um, and yet the market is still selling off.
2:12:54
Uh, Nvidia saw its shadow six more months of bull markets as high yield hairy. Although, who knows?
2:13:00
The market is tanking still.
2:13:00
Uh, the NASDAQ is down 2. 1% now.
2:13:02
And Bitcoin is down at $86,000. >> 5% today. >> Significant selloff.
2:13:11
>> Let's check in on the sailor himself. >> Um, >> also down 5%.
2:13:15
So, at least he's tracking the underlying asset.
2:13:19
>> Uh, Meltim says, uh, Nvidia earnings call first 60 seconds.
2:13:22
We have line of sight to half a trillion in revenue in 2026.
2:13:26
The bubble hasn't even started yet. Let's go.
2:13:30
>> Uh, Michael Bur still going incredibly hard.
2:13:33
This is the circularity chart that I was calling out as like >> and I think that nano banana could pretty much oneshot this.
2:13:41
>> I don't know if it would be as as over overlapped and nuanced and like the it's not >> you can't just say make it more overlapped.
2:13:50
>> Uh no, I I don't think you can yet.
2:13:50
I mean this is I mean we're really really close.
2:13:54
>> You got to talk to somebody that has been making graphics for media companies like this forms.
2:13:59
It seems so easy but if you especially >> it takes a lot of takes a lot of >> Yeah.
2:14:05
And especially if it's like >> it takes a lot of deep thinking and reasoning. >> Exactly.
2:14:10
No one has >> it takes so much deep thinking and reasoning.
2:14:12
A model could never do this.
2:14:15
>> No, they will be able to. They will be able to.
2:14:18
>> But Michael Bur says every company listed below has suspicious revenue recognition.
2:14:21
The actual chart with all the give and take deals would be unreadable.
2:14:25
>> The future will regard this this a picture of fraud, not a flywheel.
2:14:27
True end damage is ridiculous demand.
2:14:30
True end demand is ridiculously small.
2:14:32
Almost all customers are funded by their dealers.
2:14:37
If you can name OpenAI's auditor in one hour, you win some pride.
2:14:42
>> What do What does he mean?
2:14:42
True end demand is ridiculously small. That's just not true.
2:14:48
>> Like there are tons of companies that are paying for subscriptions for all sorts of AI products. And I I don't know.
2:14:54
I >> He's a del with a crazy Poom.
2:14:55
He he's a D cell with a zero pdoom, I guess. Um I don't know.
2:15:02
I mean, >> no, I I I do think he he has uh >> Yes.
2:15:08
If you're looking at the amount of investment happening now in comparison to the demand and you don't believe that the products will get better at all, if you don't believe that >> it just flipped so much.
2:15:18
Like there was a moment where it was like, wow.
2:15:20
Like demand for this new thing went from 0 to10 billion in just a few years. This is remarkable. Yeah.
2:15:28
>> And then people were like, let's invest a trillion dollars in that.
2:15:30
And it's like, okay, well, at that price, it's like it's actually kind of crazy. I don't know. It's a lot to deal with. >> Yeah.
2:15:36
But if you think about any industry on Earth, >> Yeah. Yeah.
2:15:40
>> Do we think every industry on Earth will be using 50 to 100 times more tokens within five years, 10 years? trap.
2:15:48
You don't even have to be that much of a >> no >> of a of a >> of a permable >> of a permable to believe that. >> Yeah.
2:15:58
>> In fact, it's like hard hard you >> Tyler's permabulling.
2:16:00
He's never lost sight at any moment. He's always been long. I love it.
2:16:05
Uh let's read this uh from a Capital.
2:16:08
But first, let me tell you about Vanta automate compliance and security with the leading AI trust management platform.
2:16:13
So A Capital says, "Of course that your that's your contention.
2:16:17
Of course, this is Do you know what movie this is from, Jordy? Top quiz. Hot shot.
2:16:22
>> Uh, >> do you know what movie that's from? >> You got two quizzes. >> This one.
2:16:27
>> Do you know what this is from? >> Uh, no. I don't. >> No. Goodwill Hunting. That's it.
2:16:31
>> That's the Goodwill Hunting image. Uh, you don't know that.
2:16:33
Do you know what Pop Quiz Hot Shot is from? >> No. >> That's from Speed.
2:16:37
It's issued to Keanu Reeves.
2:16:40
Uh, he gets on the phone with him.
2:16:42
Pop Quiz Hot Shot >> worth seeing.
2:16:44
>> Speed is definitely worth seeing.
2:16:44
It's a crazy It's a It's a great movie. It's just a thriller. They they crash a bus.
2:16:48
This is a great great great movie.
2:16:50
Anyway, >> I'm in >> back to the Goodwill Hunting image meme that I love this format. It's a very fun.
2:16:57
It's a very fun way to uh illustrate and like kind of tell a whole story.
2:17:01
Uh and so a Capital says, "Of course that is your contention.
2:17:03
You're a first year AI skeptic.
2:17:04
You just finished reading Andrew Ross Sorcin's 1929 and now you think you're reliving the roaring 20s with GPUs.
2:17:10
You will cling to that until next month when you hear Jim Chenos talk about unsustainable capex and then you will start paring that the entire AI ecosystem is about to collapse under the weight of its own spending.
2:17:23
That will last until someone posts a core CDS chart and you'll repeat that too without realizing that it was just dealers hedging credit portfolios, not some cosmic warning sign.
2:17:32
then you'll probably start lecturing people about global crossing because you heard someone say 1999 fiber bubble and it made you feel informed.
2:17:41
Meanwhile, Nvidia just printed one of the biggest sequential growth quarters the sector has ever seen and guided higher again.
2:17:49
The workloads are real, the demand is real, and the cafeex is already contractually locked.
2:17:53
None of that came from a cash crash narrative paperback or a cha chaino soundbite.
2:17:58
Uh but sure, keep borrowing other people's takes and pretending they're your own.
2:18:03
One day you might actually you might look at the actual numbers and realize this is not a bubble.
2:18:08
It is the early the early stage of the largest infrastructure buildout in decades. I love it. Very fun.
2:18:15
Uh let me tell you about Figma.
2:18:18
Think bigger, build faster.
2:18:19
Figma helps design and development teams build great products together.
2:18:23
[applause] Oh, Sunday robots is uh coming on the show.
2:18:27
So, we'll get uh we maybe we should uh >> we should play the video now.
2:18:31
>> Let's play the video now. And >> little teaser. Pull it up. >> Understand. >> It's very cool.
2:18:37
>> So, this is kind of a combination of the R2-D2 form factor with a humanoid. >> Yes. Look at that.
2:18:41
Picking up two wine glasses. Insane.
2:18:46
>> I love the way it just bounces around.
2:18:48
>> So, this is sped up presumably. >> Yes.
2:18:49
I think it's at like a I think it's at like a 10x speed.
2:18:52
>> Something about the lighting here makes it look CGI to me.
2:18:56
I know it's not, but [music] it looks CGIish. I I'm fascinated. So many questions.
2:19:02
Uh, [music] says it's in autonomous mode. Sunday has done it. >> Sunday has motion.
2:19:09
I think that the uh I think that the design here is fantastic.
2:19:15
I I will have to debate it and you have to tell me what you think, but um >> definitely beating the like creepy uncanny valley in my opinion.
2:19:22
doesn't feel like oh that thing is about to pick up a knife at least to me I'm pretty I'm pretty into this design and I think the internet was as well since it got over a million views and over 3,000 likes and Charlto Douglas over at Anthropic says this is insanely insanely impressive and I agree it is.
2:19:46
>> I'm excited to ask Tony how much they had to spend to get to this point.
2:19:49
That would be interesting >> because I think it I I'm assuming it will be quite a bit less than many of the other players that are kind of competing here.
2:19:59
>> The little telescoping pole is very cool.
2:20:01
The art >> Taylor says it's the hat non-threatening lid. I agree. Just throw a cool hat. >> The hat looks Yeah.
2:20:08
So Scott, I think the hat does look kind of dumb, but that's like kind of okay.
2:20:13
I'd rather it look dumb than scary or menacing or or or weird, you know, like >> think about how scary like some of these humanoid robots would be to a one-year-old looks kind of dumb.
2:20:25
R2-D2 looks kind of dumb, but it's still like a friendly, you know, you don't want it to be >> like the Optimus or or figure would be like traumatizing to a one-year-old. >> Yeah. Yeah, for sure.
2:20:37
Well, uh, before we move on to our next post, let me tell you about Julius AI, the AI data analyst that works for you.
2:20:44
Join millions who use Julius to connect their data and ask questions and get insights in seconds.
2:20:50
Andrew Reed says, "Every deal is a special situation if you're enthusiastic enough." Very funny. Uh, let's move on.
2:20:55
Um, SAM3 video tracking is so good. Yesterday, collect data.
2:21:01
Train custom object detector.
2:21:03
Use tracker to estimate object motion days.
2:21:05
Now track anything with a text prompt in seconds. Uh, who put out SAM 3? Is that Google as well? They are That's meta. >> That's meta.
2:21:16
>> Oh, this is segment anything. >> Yeah, segment. >> Oh, okay. Okay. Okay. Got it. Okay. Wait, why?
2:21:20
Why is it on Google research then? That's funny.
2:21:23
Oh, it's how to segment videos with segment anything. SAM 3.
2:21:27
And they just happen to be hosting this in a Google Collab notebook.
2:21:30
That makes sense because Meta does not have a a Google Collab competitor that I'm aware of. Um, interesting.
2:21:37
Well, that's very exciting. >> Very cool.
2:21:39
>> Um, >> you can track the all of our gong hits potentially for velocity and understand [laughter] velocity relative to the audio volume.
2:21:50
You can understand how the production team is doing their job to lower the levels for you >> so they don't blow your ears out.
2:21:58
>> No, but if we had a live like speed tracker. >> Yes.
2:22:01
>> Like as you're swinging it. >> Speed. Yeah. >> Very cool. >> We could do it.
2:22:04
Diet Coke tracker as well.
2:22:06
We could automate all of this.
2:22:06
Um, >> Sheil has some great coverage.
2:22:09
Grock says Elon is more fit than LeBron and would win a fight against Mike Tyson. >> Fact check. True. You're absolutely right.
2:22:18
>> I'm going to ask Grock if this is true. >> Is this true? >> Grock, is this true? >> Did somebody do that?
2:22:22
I bet somebody did that in the in the in the replies here. Is this true? Um, yes. Very funny.
2:22:27
Uh I wonder how much of this is like in the pre-prompt or just in the X data set.
2:22:36
You know, uh Elon's obviously like there's just there's just an incredible amount of Elon fans in the X ecosystem still since a lot of people that weren't Elon fans left.
2:22:45
But even the Tesla bulls don't glaze to this level usually.
2:22:49
Like I So I don't know where this would come from.
2:22:54
This must have been in the pre-prompt or something.
2:22:56
But it's a very silly >> it's a very silly like >> many people are doing this.
2:23:01
I mean you you went in Sora and you said depict me as a bodybuilder. >> Yes, that's true.
2:23:06
>> And then somebody tried to hack it to give you small legs.
2:23:09
>> They did successfully prompt engineer me.
2:23:12
>> Uh they got you they did get me.
2:23:12
Um but yes, I mean I feel like I feel like at this point like we're we're past uh we're past this level of like novelty being relevant in a purchasing decision for an LLM.
2:23:26
In fact, it might work against you.
2:23:27
Um, especially in light of Gemini 3.
2:23:30
Uh, very benchmark driven.
2:23:33
They put out the model card.
2:23:33
They there were a bunch of demos that went out.
2:23:37
There were some clear examples of next um next value coming from the model.
2:23:39
Um, sort of a buy the book launch and uh >> uh unclear how much this helps the Grock brand to have something like this leak out, but certainly funny.
2:23:53
Kevin Wheel, friend of the show, says, "Today we say hello world from OpenAI for Science.
2:24:00
We're releasing a paper across 13 examples of GPT5 accelerating scientific research across math, physics, biology, and material science.
2:24:07
In four of these examples, GPT5 helped find proofs of previously unsolved problems.
2:24:11
Uh, a lot of uh a lot of this type of uh posting has been heavily uh contentious uh in the past.
2:24:20
Um but uh they are continuing to share their work. >> Yeah.
2:24:26
And I think that this stuff will eventually be, you know, fully, you know, peer reviewed and uh and also there's just this interesting dynamic where like the other labs uh they they won't really let you get away with anything.
2:24:44
anything. like they'll fact check you so fast but uh if this is if this is seriously uh impressive like you'll probably see some congrats from other >> there's also a dynamic where if you are using chat GBT >> to accelerate your own research >> are you going to are are is everybody going to stand up and yell hey I use
2:25:05
chat GBT for this or are they going to be like my research >> you know uh who you know I'm not sure that a lot of people that are leveraging ing the tool are going to be quick to give uh OpenAI credit uh or an AI credit for for something that >> Yeah, but if you're doing it something in some sort of controlled environment, go after some specific problem. >> Um before we move on, let me tell you
2:25:25
>> Um before we move on, let me tell you about Privy.
2:25:27
Privy makes it easy to build on crypto rails, securely spin up white label wallets, sign transactions, and integrate onchain infrastructure all through one simple API.
2:25:34
Uh Burn Hobart says, "Preentant New Yorker cartoon that saw prediction markets coming more than half a century in advance." Wow. June 27.
2:25:42
If you can't see this, it's uh the arrivals at uh an airport and there's flights that are arriving from Chicago, Detroit, Philadelphia, Pittsburgh. They depart at 8:00 a. m. They arrive at 10:20 a. m.
2:25:56
And then there's odds listed there because of course you'll want to bet on when the plane lands.
2:26:02
And now you can maybe you're you're close to being able to with the with the prediction markets uh on their relentless march to take over the world.
2:26:12
Uh also before we bring in our next guest um we have to talk about uh group chats in chatbt.
2:26:18
We mentioned this earlier. It's official.
2:26:22
They're rolling out globally.
2:26:22
There was a successful pilot with early testers.
2:26:26
Group chats will now be available for all loggedin users on chatbt.
2:26:28
Free go plus and propens.
2:26:31
I didn't know there's a chatbt go plan.
2:26:33
We got to figure out what that is. >> India plan.
2:26:36
>> Okay, that's interesting.
2:26:36
And then um uh also I mean it just says it's like of course it's rolling out to uh to the everyone because uh this one doesn't set the GPUs on fire.
2:26:46
This is good oldfashioned stuff the text in the database um and uh reduce churn in your product.
2:26:55
So uh makes a ton of sense.
2:26:55
uh very uh you know, we'll have to test this out and see if it's actually uh that useful.
2:27:03
Um >> yeah, I think this I mean this this is the kind of thing that uh can uh help OpenAI build more of a moat outside of a of a brand and just general distribution mode.
2:27:15
>> Chach is turning into a social app.
2:27:15
Sam pulled it off before Zuck could make the meta AI app good enough to compete compete with chat GPT says Euchin Jinn X and Grock could have a real chance to do it too, but it's rough watching DMs and chat keep breaking.
2:27:29
And this is from all the way back in February.
2:27:32
Uh CNBC said Meta plans to release a standalone Meta AI app in an effort to compete with OpenAI's Chat GPT.
2:27:40
And Sam Alman said, "Okay, fine.
2:27:42
Maybe we'll do a social app." And he did.
2:27:44
He did Sora and now he's adding social features to uh chat core.
2:27:49
Um I am interested to see like how we I I send if I do a deep research report I'll send it around to people in the organization here at TBPN every once in a while.
2:28:01
I'm wondering uh how much it makes sense to keep the chat running in chat GPT.
2:28:06
Um since I certainly do get value out of like putting together the query, sharing that, sharing the whole theory.
2:28:12
Um, I I I don't know how much I'll be a DAOU of this in a month.
2:28:15
I'll I'll need to test it out.
2:28:17
Um, but we have our next guest in the restroom waiting room.
2:28:23
>> Tar from Stout here in the studio.
2:28:27
[music] >> How are you doing?
2:28:30
>> We're saying it correctly, right? >> We Yes.
2:28:31
I mean, first off, we have to say we love the brand because we love Day Job. Um, Legends.
2:28:35
Thank you for supporting them and excellent uh excellent taste in branding agencies of course.
2:28:43
Um but please introduce yourself and introduce the business as well.
2:28:47
>> Hi, I'm Tara Kari and uh you got it right.
2:28:49
It is uh it's actually played rugby after college and it uh it means prop in South African. >> Okay.
2:28:56
After college that means you were on the on the pro track or >> No, I was I was like just guy who wanted to drink some beers every now and then.
2:29:04
>> Okay, [laughter] let's go.
2:29:04
Let's give it up for those guys. Anyway, underrated guys. >> Thank you.
2:29:09
I'm I'm here to announce our series A led by Andrea Haritz for $29. 5 million. >> Boom.
2:29:18
>> Uh, also participating [clears throat] is Activan and Coastal Ventures. Um, >> there we go. >> There we go.
2:29:26
>> We we actually saw a preview of this uh of this brand design uh when we were hanging out with the day job folks.
2:29:30
And what I thought was interesting was the positioning of how AI comes through in the messaging to the customer.
2:29:39
So maybe let's start with like the problem, the solution, what you're actually building, and then how you message it to, you know, an audience of investors or what you're building, but then also how you message it to the actual end consumer who might not care that much about the particular technologies that you're using. >> Yeah. Yeah.
2:30:00
And I think there's a lot of slop in AI are thrown around with brands and that's why we use day job similar to you.
2:30:06
>> Um you know our customers just for context what we do is we help with accounts receivable.
2:30:11
>> Uh so if you don't know what that is we help collect what you sold >> and most of our customers are the kind of like Perk and Elmer Bishop Lifting organizations you might not know but they're flyover states kind of where I'm from which is Indiana.
2:30:22
Uh, and what we do is we use AI as a platform and we help customers collect 40% of their overdue invoices in the first 6 months of using our our tool. >> Mhm.
2:30:34
>> And it's not like a traditional software where, you know, they're promising you more seats, more people.
2:30:38
Uh, we're live in 3 days for large Fortune 100 companies, which is crazy.
2:30:43
Most people don't believe us.
2:30:46
>> Uh, but then they start seeing the results.
2:30:48
And, you know, most of these people we work with, you know, they have a 9 to5.
2:30:51
you know, they're not a startup hustler, they're not grinding, they're not, you know, working in New York on Wall Street.
2:30:57
>> Uh they want to go see their kids game.
2:30:59
And so we plug in at 5 to nine so you can punch out and go to that game.
2:31:02
And that's really the, you know, what we're about here at St.
2:31:09
>> And then on the messaging side, do you feel like your customers uh want to know details about the technologies that you're implementing? Do they care about that?
2:31:19
Well, you have everybody claiming AI.
2:31:22
Like, I'm not Matthew McConna, you know?
2:31:24
I got a bad crush on Matthew McConna. Yeah.
2:31:26
>> And I I have to compete with an AI technology versus Matthew McConnA. It's almost impossible.
2:31:31
>> Uh and so, you know, our branding reflects our customers. >> Sure.
2:31:35
>> Uh you know, think Clippy, which Day Dog did a great job with. >> Yeah. Totally.
2:31:39
>> And really helping Yeah.
2:31:39
helping them be nostalgic, but it's like what software first promised you.
2:31:42
It was going to automate things. >> Yeah.
2:31:45
>> But instead, 40 years later, >> you know, you need professional services, consultants.
2:31:48
it just doesn't do the job.
2:31:50
>> What's the best uh business model for this type of business these days?
2:31:55
Consumption based, seatbased, >> success based, >> success based, percentage based.
2:32:00
>> It's almost like you talk to the team at day job to team me up with these questions, but >> we did. We actually did.
2:32:05
>> Yeah, we actually didn't talk about business model with them.
2:32:07
>> You all I mean, you all buy software. It's so confusing. I'm not a smart man.
2:32:11
And I go on and they got [laughter] multiple spreadsheets.
2:32:14
>> You got a various version.
2:32:14
You got like a pricing guide. My head spinning.
2:32:16
We just charge a monthly fee like you would a co-orker.
2:32:21
The average accounts receivable person in the United States has paid $60,000 without benefits and it takes three to four months to hire them.
2:32:29
>> We can plug in the next day. >> Mhm.
2:32:30
>> For a fraction of the cost. >> Sure. That makes sense.
2:32:34
>> Uh what what h how does it uh like how how is the actual like product design work?
2:32:40
Is this like an agent that gets integrated into communication channels?
2:32:46
Like what what does it actually look like?
2:32:47
Yeah, I'm glad you asked.
2:32:47
Uh, so we do audio, so we'll actually place AI phone calls.
2:32:53
Uh, we'll we'll do emails, we'll, you know, even do SMS and WhatsApp in different areas of the world. >> Mhm.
2:33:00
>> So, you know, the way I always tell customers is we have two forms of communication, which is like outbound, hey, you need to pay me or inbound.
2:33:05
If you're a bigger customer and somebody calls you and you work in the finance team and you're like, hey, I got a question about invoice 234. >> Mhm. >> It's pretty hard.
2:33:17
uh you have to pull up multiple systems, you have to answer questions.
2:33:19
AI is great to live behind that IVR tree and just answer it immediately.
2:33:23
On the flip side, if we reach out to a customer and they might have like their generic invoice template that goes out, they'll have a question, hey, where do I send the check to?
2:33:32
AI can instantly reply without a human being and even sit in the flow of funds where we'll send them a payment link. >> Yeah.
2:33:39
Uh, I wanna I want to dive deeper into your what I think is a hot take uh about um basically sticking with a seatbased pricing model.
2:33:47
Uh Alex Karp was on the show and he was saying like in the future all companies will be paid on the value they deliver.
2:33:55
And I'm just wondering what the difference is if you wind up going to a company that has a thousand times as many invoices, you collect a thousand times as many payments.
2:34:08
you deliver a thousand times as much value.
2:34:10
Should you not get paid at least a little bit more?
2:34:15
>> Well, I mean, Alex is an amazing entrepreneur.
2:34:17
Uh, and they're an established brand.
2:34:19
Hopefully someday, if we keep winning each day and executing, we'll be where Palunteer is. >> Okay.
2:34:25
>> Uh, you know, right now our customers when we talk to them, >> you have companies that have been around since 97 saying they're AI now. >> Yeah.
2:34:33
>> And so, you know, we have to differentiate ourselves. Mhm.
2:34:37
>> And one of the ways we differentiate ourselves is with something very simple, very easy.
2:34:40
It's like if you went to Chipotle for the first time. >> Yeah.
2:34:44
>> You line up, you get a burrito, you're like, "Wow, this is amazing back in the day." Not anymore. >> I know.
2:34:49
>> So, [laughter] we we want to make things as simple fall off of all time. >> I mean, look at me. It's brutal, right? I love Chipotle. >> I know.
2:34:57
>> But the the the tough part is a lot of the stuff our customers are looking at isn't simple. Yeah.
2:35:02
and they're looking at evaluating multiple days of presentations.
2:35:06
They're getting grilled by salespeople. >> Yeah.
2:35:10
>> You know, we want to get in, demonstrate value, and see a really quick ROI with these customers.
2:35:14
And that's what we're helping them achieve.
2:35:16
So, great example is one of our customers, Bishop Lifting, reduced their invoices by 35% past due >> and have been able to free up that cash flow for other things.
2:35:26
M >> it could be like the holidays around bonus time, you know, and they have these people across America in locations and AR is not or receivables isn't their first job.
2:35:38
And so being able to offload that and get them a little more money in their pocket is something we try to achieve for our customers.
2:35:44
>> Yeah, that makes sense. >> That's amazing.
2:35:45
Well, I got to say the chat absolutely loves you. Say, "This guy is great.
2:35:48
This guy Midwest Sensibilities in Manhattan is extraordinarily powerful.
2:35:56
He's even drinking yerba mate. What a freaking legend.
2:35:59
>> I'm liking the sound of this sto. >> We Let's give it up. >> No, I love it.
2:36:04
I mean, I I love an idea that uh that when you hear it, it's just totally obvious.
2:36:09
It's like applying, you know, the same there's like the capital war happening in like uh customer experience right now.
2:36:14
They're using a lot of the same technology.
2:36:16
You're applying it in a very >> uh clear way uh in a different part of the org. And uh I'm bullish. So, thank you.
2:36:25
Thank you so much for joining.
2:36:26
>> Really appreciate you having us on and >> uh have fun on the show. >> Have fun out there.
2:36:29
We'll see you back for the beef.
2:36:31
>> Yeah, we'll talk to you soon. Cheers. Have a good day.
2:36:34
>> Let me tell you about adquick. com.
2:36:34
Out of home out ofome advertising made easy and measurable.
2:36:38
If you're launching a new company, growing, get on adqu.
2:36:41
com, get some billboards.
2:36:43
Uh our next guest is in the reream waiting room.
2:36:46
Let's bring them into the TVPL.
2:36:48
We have Nikita from Flexion is also a day job. Is this a day? >> No, no, no. We got This is Tony. >> Oh, Tony. Hey, sorry.
2:36:59
>> We got We got mixed around.
2:36:59
Uh Tony, so great to have you on the show.
2:37:01
Uh I'm sure your 24 hours, last 24 hours have been absolutely crazy.
2:37:06
We uh played your demo on the show earlier today and uh we're absolutely blown away.
2:37:13
It's it's really uh tremendous progress and we're we're excited to meet you.
2:37:17
So before we talk Sunday would love a intro on yourself, background and all that good stuff. >> Yeah. Yeah. Absolutely. So excited to be here.
2:37:26
So before that I was actually a PhD student at Stanford working on robotics.
2:37:33
>> So some of the works are like Aloha.
2:37:33
You saw like the two robot arms clamped to a table.
2:37:37
And it's not just about the hardware but how it learns.
2:37:42
>> Uh how can we learn from human demonstrations?
2:37:44
How can we learn through reinforcement learning and all these things?
2:37:48
>> And I think last early 2024 is when I have the realization that like you know pumping out more papers and doing more research may not be the most direct way to push robotics forward but starting a company and working on real product is.
2:38:05
So this is why I co-founded this company Sunday with Chung uh who is also a PhD student at Stanford and uh which leads to memo uh X1 and all these like uh new advances. >> Uh incredible.
2:38:17
What uh what has it taken to get to this demo that you released yesterday?
2:38:23
Because our our uh I have no idea how much money you've raised uh up until this point, but it feels like you guys have accomplished a ton uh in a pretty uh resource constrained way, at least compared to companies that you're competing with in the sort of like helpful humanoid in the home category. >> Yeah, absolutely.
2:38:42
So we're we're functioning in a very efficient way and I think as an early stage company we think about as a blessing that forces us to innovate and finding out like these solutions that are orders of magnitude more efficient than like 20% efficient and 30% efficient.
2:39:00
Um, and I think a big part of it is also about like the culture and the team and all the people we have that are like really experts and really believes in what we're doing. Um, yeah.
2:39:17
Uh, John, >> yeah, I I'd love to know some of the >> also the chat is mentioning you you forgot to mention you worked at Deep Mind Tesla and and Google.
2:39:24
So sort of a non-traditional background into robotics places. Yeah.
2:39:31
What what are the key trade-offs?
2:39:34
I mean there's a lot of focus right now on teleoperation.
2:39:35
Is it is it something just a step in the path towards full autonomy?
2:39:41
There are obviously some folks that are jumping straight to uh straight to full autonomy and they say oh we never use uh teleoperation at all.
2:39:51
other folks who say uh teleoperation is a really useful tool to pull forward some of the capability.
2:39:57
Where do you stand on the issue? >> Yeah.
2:40:00
>> Yeah. So I think teley operation is a really powerful research tool >> but it's not necessarily the best tool to get to a product >> because if you think about robotics and you know put that right next to autonomous driving right Tesla has like millions of cars collecting data for them every single day and it still took
2:40:19
almost a decade to kind of see a light at the end of the tunnel that things are starting to work very well >> in robotics if the only thing we can rely on is teleoperation to gather the amount training data it will take like decades for sure because robotics is a harder problem than self-driving. >> Yeah. >> Yeah.
2:40:36
>> So the way we think about it is that how can we use human data to train the model?
2:40:43
We have like 8 billion humans in the world.
2:40:45
Like if you use like 1% of that that's already huge. >> Yeah.
2:40:49
>> So what we designed instead is um I actually have it here is called skill transfer glove.
2:40:54
>> Uh skill capture glove. Yeah.
2:40:56
>> That is one to one to man's hat. >> Oh interesting.
2:40:58
And yes, the idea here is that if you can wear the glove and do a task, memo can also do it. >> Okay.
2:41:05
>> And that essentially decouples this whole like you need a robot to be deployed in the wild before you can gather the data to train the AI.
2:41:12
We can train AI just by having people wear a glove and collect data. >> Yes.
2:41:18
But uh I mean just to go back to the question of like capital intensivity, 1% of 8 billion people, that's 8 80 million gloves.
2:41:25
if the glove costs even 10 bucks were back in, you know, you need a billion dollars to get your data set or something like that.
2:41:34
Uh, >> you don't you don't think Tony can I'm not I'm not saying you can't do it.
2:41:37
I'm just saying like like I is there a smoother path here?
2:41:40
How many gloves have you shipped?
2:41:42
Is is is there a scale thing?
2:41:44
And then also I'd be interested to know about like transfer learning.
2:41:48
Are you having luck with simulation?
2:41:48
Are you having luck with uh there's a lot of video uh just content out there of people doing tasks.
2:41:54
Is there any signal that you can pull from just a YouTube video of someone doing the dishes or do you need to simulate something in uh Unreal Engine or use a world model like what are the other tools in the tool chest?
2:42:10
>> Yeah, I think robotics is at a point that there are so many of these ideas that we haven't converged to this like one single thing which is like pre-training and post- training for LMS.
2:42:19
>> And the way we think about it is that out of all these methods some will be better than others. Mhm.
2:42:25
>> And as a startup, we should focus on that one thing that we believe in and build the best system and stack around it.
2:42:32
And what we chose was using human data like using gloves to gather data. >> Yeah.
2:42:38
>> And uh actually for all the models that we saw, we of course pre-train uh on like internet scale data, but all the specific behaviors are learned only from the gloves that we make.
2:42:50
We don't do tally operation.
2:42:52
We don't do simulation and we don't have role models. >> Whoa. Okay.
2:42:55
Uh then then how do you see the the uh data capture from the glove scaling?
2:43:03
Like do you think that there will be 80 million people in five years using this to create more training data or do you think it's a little bit more tractable of a problem where uh at a certain point? Okay. Yeah.
2:43:16
It's been a big operation but it's more like 10,000 people that you're employing or something like that. >> Yes.
2:43:24
So I think this question is more about like for us how can these data generate value >> so that we can keep this loop going now >> right >> and it's kind of similar to the whole large language model space that we need to spend a lot of money into compute but the model itself is generating like tremendous amount of value >> and for us we don't need to solve robotics to ship a product that's a lucky part. >> Sure.
2:43:48
>> And in the homes there are lots of like it's one of the few places you can do relatively simple tasks. Yeah.
2:43:54
>> But give people a huge amount of value both emotionally and functionally. >> Yeah.
2:43:59
>> And and it's much more low stakes tasks than self-driving. Right.
2:44:01
So self-driving you said it's a harder you said that uh home robotics is a harder problem earlier if I heard that right. But at least Yeah.
2:44:10
But but but at least it's lower stakes and that if you have an air if you drop a dish it it's annoying and you want to avoid that.
2:44:16
But there's not like nobody's going to like die. >> Yeah. >> Yes.
2:44:21
It's like the newer start of the pro like of the company is to solve robotics. Yeah.
2:44:27
>> But we don't need to solve robotics before we ship a product. >> Yeah.
2:44:30
>> So yeah, talk talk about uh timelines. >> Yeah.
2:44:33
So we've been around for a year and a half and our next milestone is the beta program that will run late 2026.
2:44:42
That is when we'll put memo and like tons of them into people's home and actually see how people interact with the robot and what do people want from the robot.
2:44:53
>> Um and the general availability of memo will be either 2027 or 2028 depending on the progress we made uh through the whole beta program.
2:45:04
>> Uh talk about form factor.
2:45:08
Why not uh why not give it legs?
2:45:08
I'm sure you have a a a reason for that and I and I'm curious uh because I think I think people's immediate question is okay I can see how a wheeled system it makes a lot more sense in a lot of ways but what happens if I have stairs? Yeah, absolutely.
2:45:25
So the way uh we designed this robot is super safety at a really high priority and the way we define safety is we call it passively safe that if the robot arm and torso is fully stretched out and at that point you cut the power of the robot.
2:45:41
Can it stay stable or not? >> Interesting.
2:45:45
>> And a wheel robot is actually like one of the few ways >> fall over and crush your dog or or even your foot basically. Yeah.
2:45:52
or wreck just like smash the floor, all sorts of stuff.
2:45:57
That makes a ton of sense.
2:45:58
>> And then also, I imagine that there's you can just have more battery power, maybe dock easily, and there just aren't that many tasks that require it.
2:46:05
I feel like uh every demo is the I mean, the wine glass demo is remarkable.
2:46:11
Um holding two wine glasses is hard as a human, let alone as a robot with kind of odd fingers.
2:46:18
But uh just the tidying up use case is potentially underrated because that feels like that feels right around the corner.
2:46:26
Even if the like dealing with all the racks and and different spoons and knives and wine glasses, doing the full dishwasher feels a little bit harder.
2:46:36
But there's a willingness to pay, at least for me, just to go around the house and pick up the ball that's needs to be in the toy basket and pick up the shirt that's on the floor.
2:46:45
Like that's that's valuable. That is actually value.
2:46:47
uh if you can get the price right.
2:46:50
>> What was the uh what was the like key design inspirations?
2:46:53
What matters to you with design?
2:46:55
Somebody in the chat was asking if you were influenced by Homestar Runner or what? >> Oh, yeah. It is Homestar Runner. That's hilarious. >> Yeah.
2:47:03
So, the way we think about design is we kind of think backwards of what do we want the world to be like?
2:47:08
If the robots are ubiquitous, if you need to see it like every single day, what should it look like?
2:47:14
and we lean quite heavily towards building a robot that is friendly but also functional.
2:47:21
>> And these two things there's actually a small overlap in between them.
2:47:23
Um so when we designed the robots uh one I think detail that we um decide to do is we do not put camera into the robot's eyes.
2:47:36
The camera is actually right underneath its head. >> Yeah, I saw that. >> Yes.
2:47:40
So the reason is that like you're going to make eye contact with the robot.
2:47:44
you're going to like look at his face, but if you look at someone's face and his eyes, you see like a camera watching you.
2:47:50
It's a little bit creepy. >> Oh, interesting.
2:47:52
>> So, we kind of intentionally avoided that.
2:47:54
Um, and yeah, >> that's very interesting.
2:47:58
Yeah, the Yeah, the design I we were talking about earlier.
2:48:01
It feels like it really uh just it avoided like the uncanny valley, the creepiness.
2:48:07
Like there's a lot of risk factors when you're designing humanoid robotics right now.
2:48:10
We've seen all sorts of them.
2:48:11
uh or they can look cool sci-fi but maybe weirder in certain context.
2:48:14
I think this one came across very well.
2:48:18
>> Well, super super excited for you.
2:48:20
Thanks for coming on and breaking it down. This is really fun. Thank you so much.
2:48:22
If we'd love to be in the in the demo program, >> we got flat floors here. >> We got flat floors.
2:48:27
We have uh huge messes >> and we have a team of people that we will make wear these gloves all day long.
2:48:34
>> And we will take care of we will take care of Memo.
2:48:37
>> We will >> because he's because he's cute. >> Yes.
2:48:39
We love >> and we want to see him win. So, thank you so much.
2:48:42
Congrats on all the progress. >> Congratulations. >> Thank you, guys.
2:48:44
>> We'll talk to you soon. Goodbye.
2:48:47
>> Uh, you know what we got to do?
2:48:47
We got to get Memo a watch on get bezel. com. >> Memo 6,000. Ice it out. >> Luxury watched out.
2:48:59
>> Fully authenticated by Bezel's team of experts. >> It needs an RM.
2:49:04
>> He definitely and some chrome hearts.
2:49:05
>> Definitely needs a reshard mill. Why not? Why not?
2:49:07
Uh uh next up we got Nikita from Flexion. Excited for this one.
2:49:13
>> Thank you so much for taking time to talk to us today. Thanks for waiting. Good to meet you.
2:49:17
[screaming] How you doing? >> Hi.
2:49:19
I'm really excited to be here.
2:49:21
>> Um so >> I'm the CEO and co-founder of Flexion. >> Cool.
2:49:26
>> Uh where we're building the intelligence layer or the brain that powers all kinds of robots from humanoids to mobile manipulators. >> Yeah. I mean fantastic.
2:49:33
We were just talking about humanoid robotics.
2:49:35
uh how do you see the the market playing out?
2:49:39
>> Okay, so so so you uh hopefully you caught at least the end of our of our conversation with with Sunday robotics.
2:49:45
But something that I was thinking about like a real challenge is when Sunday gets good enough at uh picking up, you know, manipulating objects.
2:49:51
Uh what happens if Sunday walks up to our table here after the show?
2:49:55
We typically have lunch and Sunday needs to figure out what's trash and what's what like what should be taken and thrown away and what's actually should just stay there, right?
2:50:07
Because that's actually like somewhat of a like it requires some memory.
2:50:10
It's like okay, this is an item that that is uh it needs to be able to identify objects, figure out uh what what is like what is something that is worthy of just throwing away?
2:50:19
What is something that like I don't want thrown thrown away?
2:50:23
And if it gets thrown away, I'll be frustrated.
2:50:24
So, I feel like there's like a lot deeper uh uh more levels of complexity to a lot of these robotic tasks than than just like object manipulation and kind of understanding uh understanding the general environment and and really having like intelligence around the environments that it operates.
2:50:41
And I feel like that might be something that you're solving, but uh tell me if I'm wrong or correct. >> Absolutely.
2:50:48
Uh let me just say that Sunday is amazing.
2:50:50
I think their videos are really, really impressive.
2:50:51
probably the most impressive uh demos I've seen so far.
2:50:56
So, just start let's just start with that.
2:50:59
Um I don't know if you're doing it on purpose, but it's a great reference to the video we released this morning where we have a robot walking around and picking up trash and bring it to to a garbage can. >> Yeah.
2:51:11
>> And and the way we're doing that is actually splitting the problem into two parts.
2:51:14
The first one has nothing to do with robotics.
2:51:16
It's about um common sense and understanding.
2:51:22
And the for that part, we don't really need to train a a specific model ourselves because that knowledge is already contained in in large language models.
2:51:30
Think of it as GPT5 or all of these models.
2:51:33
If you take a picture of that table in front of you and you ask GPT what is garbage, what is not, those models are already really good at understanding that.
2:51:41
And once you have that, then the next part is actually the object manipulation.
2:51:45
Um, which we're also solving in a slightly different way compared to Sunday.
2:51:49
We we bet that the vast majority of data needed to train those models will come from simulation. Uh great.
2:51:57
You can have a look at the video. >> Uh yeah.
2:52:01
So, so say more or maybe even just narrate the video. >> Sure.
2:52:06
So, let me just quickly come back.
2:52:09
We're bad on on simulations.
2:52:09
We train robots using reinforcement learning not to humiliate humans but to solve specific tasks and just through trial and error.
2:52:17
So we have robots trying millions and millions of times and get tens or hundreds of years of simulated data and then they come up with very specific ways on how to walk across complex terrains but also use their whole body to manipulate objects. >> But problem Yeah. Yeah.
2:52:34
problem Yeah. Yeah. Isn't there a little bit of a problem there where uh to perfectly simulate that forest uh path uh requires incredible you know uh just like CGI just to I mean you need like Unreal Engine cranked to max on every physics calculation because
2:52:54
yes you can model it like a video game like it's all just one smooth surface but it's not actually that in reality there's tons of different blades of grass there might be slightly more friction over here on this blade grass versus that one. You have to simulate
2:53:05
You have to simulate all of that to actually recreate the real world. Is there not a gap? >> Yeah, absolutely. That's a great point.
2:53:15
Usually we call the sim to real gap. >> Yes.
2:53:17
>> And once you train in simulation, the whole challenge is to cross that sim to real gap. >> Mhm.
2:53:23
>> For example, here in this video, everything is trained in simulation.
2:53:24
And we were actually not even thinking about forests or mountains, but we were training the robot. >> Mhm.
2:53:30
So you don't need to simulate every single piece of grass or every single rock.
2:53:36
As long as you train on general enough scenarios, but somewhat encompass what is happening here. >> Mhm.
2:53:43
>> Then you can deploy and and the other thing is that we're not training our policies directly from RGB camera inputs.
2:53:50
Otherwise, you would actually need to simulate exactly how a forest looks. Mhm.
2:53:54
>> So we're doing some processing on top once again using some other models but we're actually trained on internet scale data. >> Okay.
2:54:01
>> A good example I think is if you want to train a robot to open a door. >> Yeah.
2:54:05
>> Um either you have to simulate every single possible door that exists in the world with all the textures, the lighting, etc. >> Sure.
2:54:11
>> Or you can use a model like segment anything.
2:54:14
Uh and then you paint the door in red and the handle in let's say green.
2:54:19
Then all doors kind of started to look the same. >> Look the same. Interesting.
2:54:22
And then you're b you're basically training the motion against the segment anything version of the door of the world. >> Yeah. Something like that. >> Okay. Yeah.
2:54:32
Uh uh what what technologies are are you most excited about across uh these generative world models uh these Gaussian splats uh just Unreal Engine getting better like traditional 3D workflows uh Houdini and Cinema 4D.
2:54:45
uh are are of those tools which ones will be most useful to you uh in the future or or is everything kind of bespoke in its own world for you?
2:54:57
>> All of this is super important.
2:54:57
It's all about the time frame.
2:54:59
>> So today we're using physics based simulators just like Unreal Engine.
2:55:06
>> And that's actually my take this is enough for for way more than what most people think.
2:55:11
people think. we can go a long way with with just those simulators >> and and the logical >> and and explain explain that is it is it that uh if you have a physics simulation that's running fine and let's say Unreal Engine you might use something else but
2:55:24
Unreal Engine is do you think we're on a scaling curve where if you had a million GPUs running a million instances of of uh Unreal Engine generating simulated data that that would actually result in better progress on the on the robotics side on the actual decision-making and planning side. >> Yeah, exactly. That mixed with one more >> Yeah, exactly.
2:55:43
That mixed with one more thing which is generative models that can create assets for simulation. >> Okay, got it.
2:55:50
>> That you don't need have humans coming up with a million different versions of all the things that the robot needs to interact with. >> Yeah.
2:55:57
So, previously that was uh programmatically like like to try and get to something with a varied world like that where there's, you know, a little hill over here and a and a rock out of place that the robot might trip over.
2:56:09
over. You would have to do all that programmatically uh maybe through some nodebased workflow in Houdini or just kind of uh or or just uh just inject just randomness just random number generators and then rotate this rock over here change the geometry etc etc
2:56:26
but you're saying that generative AI can can create even more variation is that the idea >> yeah exactly you actually have two ways to add more variation [clears throat] easily one is something like gshian splats where you go outside You collect real world data. >> Sure. >> Sure.
2:56:41
>> And suddenly you have a lot of assets.
2:56:43
And the the second version is you ask journi to to do it for you. >> Yeah. Yeah. That makes sense. That's cool. Uh >> any news? >> Yeah. What you got? >> Give us. >> Yeah.
2:56:53
So we announced this morning that for the first time that we raised 50 million uh just >> congratulations. >> Uh who participated? >> Thanks. >> A bunch of masters. Um DST Global. Okay.
2:57:08
Nvidia Adventures, Versipated, Proas, First Moonfire. >> Awesome.
2:57:14
And then where where are you building where are you building the company?
2:57:17
You're uh in Europe or have you moved uh over to the West Coast?
2:57:23
>> So, right now we're all in Zurich in Switzerland.
2:57:26
>> This is why you see the robot walking in our nice Alps. >> Oh, yeah.
2:57:29
>> But actually, right now, I'm I'm in San Francisco right now for a few days and I'm here to to find the right team to start a second office here. Okay, that's great. Well, good luck.
2:57:39
>> Yeah, I would, if I were you, I wouldn't. It'd be tough to leave. Uh, >> pretty nice.
2:57:44
>> Switzerland completely.
2:57:44
Second favorite country in the world for me after America.
2:57:48
So, >> uh hopefully uh next time next time I'm in uh Switzerland, I'll I'll definitely would love to stop by the office and and and meet.
2:57:58
>> Um but uh congratulations on the milestone. Super exciting.
2:57:59
And if you ever have uh hot takes on robotics, feel free to let us know. >> Hop by. Thanks. Thanks so much. >> Awesome. >> Thank you so much. >> Cheers.
2:58:11
[applause] >> If you're planning to go to the Alps, book a wander with inspiring views, hotel, great amenities, dreamy beds, top tier cleaning, and 24/7 concier service.
2:58:21
>> It's a vacation home, but better.
2:58:21
Did you see this that we don't understand how ice is why ice is slippery?
2:58:24
I don't know if this is uh fully confirmed at this point, but uh Masimo Rain Maker 1973 shares new research. So misspelled.
2:58:37
New research shows ice is slippery because of electrical charges, not pressure and friction.
2:58:41
For almost 200 years, the prevailing explanation for ice is slipperiness, was that friction or pressure from a skate, boot, or tire melted a microscopic film of water on the surface, creating a lubricating layer.
2:58:55
A new study from Sar Land University has overturned that long-standing idea.
2:58:59
Boris Power here says, who's the head of applied research at OpenAI says, uh, wow, this is one of the bigger firm beliefs I held that got overturned.
2:59:12
>> Like I I really [laughter] I like it's like this is the one thing I knew was true.
2:59:17
I knew that the world is round.
2:59:19
The sun rises in the east and it sets in the west.
2:59:22
And I know that the reason ice is slippery is when you layer of of water.
2:59:29
Uh, but it is a good point.
2:59:29
If it's actually about electrical charges, then it begs the question which he's asking.
2:59:34
I wonder how long until we get non-slip shoes for ice.
2:59:36
So, you could have a shoe that has a battery in there creates some sort of electrical field that cancels out the electrical field or something.
2:59:42
Maybe that does something. I don't know.
2:59:45
>> Ice is is brilliantly humbling.
2:59:45
You know, you think you're walking, you're confident, you know, you're like, I I'm handling this ice and then you just and then and then uh suddenly it feels like you got a banana under your foot.
2:59:56
One of my one of my friends uh was worried about getting cancelled because the first tweet he ever posted like decades ago and when when he first got on Twitter was I just slept on some ice # f ice like FU ci Ice >> and he was like [laughter] am I being like rude or uncooked?
3:00:16
Should I delete that post?
3:00:17
Uh Jackson Doll pulled out 12 lessons uh from our interview on uh dialectic >> his podcast.
3:00:27
Uh I I don't think we'd ever written down a lot of these ideas.
3:00:31
I think we've certainly talked >> uh you were talking about the need for principles and the need for uh some sort of uh you know culture.
3:00:40
>> That was more like operating principles within the company.
3:00:41
Some of these are relevant.
3:00:44
>> These are relevant but yeah this is more about the style of content. Good summary.
3:00:48
You can't copy compounding.
3:00:50
>> If you want to know more about us uh and how we think about the show behind the scenes, you can go listen to Jackson Doll's latest episode with none other than yours truly.
3:00:58
And >> on our own very set >> on our own very set. Yeah, we filmed it here. Uh the dialectic pod.
3:01:04
>> Uh Aiden says, "Just so we're clear, anti-gravity is a Windsurf rapper.
3:01:09
Windsurf is a VS Code wrapper.
3:01:09
VS Code is an Electron wrapper.
3:01:12
Electron is a Chromium wrapper.
3:01:14
Chromium is a C++ wrapper. C++ is a C wrapper.
3:01:16
C is an assembly wrapper.
3:01:19
Assembly is a machine code wrapper.
3:01:21
Machine code is a binary wrapper.
3:01:22
Binary is a physics wrapper.
3:01:25
Physics kind of a big jump there.
3:01:26
[laughter] >> Math is a logic rapper.
3:01:27
Logic is a philosophy rapper.
3:01:29
Philosophy is a humans rapper.
3:01:31
Humans are a carbon rapper.
3:01:32
Carbon is a star forge matter rapper.
3:01:35
Stars are a gravity rapper.
3:01:37
Gravity is definitely not an anti-gravity rapper.
3:01:40
>> 19k like people enjoy it. Uh this is very funny.
3:01:44
Everything's >> This is funny.
3:01:46
Uh Robin Hood had a post, trade the forecast, weather market predictions, and Augustus says, "Haha, this is how this is how uh Augustus can can one way that he can monetize is just, you know, getting a hedge fund betting on the weather outcome.
3:02:03
>> It's not going to rain.
3:02:03
I'd like to say it's not going to rain with size."
3:02:07
[laughter] >> Like, is it where is Augustus? Is he in town? Is he batting?
3:02:12
Uh he would be I don't know.
3:02:12
Would that be in uh would that be um investigative?
3:02:18
Would that be insider trading?
3:02:18
We'll have to figure it out.
3:02:20
Uh Satcha Nadella had a banger.
3:02:22
Um uh Barely AI says, "Never forget Sacha Nadella in 1993 as a Microsoft technical marketing manager showing how Excel works." We can play this clip. This is funny.
3:02:34
Um, as you can see, the most important architectural requirement for this piece is to be able to integrate data which exists on a host or a mainframe right now into Excel.
3:02:48
Excel being our front-end tool and the AS400 in our case being the data repository.
3:02:53
So, what I'm going to do now is exit out of this environment and try and show you how we can better integrate this data into Excel.
3:03:01
and I'll go ahead and >> call in questions.
3:03:05
Now, [laughter] >> he's doing a live stream >> basically.
3:03:09
I mean, it's on TV, but >> at this point, what it did was it talked to the uh MSquery went ahead and talked to the DRDA driver uh and went and connected to the mainframe, brought down the relevant data and populated my sheet here with the relevant data going to using Windows NTSN server connecting to the database.
3:03:27
It sounds like >> Agentic AI sounds like a workflow for numbers.
3:03:34
>> This guy's been automating workflows since day one.
3:03:36
Now he says he has less hair, but the same love for Excel and he's uh posted a photo making sheet happen since 1985. Uh he's looking great.
3:03:48
He's having he's on top of the world.
3:03:49
Um uh Sunno raise more money.
3:03:49
world. Um uh Sunno raise more money. we had the founder on the show uh like just a week or two before the uh the round so we didn't have him back on but uh congrats to everyone over there and uh there's uh the the the the regulatory deals are getting worked out though >> yeah so company called Clay is the first
3:04:10
music AI service to reach a deal with all three major record labels Universal Sony and Warner Music >> Clay plans to announce its agreements in the coming days I guess they kind of front running them there Klay is building a product that will offer the features of a streaming service like Spotify amplified by AI technology that will let users remake songs in different styles. I knew a founder that was
3:04:31
I knew a founder that was working on this exact service and uh ultimately thought that it would be impossible to get all these deals done.
3:04:40
So, I'm glad that somebody uh persisted and uh built this product because I think it's going to be uh pretty fun to play around with.
3:04:48
Um, Clay apparently has licensed the rights to thousands of hit songs so that it can train its LLM.
3:04:56
The company has positioned itself as a friend of the industry.
3:04:58
Kind of letting a fox into the hen house maybe.
3:05:02
[laughter] Uh, offering assurances that the artists and labels will have some control over how their work is used.
3:05:06
Clay is led by music producer Ari Addi and also employs former executives from Sony Music and Google's Deep Mind.
3:05:14
Um, and uh, anyways, so this uh, I'm excited to play around with the product when it comes out.
3:05:22
>> We should close out with eight. com.
3:05:25
Exceptional sleep without exception.
3:05:27
Fall asleep faster, sleep deeper, and wake up energized.
3:05:29
Uh, and I want you to tell me which Ferrari do you like?
3:05:35
Because we finally have the Ferrari bench results in a Ferrari in Minecraft. This one's from GPT 5.
3:05:41
1 Pro with the same prompt as if we scroll down, we can see what Gemini 3 Pro did.
3:05:48
Which one do you think is better?
3:05:50
Which do you think is is more Ferrari?
3:05:53
>> I mean, uh, the Minecraft Ferrari Gemini 3 actually looks something like a >> I think I like the Gemini 312.
3:06:02
>> GPD 51 Pro doesn't look anything like >> it got red.
3:06:06
It's missing just like there with the Gemini 3 Pro.
3:06:08
You can see uh what I like about it is you see that little yellow dot on the hood.
3:06:13
It's clearly like that's where the Ferrari logo goes on an actual Ferrari and it knew to put that in there.
3:06:22
It's just a little bit more the wing is a little more articulated and opinionated.
3:06:25
It feels like it feels like it's more disconnected from the overall structure.
3:06:32
Uh but still an interesting challenge and uh I'm very excited to see where this benchmark goes because um it is uh it's just so visual.
3:06:40
It's so tangible like okay I understand uh what this should look like and uh it it really illustrates all the hallucinations.
3:06:48
Anything else you want to close out with?
3:06:50
>> Uh I will close out by saying it's pouring rain so hard that I'm hearing it through uh through our earbuds. >> Okay.
3:06:57
>> Um so if you are in LA >> Yeah. Yeah. Be safe.
3:06:59
Be safe out there wherever you are in the world. We love you.
3:07:03
Thank you for tuning in with us today.
3:07:04
We will be back tomorrow for a Friday show. We got Sagg coming on.
3:07:11
>> It's going to be a fun one.
3:07:12
>> I'm sure he I'm sure he'll have fully 180ed on AI.
3:07:14
[laughter] >> Our biggest AI bull.
3:07:17
We got semi analysis and then breaking points.
3:07:21
>> We're trying to bring you diverse perspectives. >> We really are.
3:07:25
>> We really do care about that.
3:07:25
We don't want this to be an echo chamber.
3:07:27
We obviously have strong views ourselves on many topics, but uh we're here to bring uh to to have real conversation.
3:07:34
So, >> thanks everyone for tuning in.
3:07:38
>> Thank you for tuning in.
3:07:38
>> Thanks for dealing with the chaos in the chat. Wow.
3:07:40
[laughter] >> Yeah, very chaotic day in the chat.
3:07:41
Uh that was our first time being like raided.
3:07:46
>> Yeah, we got we got raided a lot.
3:07:48
>> It was really funny because they were they made they made like seemingly like 20 fake accounts.
3:07:53
>> Had 20 accounts >> and then uh they they were really angry at Ben. Yeah.
3:07:56
for some reason, which was funny.
3:07:59
And then and then they also were really angry at Merkore. They kept dunking on. >> Yeah. Why?
3:08:04
Well, why are they mad at Merore?
3:08:05
[laughter] It was very distracting.
3:08:06
I I had to wind up turning off chat.
3:08:08
But uh thank you to everyone who stayed the course, stuck with us for the show, and made it through uh while the chat was getting wild.
3:08:16
But we appreciate you all and we will see you tomorrow. Goodbye.