0:13
Hey, [music] hey, hey. Heat. Heat. N. Heat. Hey, heat. Hey, heat. Heat. [music] Hey, heat. Hey, heat. Heat. Heat. Heat. [music] Heat. Heat. N. Heat. Heat. [music] Heat. Heat.
Search this video
2:41
[music] >> [music] >> We came [music] to this world. to reach the stars.
3:25
[music] We came to this world to see our future.
3:35
[music] We came to this world.
3:44
[music] to shape our future.
3:52
[music] We came to feel the new Heat. Heat.
4:15
[music] [music] You let me [music] [music] feel.
5:00
You're [screaming] watching TBPN.
5:00
We are live from the TBPN Ultradum.
5:03
It's Wednesday, December 3rd, 2025.
5:05
You probably thought we were at YC Demo Day in San Francisco.
5:09
We got to go to New York City tomorrow.
5:11
We're interviewing Jim Kramer, a bunch of other folks today.
5:15
Actually, yeah, we're traveling today.
5:17
So, we couldn't uh we unfortunately couldn't be in San Francisco uh at the Palace of Party Rounds, but we still have a ton of YC Demo Day content lined up for you folks.
5:28
Uh, we got Harsh Digar coming on at 11:45.
5:31
Then we got Clad Labs, the Chad makers of Chad IDE, the company that Yeah, the company that sparked they by their own uh by their own definition, they call themselves the Brain Rot IDE.
5:45
We're getting to the bottom of that story.
5:47
And then we're talking to uh probably 10 or 20 other founders.
5:49
Going to be asking them how they're building their businesses, what they're building, what they're seeing.
5:54
It's always a fun time to check in with the good folks over at uh at YC.
5:58
And of course, we will be telling you about ramp. com. Time is money. Save both.
6:03
Easy use corporate cards, bill pay, accounting, and a whole lot more. Ayay.
6:10
>> And I will also be telling you about fall, the generative media platform for developers.
6:13
Develop and fine-tune models with serverless GPUs and ondemand clusters.
6:18
Um, so, uh, today I wrote about will AWS buy TPUs from Google at in the, uh, in the front page of the Wall Street Journal's business and finance section.
6:29
Uh, they're singing the tranium chips praises.
6:33
Amazon chips, Amazon's chips pose risk to Nvidia.
6:36
The whole week we've been talking to people. >> Is that clickbait? >> I don't know.
6:43
Well, we're going to find out. We'll see.
6:44
uh it it certainly doesn't seem you know good to have more competition in the uh in the market and Tay Kim came on the show yesterday to talk about uh how Nvidia was strong and and really was not going to face significant headwinds from the TPU threat.
7:03
Of course, Dylan Patel over at semi analysis wrote a 10,000word piece all about how the TPUv7 was pretty good and Anthropic was going to be buying some and they were also going to be leasing some and they maybe had and that sparked a lot of backlash from uh Nvidia Bulls and also folks who are really tied to AMD, they're upset about it.
7:25
Uh there's a lot of there's a lot of losers if Google winds up winning with TPU.
7:29
And so um the the losers came out to uh to fight apparently.
7:36
Um but let's read through let's just get the facts down from Amazon's uh Tranium 3 launch.
7:40
We of course had the CEO of AWS on the show yesterday and I asked him about this question.
7:48
Will Amazon be buying TPU?
7:48
I think that's an interesting question.
7:53
But first, let's see what Amazon's actually planning with their own AI.
7:57
>> He didn't no cliffhanger here.
7:57
He did not say yes or no.
8:00
He just kind of >> I think I I think you can read between the tea leaves and understand how the decision will be made even though the decision has not been made yet, but we'll go through that. So, um, Amazon.
8:11
com is the latest big tech company to muscle in on Nvidia's turf.
8:16
Give me a sound cue from the fall. >> Muscle in. How about this? >> There we go. Oh, that's right.
8:20
On Tuesday, Amazon Web Services announced the public launch of its Tranium 3 custom AI chip, which it says is four times as fast as its previous generation of artificial intelligence chips. 4x speed up.
8:34
That's actually very significant. That's great.
8:36
Uh the company said Tranium 3 produced in by AWS's Anaperna Labs, fascinating company, acquired a decade ago for around 350 billion or 350 million.
8:45
So, it's pretty small acquisition actually, 350 million in AI. You never know.
8:49
But back then you start a custom silicon company.
8:54
Uh you could barely clear nine figures on the way out the door.
8:57
Um but >> Inepernal Labs has been uh working on custom silicon for Amazon for a long time.
9:05
They actually do have a custom CPU at AWS uh to accelerate CPU based workloads.
9:10
Then for the last few years they've been working on GPUs or you know A6 for uh accelerated workloads.
9:15
Um and so this custom chip uh design business Annaperna Labs uh can reduce the cost of training and operating AI models by up to 50% compared with systems that use equivalent GPUs.
9:28
The chips are meant to pro provide a stronger backbone of computing power for software developers like Dean Lighters Lighters Dorf the co-founder and co- executive chief executive officer of the startup Decart who we had on the show and Deart uh had that is valued now at $3. 1 billion. Let's go.
9:50
So, if you don't remember, Dart came on >> and Dean uh was doing live AI video generation while he was doing the interview with us. It was really crazy.
10:02
>> Yeah, he basically Yeah, it was real time.
10:03
He looked like he was in a video game, but it was happening with uh little to no delay.
10:08
Uh really, really cool demo. >> Yeah.
10:11
Um before we move on, um let me tell you about Reream.
10:13
One live stream 30 plus destinations.
10:14
If you want to multiream, go to reream. com.
10:16
So, uh he said his company had a breakthrough enabled by a tranium 3 chip by the tranium 3 chip after trying out several other competitor chips including Nvidia's processors.
10:29
Dozens of programmers and AI researchers from his San Francisco based company had been trying for months to train a version of Daycart's flagship AI powered uh video generation application known as Lucy uh that would be able to render footage in real time without bugs or hiccups.
10:44
AWS gave Deart early access to Trinium 3 after meeting with the startup and being impressed with founders.
10:52
The company was two weeks into a marathon coding session in a rented house in Silicon Valley, which I think he took us on a tour of while he was in uh Wizard Land, an AI generated sci-fi world. It was very fun.
11:05
Uh it uh that a few of his employees were celebrating wildly behind him. Wait.
11:10
Oh, that's like a ref I think that's a reference to the actual call that I'm referring to. Weird.
11:14
Uh this is very weird reading the journal.
11:17
Oh yeah, I've experienced this.
11:19
Um the moment that I saw it worked, I saw four people just start jumping up and down said Dean.
11:26
The next question was how fast can we get it to market and start changing industries with it?
11:31
The launch of Tranium 3 is the latest broadside against Nvidia which dominates the GPU market.
11:37
A flurry of deals in recent months have caught the attention of investors in indicating that more AI firms are seeking to diversify their suppliers by buying chips and other hardware from companies other than Nvidia.
11:50
So Meta Platforms is in talk with Google to buy billions of dollars worth of advanced AI processors known as TPUs.
11:56
Uh and OpenAI has struck deals with rival uh Nvidia rival AMD as well as Broadcom.
12:02
And so, um, very exciting that Dayart got good results out of the Tranium chip. That's awesome.
12:07
Obviously, I'm sure everyone over at Amazon has been working very hard on that.
12:11
Uh, at the same time, we've heard that Enthropic maybe didn't have that great of an experience with Tranium and that's why maybe they're moving over to TPU a little bit more.
12:19
Um, but >> even though Amazon remains a major >> and so my question is, will AWS buy TPU from Google?
12:28
Uh, I asked Matt Garmin that question.
12:31
I >> you asked me that question. >> Yes.
12:34
>> I said they will be mocked. >> They would be mocked. >> They would be mocked. >> Which is ridiculous.
12:38
Uh and we'll get to why that's ridiculous.
12:41
I mean, first off, it's it's just it's it's funny to mock uh anyone for something like, you know, related to their semiconductor supply chain and what they rack in their massive data centers.
12:50
AWS is a massive business.
12:52
uh where I can say is like please uh please my my arch rival can I please get some chips for my data center to compete with your data center. >> Okay.
13:02
Well, let's actually go to what Matt Garmin, the CEO of AWS, said on TBPN yesterday because I asked him, "Will you be buying TPUs?"
13:09
And he said, "Hey, look, we're very excited about Tranium and I think it has and we think it has enormous potential and we absolutely think there's a benefit to optimizing every layer of that stack."
13:22
And so he uh you know people were joking on the timeline um you know oh there's this new tranium chip and somebody was like uh uh all five people using tranium are ecstatic you know that that there's that there's this new news uh but uh probably ballistic here says Amazon's so
13:39
bad at hype tranium is used by 500 million people through bedrock but their marketing team just can't AWS is undervalued blah blah blah and he's obviously a bull on the stock but what's interesting is that like >> it is it is deployed I met some of their GTM staff today. Let's just say you'll
13:53
Let's just say you'll have years to accumulate stock at cheap prices. >> Very funny.
13:57
Uh and so and so like yes, there there obviously is value.
14:01
Even if Tranium winds up being for a particular niche, like maybe it's for real-time video.
14:08
Like maybe that's the maybe that's what it gets really good at.
14:10
It could get really good at diffusion.
14:11
It could get really good at it.
14:12
Doesn't need to just be like your your ASA can be honed and honed and honed to real video. That's interesting.
14:20
Something that Dart is focused on is working with live streamers specifically on Twitch. >> Amazon owns Twitch.
14:28
So that that >> that makes that uh that kind of partnership uh more interesting. >> I like that.
14:34
Um and uh and so but so so obviously there is value to saying, "Hey, if you go to AWS, you can get Bedrock and some services that have been fine-tuned specifically for Tranium.
14:44
you go all the way down, you're going to get very good performance because we have a stack from top to bottom that's very efficient.
14:52
But at the same time, if you're trying to do something that's sort of like not within the training ecosystem, you might have a rough go.
14:58
You might wind up on a different chip.
15:01
Um, but he did say something.
15:01
He said, uh, we are going to support choice for our customers as well.
15:08
>> And so we'll continue to offer GPUs from Nvidia as an example and we'll have and we have a very tight partnership there.
15:14
So this idea of customer choice I think is important.
15:16
And if you go back to Jeff Bezos, he said, "We're not [snorts] competitor obsessed."
15:21
This idea that Google is their archrivval, that's not in Amazon's DNA.
15:25
Jeff Bezos said, "We're not competitor obsessed. We're customer obsessed.
15:31
We're customer obsessed."
15:31
And so if the customer says, "Look, >> it's great that you acquired Anaperna Labs for $350 million.
15:38
I'm really happy with what you've done with Tranium 3. It doesn't work for me.
15:42
I'm the customer and I want you to give me an Nvidia GPU in your server or or or in your data center or I want you to give me a TPU in your server.
15:51
Uh they might do that because that's actually in Amazon's DNA. >> Yeah.
15:55
And then the follow-up question is is there any world where Google sells TPU to Amazon >> maybe? I don't know.
16:02
uh already they are partnering like this was another partnership that came out uh that Ben Thompson actually wrote about in instate techy which you should go subscribe to.
16:13
So uh separately there was an announcement of an AWS partnership with Google Cloud.
16:18
Uh now they aren't buying TPUs but what they're doing is they're enabling customers to establish private high-speed lengths links between the two companies computing platforms in minutes instead of weeks.
16:28
And so the general here the general idea here is that Google has some amazing AI capabilities that customers are just struggling to match on AWS at this point.
16:38
And the same thing is happening on Microsoft as well because on Azure you have access to open AI models that you might not have access to on uh on AWS.
16:46
And so even though your whole infrastructure might be on AWS, you might be going back and forth to GCP constantly or you might be going back and forth to AWS all the time being like I got to go over to AS.
16:57
I got to go got to go back got to go Azure back to AWS back to Azure back to AWS.
17:01
And so Amazon finally just said like hey look we have a partnership and we're just going to create a like a a dedicated pipe that put puts these two systems together.
17:09
Um and uh and so um companies used to think about AI as a special piece of their application.
17:18
So it would be fine to bounce around to another cloud to get be the best possible results.
17:22
But if the next generation of companies, I'm sure we'll talk to some of the AI focused YC demo day companies today about this.
17:30
>> I hope there's at least one >> I hope there's at least one company that's doing something with AI, that would be a real treat.
17:33
Um but >> and if you're just tuning in, uh YC Demo Day coverage starts in 30 minutes. Yes.
17:41
So, >> um, but, uh, so it used to be fine to bounce around.
17:44
Now, the next generation companies, they're maybe making their entire infrastructure decision based on who has the best AI products.
17:51
What are you laughing [laughter] at?
17:55
>> I'm laughing because uh, I texted Simon. Yeah.
17:58
>> Uh, they have Turopuffer has a booth at AWS.
18:01
I said, "How's it going at at reinvent?"
18:03
And he says, "I'm not there.
18:05
I just make it seem like I'm there as a joke because the VCs keep going to the booth and then our growth intern is like, "Oh, Simon, I don't know.
18:11
I think I saw him over there [laughter] just continuing continuing to mog while while ARR skyrock hits."
18:19
Shout out to Will, the growth intern at Turbopuffer, holding it down at uh at reinvent. >> That's fantastic. I love it.
18:30
Um but uh so let me go back to AWS.
18:35
Uh Amazon needs to fire fight back against this and allowing high-speed interconnect between AWS and GCP solves a piece of that but will they go further?
18:43
Um back on Tuesday, October 21st, 2025, I wrote in the daily update in our newsletter at tbpn.
18:49
com um about increasing supply uh competition in the AI supply chain. Here's what I said.
18:55
I said not every link in the supply chain can be completely commoditized.
18:59
This is about Open AI trying to dual source from every part of the stack.
19:02
Y and I said Nvidia has an insane amount of power right now.
19:07
Uh they've just ramped fullyear revenue from 27 billion in 2023 to 60 billion in 2024 to 130 billion in 2025.
19:17
That's like one of the greatest revenue ramps at scale in history.
19:20
Uh, and then also they grew their net profit margin from 16% to 56%. That's insane. Insane. Yes. Goat.
19:30
That's why Jensen Wong is on Joe Rogan and I'm sure it's going to be a fantastic episode because he's got a lot to talk about.
19:39
Um, all the hyperscalers and OpenAI, but that creates problems, right?
19:44
because all the hyperscalers and OpenAI are now sort of incentivized to form a bit of an anti-NVIDIA alliance to commoditize the accelerator market and drive down those margins a bit.
19:53
So 56% net profit margins on 30 billion of revenue.
20:00
People are just sitting there and they're like there's $50 billion of profit over there.
20:04
Like that's a lot of acquisitions at Pernal Labs. >> That's our cost.
20:10
>> Yeah, that's our cost.
20:10
like you're just you're you're just eating a lot off of these plates.
20:13
And so, um, CO2, I think, has done a good job explaining the current state of the anti-vidia anti-invidia alliance.
20:19
They call it the Google complex, which is probably a little bit better.
20:24
Uh, that consists of Google, Broadcom, Celestica, Lumenum, and TTM Technologies.
20:28
This coalition stands in contrast to the OpenAI complex that consists of NVIDIA, SoftBank, Oracle, AMD, Microsoft, and Coreweave.
20:37
But you know who they left off the chart entirely? Amazon.
20:39
Amazon doesn't fit neatly into either of this.
20:43
>> CO2 just loves I think I think they just love leaving a major player off any sort of graph or chart that they make, right?
20:51
They left Google off of their their fantastic 40 AI companies.
20:53
So, I think that's just a little that's just that's just them messing around a little bit. >> But there's Yep.
21:00
I mean, I I think it's accurate.
21:01
I like if you said uh is is Amazon more aligned with OpenAI or Google?
21:06
You'd be like, what are you talking about? Neither. That's correct.
21:10
They're not in one of the complexes.
21:12
Maybe they need to be, maybe they don't.
21:14
Maybe they will uh, you know, form their own complex outside of it.
21:17
Um, but I just think it's interesting that, uh, I agree with you that it's like it's ridiculous to consider the idea of them buying TPU.
21:25
That feels so uncharacteristic.
21:28
Um, and yet they serve up plenty of competitor products within AWS and they're they they will you go back to the early days of Amazon.
21:37
You can get Amazon basics paper towels.
21:39
You can also get name brand paper towels.
21:42
And that's and that exists within the AWS stack from the databases that they have on offer.
21:50
There's a lot of >> you should rebrand Tranium to Amazon basic [laughter] >> GT GPU Amazon Basics Accelerator >> Basic Basic Acceler Basic Chips. Amazon Basics chips. It would be good.
22:01
They [laughter] really really are.
22:03
They're like actually it's like one of the greatest things ever.
22:07
It's the most incredible thing that America or that humanity has ever created.
22:10
It's extremely difficult to make.
22:12
We we taught sand a thing.
22:14
[laughter] Uh anyway, uh I I I just don't think Traium 3 is the, you know, obviously everyone at AWS is like excited about it and it's a big it's a big deal.
22:23
Um but it's just not the backbone of their business and in the long term they might just retreat to supporting choice for their customers.
22:31
And so, you know, I I keep going back to that Jeff Bezos line.
22:35
We're not competitor obsessed.
22:35
were customer obsessed and so I wouldn't be as surprised.
22:40
>> How much do you think it hurts Amazon that they don't have a dedicated podcast guy?
22:45
>> Like they don't have a Scholto, they don't have a Sam, they don't have a Satia.
22:49
>> You know how much that hurts?
22:49
Because they definitely have someone in that role.
22:51
You just don't know them.
22:52
>> That's what I'm [laughter] saying.
22:54
>> They might have they might have the title, but they're not really in the driver's seat, right? >> Yeah. They don't have a rune.
22:59
>> They don't have a rune, right? They don't have a schol.
23:02
>> Yeah, they should step it up.
23:02
They should They should definitely get someone. I'd love to see it.
23:04
Um well uh fortunately I mean the semi- analysis crew was over there taking pictures sharing photos uh in the timeline of the tranium 3 ultra server liquid cooled with a lot of hard eyes.
23:18
That's some good news from uh that's a glowing endorsement from the uh uh the semi-nalysis crew.
23:25
And look at this very purple.
23:27
I wonder if that's like intentional.
23:28
Uh I wonder if they set up the uh the purple lighting.
23:30
Uh there there's a bunch of funny things going on over at reinvent.
23:34
It's also just like it's a punishing time of the year.
23:36
I guess it's like right before the holidays or something because we've just been completely to torn.
23:42
We we obviously wanted to go to YC Demo Day.
23:45
I also wanted to go to Nurips which is going on right now, the premier AI conference.
23:48
Uh there's also Dealbook Summit, Andrew Sorcin doing like all the greatest interviews.
23:54
Uh at the same time, there's reinvent. I wanted to go to that.
23:56
Uh >> crazy interviews coming out of Dealbook.
24:00
I just saw some clips this morning.
24:00
You got Scott Bess just going hard.
24:02
You got Alex Karp going hard.
24:06
No real surprises on either of those fronts, but excited to >> uh get the update there.
24:13
>> Let me tell you about cognition.
24:13
The team behind the AI software engineer, Devin, crush your backlog with your personal AI engineering team.
24:19
Uh let's let's let's close out the tranium coverage with this Zephyr post who says Google is having this kind of success with TPUs.
24:27
What about Amazon's Tranium?
24:29
Tranium is new and underpowered just 667 T flops uh BF-16.
24:34
It has lots of HPM but the bandwidth is lower than the H100.
24:39
TPU V6E is competitive H100 not on HBM or bandwidth and Ironwood is competitive with Blackwell on flops bandwidth and HBM capacity.
24:47
I expect Ironwood to quickly gain market share as it ramps up as you can see from throughput/TCO Nvidia versus tranium.
24:54
Ruben mogs tranium 3 harder than Blackwell versus Tranium 2 on TCO training flops and reduces the gap by 5% on TCO me bandwidth.
25:04
So the gap between Nvidia and Tranium is actually increasing rather than decreasing.
25:08
By the way, this math was done before CPX was introduced.
25:10
I won't be surprised if CPX plus Ruben is cheaper than Tranium for inference.
25:13
So I I do think that there's a world where there's uh where there's something specialized like what's going on with Decart uh some sort of special model that's that that that excel that that uh that thrives in what Tranium is good at and they can further niche down. But uh but we'll see.
25:33
I mean maybe they come from behind and they just destroy TP TPU and we're all talking about Tranium next year.
25:39
Uh anyway uh let me tell you about linear meet the system for modern software development.
25:44
linear streamlines work across entire entire development cycles from roadmap to release.
25:52
>> Uh we got to say a little rest in peace.
25:54
>> Rest in peace to >> Claude.
25:56
>> San Francisco's beloved albino alligator has passed away at age 30. That's a good age.
26:03
I don't know how long alligators typically live, but I'm glad it feels like >> 30 to 50 years.
26:07
So cut a little bit short, but uh Claude was of course often supported reaching 70 years or more. >> Yes.
26:18
>> Uh anyways, RIP uh uh there was you know obviously people started speculating immediately.
26:26
Anthropic of course was the sponsor of Claude. >> Yes.
26:31
>> And uh you know people were wondering was uh was there foul play involved?
26:32
Uh, was it possible this it this >> uh poor dinosaur uh not dinosaur alligator passed the day that they uh that that it that it got announced that uh they've hired IPO lawyers.
26:49
>> Some people were speculating could is it possible Claude was sacrificed to the capital markets gods and some type of >> uh ritual?
26:57
But anyways, um he look at the look at this expression he has on his face.
27:02
Can we zoom in a little bit?
27:04
What uh what a what a cool guy and he will be remembered.
27:10
>> Yeah, Dan Primac here is talking about uh Xite.
27:13
I think we might have the CEO on the show soon.
27:15
Uh the Trump administration will invest $150 million into a lithography startup called Xlite.
27:22
Uh its first chips act award.
27:22
Chatted this morning with XLite CEO.
27:24
There's a few lithography companies now.
27:26
We've had some on the show.
27:28
Um this feels like an entirely new um it's a it's a very interesting tier of investment like $150 million from the government that feels like a series B.
27:37
Um they did raise a they did raise a series B um this past summer led by Playground Global um with Playground partner and former Intel CEO Pat Gellzinger becoming Xlight's executive chairman. >> Wow.
27:53
>> And so uh makes sense that the government's investing in Intel.
27:55
Pat Gellzinger, of course, former Intel CEO.
28:00
Um, now he's getting involved in XLite.
28:03
Marshaled 40 million of capital, went and got 150 from the government.
28:05
Uh, the the story continues.
28:08
Um, there's also another AI startup [laughter] uh that wants to remake the $800 billion chip industry.
28:15
This one's in the Wall Street Journal, founded by ex Google researchers, Recursive Intelligence, raised 35 million with backing from Sequoia to automate chip design.
28:24
Obviously, this is not lithography.
28:24
This is the design process, but still um companies are >> Dylan Patel was talking a little bit about this >> the stack. Oh, he did.
28:31
I didn't hear about that. Um very cool.
28:35
>> This is AI for for AI chip design. >> Oh, that's right. Yes. AI for AI chip design. Uh everything we need.
28:40
Uh on a quiet residential street a few blocks from Stanford University, two former Google researchers are launching a startup they hope will remake the $800 billion chip industry.
28:49
Anna Goldie and Aelia Mer Hosini are trying to build software that can automate the design of cutting edge chips.
28:58
A prospect that would allow every company to build their own chips from scratch.
29:02
Working from the top floor of a suburban home, the duo recently raised 35 million to kickstart recursive intelligence with funding from Sequoia Capital and striking the >> recursive putting it in the name.
29:13
We got to add that to the list of because there's standard capital, modern capital, standard intelligence, modern intelligence, >> raw intelligence.
29:23
>> Raw intelligence was the the low hanging fruit. Applied was another one. >> Cap intelligence.
29:28
>> And then what was the other one?
29:28
Uh there's what what's locking Groom's uh company?
29:33
>> Physical intelligence.
29:34
>> Physical intelligence. Physical capital.
29:36
So it's the matrix of like capital uh what was it?
29:39
Capital, intelligence, and intuition or something like that.
29:41
and you and you multiply them all out and you get the whole thing.
29:46
>> Eventually, we're going to run out, right?
29:47
There's it's some somewhat finite.
29:49
>> No, there will always be more names.
29:51
>> Startup named >> new words. >> So, wow.
29:53
The company 35 million uh for a valuation of 750 million.
29:58
That's very low delusion.
30:00
What 5% or something like that? Uh pretty remarkable.
30:02
Uh definitely um >> VCs were high.
30:07
Yeah, I I would I would have assumed this would be a very a very capital intensive business, but I I I suppose if it's just a software that they're developing.
30:14
Um maybe maybe they have more control here.
30:17
Uh companies such as Amazon and Google have developed custom chips for AI and data center use and Apple saved billions of dollars by insourcing chips for its devices, including the M series chips that have helped revitalize its MacBook laptops.
30:32
Such silicon options can be cheaper, more >> I had a funny moment yesterday.
30:34
We got an Amazon package and it was covered with like Alienware, >> like Alienware branding.
30:41
And I I asked uh I asked Sarah, I was like, >> "Dude, did you get something from Alienware? Like what is going on?"
30:48
And it turned out to be an ad, but they were advertising that is powered by like Intel. >> Oh, interesting.
30:55
>> Which which uh didn't make me necessarily want to immediately buy an Alienware. >> What do you do?
31:02
you you put the money straight back in your pocket because you're a taxpayer. You own Intel. >> That's true. That's true.
31:08
>> You should support Intel. No, Intel.
31:08
In Intel is undisputably great for gaming.
31:12
There's no question there.
31:12
The question is, are they going to, you know, be able to build a a fab that competes with TSMC?
31:19
Like, it's a completely different question.
31:21
>> Um, I I might go build a an Alienware Intel PC.
31:25
>> Well, we're going to for the for the the office uh sim racing rigs.
31:28
>> The sim racing Intel inside. For sure. For sure.
31:31
This is just going to turn into a a sim racing show where we watch other podcasts while sim racing and reacting to it. >> Yeah. So, >> I like it.
31:40
>> Vanta automate compliance and security AI that powers everything from evidence collection and continuous monitoring to security reviews and vendor risk.
31:48
Uh Dwarash Patel has a uh a massive essay shaking up the timeline thoughts on AI progress.
31:57
He says he's moderately bearish in the short term but explosively bullish in the long term. Very interesting.
32:04
So he says he's confused why some people have short timelines.
32:08
They say AGI is coming soon, >> but at the same time they're bullish on RLVR, which is reinforcement learning with verifiable rewards.
32:17
And so he says if we're actually close to a humanlike learner, this whole approach is doomed.
32:24
Currently, the labs are trying to bake in a bunch of skills into these models through MIDRA.
32:29
There's an entire supply chain of companies building RL environments which teach the model how to use Excel to write financial models.
32:37
For example, I think we're actually talking to an AI Excel analyst for Excel power users called Crunch at 1250 YC company.
32:45
Uh I think that these are good good ideas.
32:47
I'm actually very bullish on on uh this this uh this model.
32:50
But in the context of when does AGI arrive, when does super intelligence arrive, I understand Dwarfish's point.
32:59
Uh he says either these models will soon learn on the job in a self-directed way, making all of this pre-baking pointless, or they won't, which means AGI is not imminent.
33:08
Humans don't have to go through a special training phase where they need to rehearse every single piece of software we might ever use.
33:13
Baron made interesting points about this in a recent blog post.
33:17
When we see frontier models improving at various benchmarks, we should not uh we should think not just of increased scale and clever ML research ideas, but billions of dollars spent paying PhDs, MDs, and other experts to write questions and provide example.
33:34
>> Let's give it up for answers and reasoning targets.
33:37
>> Let's give it up for the experts.
33:37
these precise capabilities.
33:39
In a way, this is like a large-scale reprise of the expert systems era where instead of paying experts to directly program their thinking as code, they provide numerals of their reasoning and process formalized and tracked and then we distill them into models through behavioral cloning.
33:54
This has updated me slightly towards longer AI timelines since we since given we need such effort to design extremely highquality human trajectories and environments for frontier systems implies that they still lack the critical core of learning that an actual AGI must possess.
34:11
Uh this tension seems especially vivid in robotics.
34:16
In some fundamental sense robotics is an algorithms problem, not a hardware or data problem.
34:22
With very little training, a human can learn how to tea operate current hardware to do useful work.
34:27
So if we had a human-like learner, robotics would in large part be solved.
34:31
But the fact that we don't have such a learner makes it necessary to go out into thousands of different homes and factories and learn how to pick up dishes or fold laundry.
34:39
One counterargument I've heard from the takeoff within five years crew is that we have to do this clu in service of building a superhuman AI researcher and then the million the million copies of automated Ilia can go figure out how to solve robust and efficient learning from experience.
34:59
This gives the vibes of we're losing money on every sale but we'll make it up in volume.
35:03
This automated researcher is somehow going to figure out the algorithm for AGI, something humans have been banging their heads against for the better part of a century while not having the basic learning capabilities that children have.
35:16
That seems super implausible to me.
35:18
Besides, even if you even even if that's what you believe, it clearly doesn't describe how the labs are approaching RLVR.
35:25
You don't need to pre-bake the consultant's skills at crafting PowerPoint slides in order to automate Ilia.
35:32
So clearly the lab's actions hint at a world where uh at a worldview where these models will continue to fare poorly at generalizing and on the job learning thus making it necessary to build in the skills that they hope will be economically valuable beforehand.
35:47
I want to go to the section on e economic diffusion.
35:50
Um but first I'm going to tell you about privy.
35:53
Privy makes it easy to build on crypto rail securely spin up lightel wallet sign transactions and integrate onchain infrastructure all through one simple API.
36:01
So you've been asking about economic diffusion.
36:03
What is the rate that we're diffusing?
36:05
Let's see what Dwar has to say about economic diffusion.
36:11
He says that economic diffusion lag is cope for missing capabilities.
36:16
And so this is also seems informed by the Tyler Cowan take that uh AGI is here the models are good but it just takes time to adopt them.
36:24
And I I'm very sympathetic to this because when I go to the doctor's office and they hand me a piece of paper, I know that a web form is good enough.
36:36
Like the capabilities of the digital form are complete.
36:42
It's not that the form is lacking in something or it's not reliable enough.
36:46
It's not like that like oh yes, like the website goes down 20% of the time and so paper makes more sense still in this case.
36:53
It's like, no, it's just a diffusion problem.
36:55
There's just someone who runs that doctor's office is like, I like doing it the old way, right?
37:01
And that's the ep and that's the economic diffusion lag problem that I think is real in a lot of scenarios.
37:07
But >> the missing capabilities thing, I mean, just just to give a pretty concrete example, right now AI is great at generating text, right?
37:15
It's great at kind of analyzing a piece of content and then generating text based on that.
37:19
And yet we still have multiple people on the team at TBPN whose job is to like find interesting moments of the show and then create captions around that and share it to Axe and Instagram and YouTube and other platforms.
37:34
>> And and Dor too where he was trying to find the most interesting pieces of a full podcast uh with one big Gemini prompt and he was trying all the different models and couldn't get it to actually find like the most salient and viral points. >> Yeah.
37:48
So one of the the the other thing that stands out is like one of the uh seeming missing capabilities is is like uh ability to like identify humor or even something like it's almost emotional.
38:00
So Ilia and Dwar talked about this where >> I think Ilia was giving the example of uh scientists studied people who had had >> various brain injuries that limited their ability to experience emotion.
38:10
Y >> and uh when you they took out emotion, it took them it would it can take somebody 2 hours to figure out which pair of socks to choose >> and they were kind of like stunned like it's just a pair of socks like you know like you know what's going on in your day. Yeah.
38:27
>> Why do you need emotion in order to make that kind of decision?
38:29
And so >> it seems like at least in AI, a missing capability is like, okay, finding out like what >> what's an interesting moment of a podcast and >> case, right?
38:41
Is it something that makes the audience member feel feel something, right?
38:46
Is it >> I mean, there's just so much to pull through.
38:48
Like I remember during the Carpathy interview, I was watching it and Tyler was watching it and there's this moment where Cararpathy says like the coding models are are are amazing and they're magical, but what they produce is slop.
39:00
And it's like that word slop is so it's the it's like the word of the year or maybe the word of last year. Like it's a huge word.
39:06
It has a huge amount of weight coming from him.
39:11
It's >> it's crazy that rage bait beat out slop for the word of the year.
39:14
>> Slop is probably the 2024 word of the year or something like that.
39:17
But anyway, um the point was like when when when I heard that when Tyler heard that that that word Cararpathy calling it slop, everyone was like, "Whoa."
39:28
And I was like, "We should clip that."
39:30
And we looked and it had already been clipped by by a human.
39:32
Like someone on the timeline had also identified that it was like that was the crazy moment that we should be like reacting to and taking in.
39:39
And and it was >> it's crazy.
39:41
The other the other thing that's that's notable is like on one of the best one of like the top jobs that people do on or or way they make their first dollar online is just like clipping for various content creators and media companies >> and and some of the clips that they make are so sloppy.
39:58
[laughter] Like it's literally just like a random segment of the show and they're blasting it out from like 20 different accounts.
40:05
>> And the fact that we're still paying humans to do that uh still I mean it just feels notable. >> Yeah.
40:10
Well, let's read Dwaresh's take on economic diffusion lab lag being cop for missing capabilities.
40:17
Says sometimes >> copium would be a beautiful name for an AI chip.
40:21
By the way, >> it would it would you got trrenium. Maybe they needium.
40:26
[laughter] >> Uh sometimes people will say that the reason that AIs aren't more widely deployed across firms and already pro providing lots of value outside of coding is that technology takes a long time to diffuse.
40:38
Dorcash thinks this is cope.
40:40
He says people are using this cope to gloss over the fact that these models just lack the capabilities necessary for broad economic value.
40:49
Steven Bar Burns has an excellent post on this and many other points.
40:54
He says new technologies take a long time to integrate into the economy.
40:57
Well, ask yourself how long how do highly skilled, experienced and entrepreneurial immigrant humans manage to integrate into the economy immediately?
41:06
Once you've answered that question, note that a AGI will be able to do those things, too.
41:10
Uh, Doresh says, "If these models were actually like humans on a server, they'd diffuse incredibly quickly.
41:18
In fact, they'd be so much easier to integrate and onboard than a normal human employee.
41:23
They could read your entire Slack and drive in minutes and immediately distill all the skills that your other AI employees have.
41:30
Plus, the hiring market is very much like a lemons market where it's hard to tell who the good people are beforehand, and hiring someone bad is quite costly.
41:39
Uh there's there this is a dynamic that you wouldn't have to worry about when you just want to spin up another instance of a vetted AGI model.
41:47
For these reasons, I expect it's going to be much easier to diffuse AI labor into firms than it is to hire a person.
41:54
And companies hire lots of people all the time.
41:56
If the capabilities were actually at AGI level, people would be willing to spend trillions of dollars a year buying tokens. Knowledge workers. >> Think about that.
42:04
We we hire someone that like we hire an AI or or we're leveraging an AI and they've listened to every single minute of TBPN ever. Yeah.
42:14
>> And watched every clip. >> Yeah.
42:17
>> And right now you'd have to fine-tune that into the model or whatever.
42:18
You you don't just get that out of the gate. >> Yeah. Yeah.
42:22
And I'm just saying like the the we we do end up hiring a lot of people that that are like previously just listeners.
42:29
>> Uh but uh getting somebody that knows every single moment that has ever happened on the show. >> Yeah.
42:35
>> Would be super powerful.
42:35
But again, there's just like missing a missing capability set that >> doesn't allow agents to deliver a lot of value internally.
42:42
At least >> the reason that lab revenues are four orders of magnitude off right now is that models are just nowhere near as capable as human knowledge workers. Yeah. I I agree with that.
42:51
I the one thing that I don't necessarily agree with here, he says, "Well, ask yourself, uh this quote from Stephen Burns, uh how do highly skilled, experienced and entrepreneurial immigrant humans manage to integrate into the economy immediately?"
43:04
I mean, they do sort of integrate into the economy immediately, but like the the immigration flow is like a slow process.
43:10
Like it doesn't just happen immediately.
43:12
It's not just like, you know, the amount of immigration went from like zero to like, I don't know, a million people or something like it's like people move around.
43:21
There is like a there is a bit of a drag, but I understand what he's saying here and it does make sense.
43:25
Um, anyway, let me tell you about public.
43:27
com investing for those who take it seriously.
43:29
They got multi-asset investing.
43:31
They're trusted by millions.
43:34
Um, uh, The Verge is trying to get on the action, trying to attack David Sachs with a headline.
43:41
It's like, it's like so funny that the New York Times went after David Sachs and then The Verge was like, "We want to go after him, too.
43:48
We want to get some of the hate.
43:52
>> Wait, let us let us cook. Let us cook."
43:54
>> We heard everyone in tech hates, you know, this this article hate, too.
43:59
Well, while I while I don't agree with uh with this uh journalistic approach, it is a pretty funny headline. >> Yeah. Oh, yeah. It's hilarious.
44:06
The the the headline is Silicon Valley is rallying behind a guy who sucks.
44:10
It's like [laughter] what does that mean?
44:17
Uh >> just pure >> pure like qualitative like just name calling.
44:22
Uh they're just like we don't like this guy. >> Pure ad hominant.
44:24
But, uh, you know, go off if people if people if your fans like it, if that's what your your audience wants.
44:31
Uh, it's it's it's rage bait. It's going to go hard.
44:33
It already got a thousand likes on a linked article.
44:36
The Verge is not putting up a thousand likes per link.
44:38
So, this is outperformance.
44:41
And, uh, it's heavily paywalled.
44:43
You cannot learn how David Sax sucks without subscribing to that thing. They did a good job. You got to pay.
44:52
>> You want to know why he sucks?
44:52
Uh, I didn't, so I don't know why he sucks, but uh, >> that'd be really funny if Behind the Pay Wall is like, "We're just kidding."
45:00
>> We think we think the New York Times missed on this one. [laughter] >> Who knows? >> Uh, Paul Graham. Yeah.
45:07
>> On the timeline, he says, "A startup told me that one of their investors didn't like that they were selling to newly founded startups and wanted them to sell to bigger companies who have more money.
45:17
If investors tell you this, write them off as idiots.
45:18
Selling to startups is the best thing you can do.
45:24
I'm sure many of the companies we're talking with today will be selling to other companies in the batch.
45:27
A lot of people uh a lot of people like say that's bad. >> Yeah.
45:35
>> Uh they try to say like a YC is a circular economy.
45:38
uh but you have to ignore the the hundreds of you know very real businesses that have you know been been created uh through YC and uh gone on to work with every kind of company in the world. Yeah. Yeah.
45:52
It certainly it certainly seems at this point uh startups tend to be smarter, less bureaucratic, uh more representative to future trends.
45:59
Like even if there's a you know some sort of insular circular economy in the startup ecosystem, like there's a pretty immense amount of pressure to actually deliver something uh that's valuable because every dollar is precious.
46:12
You >> are these every found Yeah.
46:14
They're they're being rational.
46:15
It's not like it's not like uh I'm sure there's been small instances where companies were actually, you know, had had somewhat bad behavior, but in general, it's like if I'm going to pay for your the SAS tool or the beta that you're running, it has to be good. >> Yeah. >> Has to work.
46:32
>> Uh did you see uh did you see Stuart brand?
46:37
He says, "So, there's a $ 1.
46:37
5 billion judgment against Anthropic for including $480,000 books in training their AIS.
46:44
Five of my books are among them.
46:47
Word is there might be uh $1,500 payout per book according to my agent, Max Brockman." That's a good name.
46:54
Uh he said, "I wrote them I wrote to my agent Max the following.
47:00
If any payment comes to me, please send it back to Anthropic with my thanks for including my books in their AIS.
47:06
The judgment website offers a way to opt out of the payment, but I found it cumbersome, so I didn't.
47:12
[laughter] I'm principled, but too lazy to be highly principled. I really like this.
47:16
This is uh he's the co-founder of the Long Now Foundation, which takes no sides.
47:20
In this forum, as a private person, I do take sides occasionally.
47:24
Um, so I thought that was a funny thing.
47:26
Uh, there is secondary market fraud going on left and right.
47:28
But first, let me tell you about graphite.
47:31
dev code review for the age of AI.
47:32
Graphite helps teams on GitHub ship higher quality software faster.
47:35
Uh, yeah, reading through this, Matt Grim says, "Secondary markets are rife with fraud and bad actors, and it pains me to see these bottom feeders profiting off of Anderil's growth while fleecing retail investors through unreasonable or opaque fee structures.
47:49
In this week's episode of Nonsense, uh, Ignite VC, a fund we've never taken a meeting with or had any contact with whatsoever, uh, founded by Brian, who we've never met, is soliciting investors via public Google doc to invest in an SPV that will in turn invest in another SPV that will in turn potentially enter into a forward contract with a supposedly, though unnamed, early employee. A few problems here.
48:13
First off, so-called forward contracts are notoriously hard to settle in private companies and counterparty risk is extremely real.
48:18
What about the many complicating corner cases like acquisitions where shares don't trade or marriages, divorces or deaths where ownership of the underlying shares is complicated?
48:29
Just generally a risky structure to close that I don't think most folks actually understand.
48:32
So yeah, if you enter into a forward contract and you basically buy the right to the future value of some shares and then somebody uh gets you know again married or or divorce or passes away or bankruptcy is another situation where uh you might not be actually able to collect even if the uh even if your investment uh should have generated some return.
48:54
Matt says second, this deal memo includes basically no details about Andrew's performance, no revenue figures whatsoever, no product specifics.
49:03
I guess that's good, right?
49:03
Like if they were if they were just floating around information that they had acquired.
49:07
Uh but anyways, continuing almost as if it's soliciting investors to invest on hype and momentum and not fundamentals.
49:16
Generally, I'd advise folks to be skeptical of any deal memo lacking basic details.
49:20
Third, forward contracts are explicitly disallowed by Andreal stock plan and bylaws, which means that Anderrol will never consent to Team Ignite's SPV actually taking possession of these shares while we are privately held. zero chance.
49:32
And finally, the memo spends most of its time talking about the structure and fees, which are insane.
49:38
A double layered SPV with all legal and admin costs passed through in addition to an 8% upfront fee, 3% annual fee for 2 years, 20% carried interest, and the craziest part, an implied price per share.
49:51
That is completely insane.
49:51
In this case, the implied uh PBS is 115% higher than the most recent preferred raise from nine months ago.
49:58
flattered, I suppose, but also puts these investors in an almost absurd position by paying more than double the price per share of our most recent transaction.
50:05
As stated at the top, I don't know Brian or Team Ignite at all.
50:09
Maybe they're kind of wholesome people, and this is all a big misunderstanding, but if I were an investor looking at this quote unity quote, I'd run for the hills.
50:18
Um, and I believe the founder the founder replied and said, "Appreciate the heads up.
50:23
The document reference was an internal draft prepared for discussion with an existing LP and was not intended for public circulation.
50:31
It appears someone shared it without authorization and we're looking into how that happened.
50:36
>> But do you see what and then >> there's like seven people that share a screenshot of like a direct email they got with this exact memo. >> Okay.
50:42
And the other thing is they say not soliciting investment for any and related vehicle.
50:46
Uh Matt says really the draft was written by your founder and managing partner.
50:50
I literally watched him edit the doc in real time and [laughter] he has a screen a screenshot of like the the the founder's name uh in Google Docs like you know >> uh basically >> what a mess.
51:03
Anyways, well, so don't do this. >> Don't do it.
51:06
Instead, uh why don't you start a company and apply to Y Combinator?
51:11
Um build a actual business instead of going around hustling uh SPVS and companies that don't want to sell shares.
51:18
Um but we are moving on to >> SPVS Y Combinator coverage.
51:20
We have Harsh Tagar here in the reream waiting room.
51:25
Let's bring him into the TVPL showroom.
51:27
Harsh, thank you so much for taking the time on a busy Y Combinator demo day to come talk to us. How are you doing? >> I'm doing good. Thanks for having me. >> Fantastic.
51:36
Um, take us through, uh, how's the day going?
51:39
What is the schedule like?
51:41
And then I'd love to dig into some of the trends that you're seeing, some of the standout, uh, companies.
51:47
I'm sure we're going to be talking to a lot of them, but uh, what's the what's the run of the show today and where are we in the course of the process of, you know, >> yeah, >> graduating these companies?
51:57
So we got we got started like uh almost a couple of hours ago 10 in the morning.
52:02
Um and so the founders kind of investors all gather together.
52:06
They get into um the main room here at the YC office and then the founders start giving presentations talking about like the progress, what they've built um uh themselves, their background.
52:17
They're pretty quickfire presentations, one minute each.
52:18
Uh and then there's sort of like a break in between sort of blocks or presentations where the investors can hang out and talk to some of the founders and get to meet them and um you know obviously hopefully invest in a bunch of them.
52:30
So that's kind of we're like we're just about approaching lunch so it's kind of like that part of the day where people have like listened to a bunch of companies probably got like a sense of some of the stuff that they're interested in.
52:39
I see people right now like just hanging out doing deals.
52:40
So it's kind of like a fun vibe. It's like live.
52:44
>> It's the party rounds. We love it.
52:44
I wish we could be there.
52:46
[laughter] Uh are are there any uh like hero metrics or or stats that uh the YC team shared this year to kick off demo day?
52:54
Uh how are you how are you sharing like the shape of YC these days?
53:00
Yeah, I mean um we didn't go too statsheavy this time.
53:06
um around I think I mean at a high level it's just the the continuation of the theme we've seen this whole year which is just like the companies during the batch are just getting faster revenue growth are signing like contracts with like big companies in some cases like
53:22
even like government defense tech like the the dollar value contracts that startups can close in like the first few months of their life are just bigger than anything we've ever seen and that's all like very directly from AI so it's like it Yeah, it's a very it's very interesting kind of approach. You can
53:36
interesting kind of approach. You can sign one big contract and generate enough revenue to go on the stage at demo day and feel confident in your pitch and have something that's compelling or you can go and get sign up a bunch of startups to something uh you know smaller plans but
53:52
>> post demo day like there's companies that keep growing like you're like in the SAS world you were used to sort of just consistent month overmonth like growth and now in sort of AI world you're used to like big step function growth and it might be flat for a month but then you sign like another contract and it just like leaps leaprogs Interesting. Uh yeah, help me square
54:08
Uh yeah, help me square sort of the shape of revenue with some of the YC batch that we might talk to today because uh Paul Graham was on the timeline sort of defending this idea of selling to startups.
54:19
We were in in complete agreement with that that uh selling to startups can be so much better in a bunch of different ways.
54:25
But it does feel like we're entering an era where maybe it's AI, maybe it's just the maturity of the ecosystem.
54:31
like it's also been easier than ever to sell to the government or to sell to Fortune 500.
54:37
And so are both happening in are there specific companies that are really great at one or the other?
54:41
Is there is there any advice that you've given founders on how to decide between those two paths?
54:49
>> Yeah, it's really it really depends I think on like the type of product you're building.
54:52
So I think like the the bull case for selling to startups as your customers is like the stripe or the AWS case and like it's like you get them all early.
55:02
I mean you could put gusto rippling deal into that bucket as well.
55:05
It's like if you get the startups early and you can grow with them.
55:07
That is one of the most powerful business models you can have, right?
55:11
Like the Stripe team could go on vacation for like two years and they would just like keep growing because like the cohorts would just keep going up and to the right, right?
55:18
So I don't think they're going to do that anytime soon, but they could if they wanted.
55:23
So I think if you have a product like that where you can grow with the startups and you can get in early and they will just like those startups in the future will become your enterprise customers, that's like fantastic.
55:32
That's absolutely what you want to do.
55:34
I just think like with AI what's new is you didn't even have the option of selling to a big customer until you sold to startups and you'd build up like hey like we don't have an enterprise customer yet but we got like a thousand startups and like in aggregate we're processing like X or like we're reliable we're not going to shut down.
55:52
I think now with um AI and the fact that the incumbents can't actually build the products because the engineers that work at these bigger companies don't even believe in AI.
56:01
So like the startups in the batch are able to go to a big company and actually get them as a customer because they're the only ones that can actually deliver the product.
56:10
And I think that's just new.
56:10
So like we still give the advice.
56:12
It's very dependent on the company and the product and like will you be able to scale with startups or not.
56:16
But like in general there's just more options as a founder for how you do sales than there's ever been.
56:22
>> Let's uh let's talk about themes in the batch.
56:24
two batches ago, I've it felt like a lot of the companies were at least the ones that we talked to were were uh like various like infrastructure.
56:33
It was like infrastructure for building agents.
56:35
Last batch was really felt like much more applied.
56:40
It was like applying AI uh to very specific industries and opportunities.
56:45
I'm curious um I'm sure you're seeing both of those kind of types of companies but um looking at the list of guests that we have today uh bunch of bunch of super exciting companies but curious to know kind of like broad themes across the batch.
56:57
I mean, I think you say it right.
56:59
I think what we've seen is that like maybe a year ago, just a year ago, it was like infrastructure, infrastructure to build agents like you're saying, like laying the foundation.
57:07
Then it's like vertical agents just take off, like customer support, logistics, um like name any ver like healthcare, like all these verticals and they're just like um taking off.
57:17
And primarily what they were doing is selling these agents to the companies in those verticals to make their operations more efficient.
57:24
I think what seems to be a theme coming out of this batch, you'll notice um is like the companies are going the next step and they're not actually selling the agents to the like incumbents.
57:34
They're going like AI native full stack.
57:36
They're just actually doing the thing.
57:38
So you have like um >> Fernstone being like an AI native insurance brokerage.
57:43
Like they're just they they are insurance broker and they're just going to use AI to be the best one.
57:48
Um Saba is doing that with trust.
57:50
It's like a company that sets up trust but it's doing it with AI.
57:52
So I think that um that seems to be the new trend is going like AI native and not just selling your agents but using them to build the company doing all of the stuff.
58:02
>> Yeah, we uh yeah we've talked to a couple like law firms that have done that and also like investment banks just people who have said okay we actually need to go do the do the core thing.
58:10
I'm always reminded of Justin Khan's company because yeah, uh it feels like Atrium was like just a little bit early to that model and now everyone's working on it and it's starting to maybe work and we'll see.
58:22
Um >> yeah, I mean thing is I think if you go back do you remember it was I mean it's like a decade ago now but it was Bagy that started this whole thing with like the full stack startup.
58:29
the full stack startup. Um, >> yeah, >> he like he had this blog post and like I don't know if you guys were in San Francisco at the time, but like there was this moment where there was Door Dash which was delivering food and then you had Spoon Rocket and Sprig which were like the full stack version cuz what they did is they had these kitchens like these bands which had little
58:46
kitchens driving around San Francisco cooking the food right so I think like back in that era was like it was seen as being the most ambitious thing to be a full stack startup I you didn't just sell your software you did the whole thing >> ultimately those companies And it turned out that being a marketplace or selling software was just a better scalable business in that era. But now with AI
59:03
But now with AI like I think the promise is we're kind of going back to the full stack startup idea.
59:09
But this time like you know we're all hoping and kind of seems like these things will actually scale cuz you don't need to hire like a thousand people to do the work.
59:16
You just keep improving your agents. >> Yeah. Yeah.
59:19
>> Yeah. Yeah. Yeah, I mean the the the food example is interesting because uh it feels like Travis Kalanick is maybe dipping his toe in like oh what if I did the full stack thing he's got picnic and I think it's Otter and he has cloud kitchens so maybe he's
59:33
like I can do it but maybe at his scale maybe it's a scale thing I don't know but it is it is more complicated financially >> I think if you if you're if you have Travis's like access to capital and his like background like operating like you can you can you can do that. >> Yeah. How are how are companies or >> Yeah.
59:45
How are how are companies or founders grappling with uh what's happening at the largest foundation model labs?
59:54
Like I remember there was some uh there was some Sam Alman interview where he said, you know, uh here's how not to get steamrolled.
1:00:00
Uh if the models if your entire business is just predicated on the model not getting better, you're going to have a bad time.
1:00:11
But if you're doing something completely separate with the model, um, you're you're probably good.
1:00:14
Uh, how are people thinking about it in the more modern context?
1:00:20
>> I I I think the framework people have on this stuff is that they expect, you know, >> SAM and the big lab companies.
1:00:26
I mean, Open Eye in particular to go after probably like maybe more of like the sexy consumer ideas that like capture the public's imagination.
1:00:35
Um, and it is going to be hard to compete with them on that.
1:00:38
But there are like the startups in the batch in particular that focus on just like the unsexy verticals like building an audit firm, building a legal firm, building insurance broker.
1:00:47
Like the bet they're making is that like the best people at OpenAI or anthropic are not going to be thrilled to build like auditing software or auditing agents, you know, and so like >> or actually sell the or actually sell the end service, right? >> Yeah. Exactly.
1:01:03
like doing it like going like all the way and like learning what that customer wants and how to do it really well and like iterating on it a thousand times together.
1:01:10
>> This is the whole thing with Google versus Amazon.
1:01:12
Like Google did wind up building a shopping product, but they never really had that in them to be like we're going and doing warehouses [laughter] and we're going to compete with Amazon even though we want e-commerce like we don't really want it that badly.
1:01:24
[laughter] That sounds actually sort of miserable and it's just >> don't want to do it right.
1:01:28
Like the best engineers at Google don't want to build a shopping product.
1:01:31
They like back in the day they wanted to work on search quality.
1:01:33
Now they probably want to work on Gemini.
1:01:35
But like you just >> Yeah.
1:01:37
And there's and there's also just cultural I feel like culturally there are certain companies where like if you're like we do 80% gross margin work and you show up and you're like I'm the guy who does 30% gross margin work.
1:01:48
They're [laughter] like you can leave the company actually like we don't like you at all.
1:01:52
[laughter] But so yeah, you know your margin is my opportunity both directions sometimes. >> Yeah.
1:01:58
What uh what are what are some companies from previous batches that you really feel like are hitting their stride now?
1:02:03
Uh we had Kalian on yesterday >> for their 11 billion round.
1:02:08
I I don't think a lot of people are even aware that they went through YC because it was so so long ago, right?
1:02:15
>> Yeah, there was that was uh 2019 I think.
1:02:18
So yeah, I mean obviously couch couch is like the prime example of a company that just made a bet on a space early and like had to just wait for the market to actually exist for it and like those founders like super senacious went for it.
1:02:30
for it. I think like more recently there's a company that announced around um doing customer support called Giga um which I think is like really exciting one like they're competing with Sierra and Decacaorn like superstar founders of those companies tons of capital raise but they've been able to like beat them on head-to-heads with customers like
1:02:47
Door Dash um >> through like technology really so I think like Gigo seems to be really growing um >> I mean another one like nonai that's wor like Postthog is actually like a little bit more under the radar but um >> they sort of like taking the rippling approach of >> Yeah, they're launching a new product like every week it feels like. >> Yeah, it's like really interesting to
1:03:06
>> Yeah, it's like really interesting to see like they've done that from day one and it seems to actually be compounding and working in the way that it has for Rippling.
1:03:12
So, I'm curious to see if you you start seeing more of that just like startups trying to build multiple products from day one um and have like the compound startup effect.
1:03:21
>> I I like animal themed uh companies. I like Post Hog. I like the hog themed.
1:03:24
uh when we did our first demo day stream uh we we talked to a company called Pig and we really like Pig and it stuck with me and so I'm rooting all the swine themed startups.
1:03:37
I hope they all do [laughter] very well but uh but thank you so much for uh taking the time to kick this off with us.
1:03:44
Uh congratulations on the big day.
1:03:47
>> Great great to have you on for the first time.
1:03:48
Wanted >> wanted and we got to we got to do this more often.
1:03:51
A lot of time I would love to. Thanks for having me. >> Yeah, let's talk. >> Have a good one. See you guys. >> Goodbye.
1:03:57
Uh we our first guest will be Clad Labs, makers of the Chad IDE.
1:04:03
First, let me tell you about Julius AI, the AI data analyst that works for you.
1:04:08
Join millions who use Julius to connect their data, ask questions, and get insights in seconds.
1:04:12
Um we have chat clad labs.
1:04:16
And >> while we wait, I have some other names for uh if you're launching a startup and you want a pig themed name, swine themed name, you could have Wilbur, Babe, Hamlet, >> Daisy, Peanut, and Cookie.
1:04:30
>> Okay, [laughter] >> I like that. >> Ham Solo. >> I like Babe. I think Babe. Babe, >> mud pie.
1:04:36
>> Okay, so we have the founder of Clad Labs in the Reream waiting room.
1:04:40
Let's bring him into the TVP Ultradome. >> What's going on? >> Look at the shirt. You look fantastic. >> Incredible. Incredible.
1:04:49
>> You know, you're you're winning me over already.
1:04:50
Uh, break it down for us. Introduce yourself.
1:04:52
Tell us what you're building. Good to meet you. >> How's it going, guys? Yeah.
1:04:55
Uh, I'm Richard, the CEO of Clad Labs.
1:04:57
We're building Chad IDE, the world's first brain rot IDE. >> Okay. Why? >> So, so great.
1:05:03
[laughter] So, so, so, uh, we, we exchanged some comments and wanted you to come on the show.
1:05:08
I think you get the TVPN award for the best rage bait of uh at the product level of the year.
1:05:16
Uh and I thought your response to the essay that I did was amazing.
1:05:18
You were like cool essay.
1:05:20
Unfortunately, [laughter] it doesn't unfortunately it doesn't apply to us. >> Yeah. So why doesn't it apply?
1:05:26
What are you actually building? Like why brain rot?
1:05:29
Is it just for fun or is there something meaningful here?
1:05:31
Do you think this turns into a real business? Like what's the plan? >> Yeah.
1:05:36
The general thesis is that we're able to subsidize the generation of code with affiliates >> and provide these state-of-the-art models for much much cheaper mostly for free actually to most developers. >> Mhm.
1:05:47
And and so that's why you're putting you're you're you're putting so you're acting as a funnel to you know any affiliate that so it could just be ads but you picked specifically the most controversial ones the gambling and the and the and the subway surfers like the stuff that feels more brainy.
1:06:03
Um because that would get a a reaction. Was that the plan? >> Yeah. >> Yeah.
1:06:11
I mean there's a I mean I think Jord touched on this earlier.
1:06:13
There is a difference between the marketing and the product. Sure.
1:06:17
>> We actually started out with um affiliates on these very normal sites and then a lot of our users actually requested saying hey we actually like school on rain bet.
1:06:23
We actually go to stake during our generation time like okay we'll integrate that feature and then we'll use that as our marketing campaign. >> Okay it's incredible.
1:06:30
I mean, you know, the debate was, are you making something people want?
1:06:34
Is this in keeping with the Y Combinator thesis and the values of the of the organization? Um, my >> Yeah, I guess.
1:06:41
So, so break down what's actually happening like like you have the you have the IDE and then you have this other column which you can basically fill with anything.
1:06:50
You could fill with an ad, you could fill it with videos or rainbed or whatever.
1:06:54
What are some of the most common ways that developers are using the product today and what do you think really scales and becomes um the most popular?
1:07:06
>> Yeah, the greatest thing about AI native is that it completely changes the ad unit, right?
1:07:10
So, we have these AI native ads that are in context and it's really great for code generation.
1:07:15
Here, let me give you an example.
1:07:16
So, I say I code a website. Code me a website.
1:07:18
Right now, cloud code has this multi-stage planning, right?
1:07:21
It says, well, what what do you want to code?
1:07:22
Like, how do you want to use a backend?
1:07:24
If I say well maybe I want to use a like soup base say yes superbase that's a soupbased conversion right there.
1:07:30
So the ad is actually in the context in the application layer.
1:07:33
So we have multiple ad placements but I think the most exciting one is how does ads scale at AI native.
1:07:41
>> Yeah we we had a what was the name of the company that we had on?
1:07:43
There's another company that's doing this and and actually integrating the ad so that you see an ad, you're like, "Yes, I want this functionality."
1:07:50
You press a button and the AI actually implements the uh product for you and then you're just and and and I can just see that converting at a at a really high level and companies being willing to pay >> quite a lot to get in front of people like at the right time.
1:08:05
>> I mean, yeah, it makes a ton of sense to me uh on on that level.
1:08:07
uh a little bit less on the stake gambling while you're waiting.
1:08:14
Uh that feels like that would actually reduce developer productivity.
1:08:17
Do you have any plans to actually assess whether or not this is a good decision?
1:08:22
Because most developers are not solo entrepreneurs.
1:08:25
They're employed by someone.
1:08:26
And if I'm running an organization, do I really want my most valuable, you know, resource, my most valuable human capital, uh, tuning out every other second while they're waiting for, uh, you know, the generation to come.
1:08:40
>> One of those engineers might say, "Well, I would because I'm betting on Rainbet with my personal dollars.
1:08:44
You're paying less for the IDE, I'm saving the company money." Yes. John. >> Yes. But yeah. [laughter] Yes.
1:08:50
But but but but do do you think that it would be better to show educational videos then something like that?
1:09:00
>> Oh, we have that as well. Yeah.
1:09:00
So, we have uh educational videos, learn about the code that you're actually writing. Okay.
1:09:05
>> But I think our thesis is basically that we follow the YC advice, talk to your user, and the user wanted the the gambling integration.
1:09:11
So, we made it for them and as at some point the user doesn't want it anymore, we'll take it away. Right?
1:09:16
So it's all about I think for B toc is being close to the user iterating close with the user and >> okay >> serving what they want.
1:09:22
>> Have you been banned at any companies uh yet?
1:09:25
[laughter] >> Actually the opposite we had quite a few companies reach out to us and say hey we actually really want you to integrate our notion our Jira board like the whole like productivity workflow into the generation time and we're like we really we're serving consumer right now but I mean there's a there's infinite possibilities here to scale at like various business levels. Okay.
1:09:44
>> Uh, how's the traction been to date? >> It's been great.
1:09:47
Yeah, I think I have to thank you guys for that as well, helping us go viral.
1:09:50
So, we have a great weight listed.
1:09:54
>> 11,000 people on the wait list. >> Successfully baited.
1:09:57
>> Has anyone has anyone used it yet? Have you built it?
1:10:00
Is it Is it Is it in the wild?
1:10:01
Is it Is it a BS code fork?
1:10:01
Is that >> What were kind What were the metrics that you shared at at demo day?
1:10:08
>> Yeah, so the metrics I shared at demo day were 11K on the wait list, 30k in revenue from ads.
1:10:12
We have people using right now in beta and we're gonna give out codes today at demo day.
1:10:15
So, anybody who comes up to us in demo day, we're giving you a code. Okay.
1:10:18
Um, it's going really great. >> Amazing. Find Tyler.
1:10:21
I know he's probably in the same room.
1:10:23
Let's get Let's get Tyler uh on uh on Clad Labs or Chad IDE.
1:10:29
Uh we should hit the gong. >> Yeah, we should.
1:10:33
>> Uh and uh >> uh how's the round going? >> It's going great.
1:10:38
Yeah, we filled half the round was have a lot of allocation to um to give out to people who are interested. >> Awesome. All right.
1:10:44
Well, great to meet you, Richard.
1:10:46
Uh thanks for coming on and breaking it down. >> Appreciate it. We'll talk to you soon. Have a good one.
1:10:52
>> Uh let me tell you about Figma.
1:10:52
Think bigger, build faster.
1:10:54
Figma helps design development teams build great products together.
1:10:57
Uh you can get started for free.
1:11:00
We have uh our next guest coming into the uh into the uh the Ultra Dome.
1:11:06
This next company is Absurd. >> Really? Oh, wait. It is absurd.
1:11:09
That's the name of the company.
1:11:11
Uh, they will be joining in just a minute here.
1:11:12
We might need to pop back to the timeline while we wait for them to seat sit down.
1:11:19
Jordan, you can take a you can take a a view here.
1:11:22
This is a live view into the into YC.
1:11:24
So, uh, if we if we jump ahead of the schedule, we can always, uh, check in there.
1:11:30
But, uh, we have the founder of Absurd in the reream waiting room.
1:11:35
Let's bring him into the TV. How are you doing?
1:11:38
>> Thank you so much for taking the time to talk to us.
1:11:40
>> Of course, I'm doing good. How are you guys doing?
1:11:42
I know you guys are only taking on a couple companies today.
1:11:43
So, uh, thanks for having me on. >> We appreciate you. >> Fantastic.
1:11:48
Coming up, >> uh, please introduce yourself.
1:11:49
Tell us what you're building. >> Yeah. Uh, my name is Philip. I'm CEO of Absurd.
1:11:54
Absurd makes AI marketing videos.
1:11:56
Um, an ad that we've made, you've probably seen on your feed, is Kali's Mandani versus Cuomo 1 v1 basketball match, which we did right before the elections.
1:12:05
We like to joke that we influence the New York City elections. >> Amazing.
1:12:10
So, what uh walk me through the product?
1:12:13
Uh, it sounds like you're more using the foundation models, using Sora, V3, then training your own.
1:12:18
Um, but what what what are you building?
1:12:21
How do you fit into the stack?
1:12:23
Are you more of like a creative agency that I hire and pay a lot of money for an ad and you go out and use all the tools or are you trying to build software as a service or train a foundation model?
1:12:33
Where do you sit in the stack?
1:12:37
>> The way we're seeing how we fit into the stack is that we handle everything for a company in terms of AI native distribution.
1:12:43
And the reason why we're doing it in that route instead of like making an editor that anyone could use is because we can charge exponentially higher for that.
1:12:54
So do you want to ultimately productize this?
1:12:58
>> This is what um uh this is what Har was talking about, right?
1:13:00
Basically, instead of building like an AI native accounting firm or an AI native law firm, you're effectively building uh an AI native creative ad agency where somebody comes and say, I want >> one launch video, please.
1:13:13
And you say, sure, here's the fixed price.
1:13:15
And then you guys use your internal tooling and whatever models you have access to to generate the best possible output and you deliver that end product. >> Exactly.
1:13:28
>> Um, and what are you what are you charging on on like a per video basis today?
1:13:33
>> Uh, so a lot of that's confidential, but I can say we charge upwards of 30 grand per video. Oh.
1:13:37
So, so in the same world, >> you're effectively charging the same somewhat similar to like what somebody would pay for like a full day shoot. >> Totally.
1:13:47
>> Um, >> you're in like the proper video production realm.
1:13:50
Uh, at least in terms of price.
1:13:52
What uh what are the secrets to using the uh video models appropriately to actually go viral?
1:13:57
Uh what what do you hire for? What are you focused on?
1:14:02
um making sure that the video that you deliver is actually hitting, you know, upwards of $30,000 of value.
1:14:12
>> So, in terms of the value we deliver, every video we've posted has gone viral.
1:14:16
I mean, we average 300,000 organic views for every company we work with, regardless of whether you have 200 followers on Twitter or you have like a million.
1:14:24
M >> um second thing in terms of what we're prioritizing um what we're really thinking about internally is just how many videos per person per week like what's that throughput looking like and then how do you drastically increase that week over week.
1:14:38
So 3 weeks ago that was one video per person per week. >> Mhm.
1:14:43
>> Today it's 10 Super Bowl quality ads per person per week made in parallel.
1:14:48
>> Next week it's going to be 50.
1:14:48
Following week it's going to be a thousand.
1:14:50
I mean there was a company that came to us.
1:14:52
I can't say their name, but um they said they won 1500 of our Koshi Super Bowl ads in a month.
1:14:58
[laughter] >> And that's the type of quantity that we're talking about here.
1:15:03
>> Like this is this is a lot of money that I we turned down $200,000 in the past 3 days because we just, you know, in terms of our bottleneck, we just had this huge technical bottleneck and we couldn't get it out in time.
1:15:16
>> Like >> you turned that you turned that revenue down a few days ago.
1:15:18
Why don't you just go back and say, "Hey, we have the capability.
1:15:21
We have the capacity now."
1:15:23
You just said you said it's >> ramp.
1:15:24
We still don't have capacity now.
1:15:26
We are We could literally >> So, how many $30,000 videos have you sold?
1:15:30
Did you Did you create Did you find an infinite money glitch here or something?
1:15:34
There's not even a thousand There's not even a thousand, you know, ventureback startup launches uh you know, a week. >> Yeah.
1:15:42
>> So, the the way the way we're seeing things right now, sure, we start out with launch videos that we charge 30 40 grand for, but now we're going towards more of like a retainer, right?
1:15:50
So now we're striking deals with companies like Khi, Replet, and and we're telling them, you know, we'll do a bundle deal, 10 videos a month for X price, >> right?
1:16:01
And eventually it's going to go to 50, then 100, then 200.
1:16:03
A lot of this is going to be used in ads.
1:16:07
>> Um because the more you spend on ads, the more you have to switch out ad copy because of ad, >> you know, fatigue.
1:16:13
>> And then we're going to go up and we're going to actually connect the orchestration layer to the actual metrics dashboard of all these ads.
1:16:17
Um, and then eventually we're going to get to this point where like we have this huge compounding data mode and our ad just get better and better.
1:16:26
And you can think of an ad, I think for the first time in history as like you can create a thousand different variations with one click of a button.
1:16:34
Cuz if you think about the ad in an AI ad, it's literally just like images and you're animating them.
1:16:40
And as long as you have an agent that edits the images and changes the prompt slightly, you can create a thousand different variations and then test multiple things at once.
1:16:47
Uh, will we see any absurd commercials during the Super Bowl this year?
1:16:55
>> I I can't say I I I I uh [laughter] >> You think maybe? >> That's a good answer. That's a good answer. That's a good answer.
1:17:03
answer. That's a good answer. He can't he can't uh he can't you know people are going to look up his customers and >> what do you think about uh the role of of taste of craft a lot of what's what's previously gone viral in the age uh in
1:17:19
the pre-ai age has been someone coming up with a really unique concept a really unique spin and and AI hasn't really been able to deliver those unique ideas it's really good at reconstituting what's already out there and coming up with, you know, existing ideas. >> Yeah. Historically, the best creative >> Yeah.
1:17:38
Historically, the best creative agencies have been the agencies with the best ideas, right?
1:17:43
It's like you pay uh you you pay to work with somebody that has a track record of generating great campaign concepts and then they'll often times just like outsource the work >> to people that are good at the execution layer but not at the idea side.
1:17:57
Do you feel like you guys need to develop like a like internal >> like taste or >> Yeah.
1:18:04
Just just ability to like generate a high volume of good ideas now that the actual execution >> uh in terms of like creative production is like so much faster >> uh with AI. >> Yeah.
1:18:17
I think um we think of like creativity not as a monolith but really in terms of two parts. Exactly how you put it.
1:18:24
So there's a taste layer, there's an then there's an execution layer.
1:18:27
our our job here is we want to remove that bottleneck between an idea and a finished product.
1:18:31
Um so internally what we're doing to solve that is like sure we're not going to replace human creativity.
1:18:38
We're going to automate human labor.
1:18:40
Um we're going to make it so easy for like a comedian or a script writer or someone who just wants a part-time job and we're going to pay them like a really high salary.
1:18:48
Um really easy for them to like create the seed of an idea that we can spin off tens and thousands of ads for. Hm.
1:18:57
>> Uh what how much are you guys actually spending on uh on the at on the model on the model side or within any of the applications that you're using to generate this uh this content?
1:19:10
>> Well, for like a 30 secondond to 60-second video, uh really it's like 300 400 bucks.
1:19:16
We have an internal orchestration layer uh that picks the best models to use for all these specific use cases.
1:19:20
It's paired with like a 50-page doc that has all our learnings that isn't available anywhere online and we're able to use these models really effectively.
1:19:28
So our margins like beyond just human labor because we're the ones making the videos and spinning all these things up.
1:19:34
We don't know how to price that is like close to like 98%.
1:19:36
But if you add in human labor I think it's still like above 90.
1:19:42
>> How many different models are you using on an average 30 secondond video?
1:19:44
Do you do you feel like it's worthwhile to stick to one model because you get more of a consistent look or are you jumping around?
1:19:52
How do you think about the different models, what they're good at?
1:19:56
>> Well, it's extremely obvious that some models are just really good at some stuff and really bad at another thing.
1:20:02
Um, Cream is good at specific use cases.
1:20:05
Nano Banana is good specific use cases.
1:20:07
Cling Juan all have their own >> um unique use cases that we we use.
1:20:09
Um, something that's interesting beyond just the models is just like work in terms of workflow orchestration.
1:20:17
Um, before Nano Banana Pro came out, I'll give this as an example.
1:20:21
If you wanted to swap someone else's face, like you would put in, you know, I put in John's face and I say, I want to put Kanye on that.
1:20:29
>> Um, and that wouldn't work.
1:20:29
So, the way you do it is you'd actually tell Nano Banana to cut off John's head and then get that like headless image and put Kanye's head on top.
1:20:37
And that's how we swap faces before Nano Banana Pro.
1:20:39
Um, so there are all these like little workflow >> uh things that we've learned just by experimenting and playing around with these models which play a huge role in making all our ads like the creation of our ads really effective. >> Fascinating.
1:20:55
>> Have we entered a post-slap era?
1:20:58
>> Will we enter a post slop era?
1:20:58
What what what is your post slop timelines?
1:21:03
>> I was speaking to Jess Lee actually about this.
1:21:05
Uh she was talking about how photography used to be seen as slop.
1:21:10
>> Um and because you know it used they used to say that photography was this way like artists were actually painting something. >> Mhm.
1:21:18
>> Um but photography allowed people to realize that allowed people to capture like a smile really quickly through slow motion.
1:21:27
Um, and something will emerge from this AI era where you can do something with AI video to capture some essence of human that you wouldn't be able to do otherwise.
1:21:37
And we don't know what that is, but I'm pretty sure we'll be the first to figure that out, especially if we're pushing out all these videos every month.
1:21:44
>> How big is the team today?
1:21:44
And how's the fundrais going?
1:21:48
>> Uh, it's just me and my two co-founders, Daniel and Damon.
1:21:50
Uh, in terms of the round, we closed uh I can't announce how much, but we closed uh a week and a half ago. >> Incredible. John, hit that gong. >> I will.
1:22:03
>> For Phillip and the Absurd team, uh, thanks for coming on, breaking it down.
1:22:07
Uh, I'm I'm actually surprised there's not more companies trying to do this this exact uh exact sort of playbook.
1:22:11
Uh but uh it's it's cool to hear how you're thinking about this and excited to see uh more of the work that you guys put out. >> Of course. Yeah.
1:22:23
I And by the way, before I go, I'd love to make a launch video for you guys. >> Okay. >> Let's talk. I would love that.
1:22:29
I want to see what I want to see what you can do.
1:22:32
We have we have a benchmark here, Bezel Bench, where it it involves a lot of watches on arms.
1:22:38
We like to put this to the test with a lot of different uh AI video generators.
1:22:43
It's a particularly hard uh shot to get right, but we can come up with a bunch of different ideas. Let's do it. That'd be fantastic. >> Let's do it. >> Perfect.
1:22:51
>> Well, have a great rest of the investor.
1:22:54
>> Uh he'll be in contact.
1:22:56
>> Thank you guys for having me. >> Talk to you soon. >> Cheers. >> Goodbye.
1:22:59
>> Uh before we bring in our next guest, let me tell you about Adio, the AI native CRM.
1:23:02
Adio builds, scales, and grows your company to the next level.
1:23:06
Up next we have Lightberry with Aliar.
1:23:13
Social brains for robots.
1:23:15
>> Social brains for robots.
1:23:15
Let's bring >> I like the sound of that. >> Ali uh Lightberry. Yes.
1:23:18
Uh very interesting to see what robots we're talking about, but we have him here in the studio. >> Welcome to the show. >> Hello. How are you? >> What's happening?
1:23:30
Uh light berry owning yellow. Verticalizing yellow. I love it. I love it.
1:23:35
You know, we have to wear something different.
1:23:37
Everyone's wearing like gray and blue and black and like we need to stand out.
1:23:40
So, yellow underrated color. Underrated color. >> It is. It's awesome.
1:23:45
>> Great to have you on the show.
1:23:45
Uh why don't you introduce yourself?
1:23:46
Give some quick background what you were doing before starting Lightberry. >> Yeah, of course. Um so, yeah. Hi everyone. I'm Alli.
1:23:53
I'm one of the founders of LightBerry.
1:23:55
Um we're effectively just building the operating system for all robots that any person can use a robot.
1:24:01
Um before this I ran a browser company called Sigma OS.
1:24:05
I was running product and design there and I went through YC in summer 21. Um very cool. So that's me. >> Very cool. Um talk more about that.
1:24:11
Uh this feels like a very big opportunity >> but uh I'm not using a lot of robots in my day-to-day life today.
1:24:21
I I assume that I assume that I will be >> uh much more in the future.
1:24:27
Uh but yeah, talk about what the business and the product looks like today and where you see the kind of category going.
1:24:35
>> Yeah, so um we literally have a humanoid robot upstairs right now uh MCing the entire event for demo day and you know he's fully autonomous.
1:24:43
He talks we give him some instructions about like how he should behave for the day and he's just acting like a part of the event staff.
1:24:50
Now you can go out there right now and just buy a humanoid robot from at least 50 different manufacturers.
1:24:54
But if you do that, >> who who did you buy yours from?
1:24:59
>> So ours is from Unitry. It's a Unit robot.
1:25:01
>> Did you buy it on walmart. com?
1:25:03
>> Because I know they sell I know they sell >> Un. No, no, no, not at all. No, no.
1:25:05
We actually work directly with Unitry.
1:25:07
Um, and so like, you know, if you buy a robot from them or any of the 50 others, like it it literally doesn't do anything. It can't talk.
1:25:14
You can't teach it anything.
1:25:16
You can't The only way to interact with is by writing code.
1:25:18
Uh, we thought that's insane.
1:25:20
And so we're building a software layer that allows literally anyone to use a robot by just talking to it.
1:25:27
>> Uh what uh yeah, what what does adoption kind of look like with this?
1:25:31
Like how are you actually selling it?
1:25:33
Is this something that you want Uni to uh kind of uh encourage their customers to adopt?
1:25:39
Because again, I'm sure any manufacturer of robots doesn't want to just sell to developer hacker types that uh happen to want to go through all the different hoops in order to uh actually get value out of a out of a humanoid.
1:25:53
>> I mean, that's exactly it.
1:25:53
You hit the nail on the head.
1:25:55
Like, we're working directly with the manufacturers.
1:25:56
There's like over 50 of them.
1:25:57
Uh we actually just last week closed a deal with Unitry.
1:26:01
Uh they're like they correspond to like 90% of market share in the world.
1:26:05
I'm giving you the air horn, but I have encouraged various government officials [laughter] to ban Uni Tree from the United States. >> Oh, no.
1:26:14
>> So, well, look, you know, the truth is like they're the only ones shipping.
1:26:17
Like, we want to work with the American companies, too.
1:26:19
We want to work with literally everyone.
1:26:20
Uh, but Unit Tree is shipping.
1:26:22
They have market share, so it just makes sense to ship on them.
1:26:23
Uh, we're going to be selling LightBerry powered robots with them all over the US.
1:26:27
But we're also working with other companies, some European ones, some American ones. Uh, >> yeah. What's happening?
1:26:32
Uh, so I would imagine 1X has no interest in in partnering with external providers. That would be my sense.
1:26:41
Maybe that maybe that changes in the future, but I know they're they're trying to really verticalize and I'm sure they want to create a personality and some of the same feature set, but uh what about other other players in the US, Figure, Optimus, etc.
1:26:56
I mean the truth is like they're just not shipping yet and when they want to start shipping and right now they currently don't have any software that allows you to interact with the robot.
1:27:06
There's nothing that works in a public space.
1:27:07
Um I heard that Figures deal with OpenAI just fell through.
1:27:09
I don't know if that's true but like that's the rumor.
1:27:12
We we'd love to help all of those companies get to market faster.
1:27:14
Uh it's just a race right now.
1:27:16
So it's like whoever needs software so that you can interact with the robot we're here to help.
1:27:22
What do you think the uh the most dominant form factor for robotics in daily life will be in just maybe like two or three years?
1:27:31
Do you think we're going to go through like a like a wheeled robot phase or uh or you know one robotic arm on a Roomba phase?
1:27:38
Like how do you see because the the self-driving cars are sort of here, the Roombas are sort of here.
1:27:46
um the the h full humanoid robot that feels a little bit farther out.
1:27:51
Um but is there going to be more of a transitionary phase in your mind?
1:27:56
>> I mean, if you look at sci-fi as an indicator of what people want, we don't just want humanoids.
1:28:00
There'll be different kinds of robots.
1:28:02
Uh you're going to have some like small bipeedal droids that, you know, we work with a few companies that do that.
1:28:07
You're going to have wheeled robots for like delivery.
1:28:09
That's just more practical.
1:28:09
In homes, I actually don't think you'll have humanoids cuz like why do you need locomotion in those cases?
1:28:14
Humanoids are going to be the first like general purpose form factor that's going to make it in my opinion just because you know they look like us and the reason why we're building humanoids is because they be they look like people and so we'll just be deploying them in people facing roles.
1:28:28
So like shop assistants, um manning booths at events, MCing at demo day, right?
1:28:34
Like we have done this before.
1:28:36
We deployed like a fully functional autonomous humanoid at the 11 lab summit uh like 3 weeks ago and it was just working there for 10 and a half hours like fully autonomously alongside the staff.
1:28:45
Um so yeah, that that's that's what we do and and and we think that there's going to be tons of different form factors.
1:28:50
It's going to be like a Cambrian explosion of robots.
1:28:53
>> What are the compute constraints like?
1:28:55
Do you uh do you think ondevice inference is going to be really important?
1:28:59
>> So we run a hybrid pipeline.
1:28:59
Uh we rely heavily on the cloud because that's where the best models are and people prioritize the quality of interaction more than than you know like the reliability of it. >> Sure.
1:29:11
>> Now we also run it as I said hybrid.
1:29:11
So we have an offline version that's also running in the same time.
1:29:15
So if you know connection drops or anything the robot will still talk to you. It'll still understand.
1:29:19
It'll be less smart but it'll know about it. >> Yeah.
1:29:21
>> Yeah. uh have you had any luck uh h I mean how do you think about like personality development and uh I've been very fascinated by the fact that pretty much no lab has been able to hammer out of the model like the it's not this it's
1:29:35
that like they all have this specific LLM flavor to them that I don't think most humans maybe I run into one out of a million people that talk like that but they all kind all the robots talk like robots and I'm wondering if you have any thoughts about where that all goes. I
1:29:48
I think prompting is just I mean these models are getting more and more steerable and they're better at following instructions. Yeah.
1:29:55
So as long as you do a great job of spending time on designing those interactions, you'll be able to get these robots to behave less like robots.
1:30:04
Now we're not trying to make robots that, you know, behave just like people.
1:30:08
Like people love C3PO, but C3PO is very obviously a robot. It has a robotic voice.
1:30:12
It's a bit awkward in the way it speaks.
1:30:13
And like that's the inspiration.
1:30:15
It should just be like smart enough, but it should still like behave and follow our social norms.
1:30:18
Like the robot should look at you when you're speaking to it.
1:30:21
The robot should be wearing the outfit of like the staff members that it's representing if it's at an event.
1:30:27
Um, and that's what we're here to do.
1:30:29
Like we're we're just here to make that easier for all of those manufacturers cuz they're racing on hardware.
1:30:33
They don't have time to think about the software and the interactions.
1:30:37
>> Are you are you excited about robot pets as a category?
1:30:39
I know dogs are are substantially cheaper and that feels like something that a robot pet doesn't need to necessarily add any value outside of companionship and so it feels like a potentially a an area uh where we could see a lot of growth uh in in the near term.
1:30:58
>> So we we actually have like a little pet droid in our office.
1:31:00
It's like a bipedal that kind of looks like R2-D2.
1:31:02
Uh we we brought a bunch of little robots to the event, too.
1:31:07
There's like six of them in the demo for anyone who's here.
1:31:08
Um, yeah, I think robot pets are going to be really big.
1:31:13
It's just we're we started working mostly with humanoids just because the price point is so much higher that we could just focus on quality rather than like trying to optimize for cost.
1:31:22
Obviously, as these robots get smaller, the cost gets lower.
1:31:26
And so, you know, for us, we just want to we just care about quality.
1:31:29
The models are going to get cheaper, too.
1:31:32
So, we'll be able to like deliver on like toys, pets, um, in the near future.
1:31:38
>> Yeah, the toy the toy market seems really really interesting.
1:31:39
Um, >> yeah, our first customer was a toy company actually.
1:31:44
>> Just so much lower in my opinion. >> What about security?
1:31:45
I I feel like there's a potential use case for humanoids just having a human-shaped thing just just moving around.
1:31:50
So literally the landlord the landlord of our building when he um when he he saw that we moved in he stepped into our office and on day one he just asked us like oh so these things can talk and they can walk around they can map the world I was like yeah and he was like you know what I would love to deploy them for security how much does it cost and I told him it's going to be like around 60 to 70k. He's like I want four. I was like okay deal.
1:32:13
So like he he already pre-ordered them.
1:32:15
Like people want this for security not because it can fight not because it can harm people.
1:32:20
These things can't but they're like the best deterrent.
1:32:23
>> Yeah, it's just >> it's the best deterrent and like you know we can literally talk to weird people in the evening and say like who are you and like run facial recognition like are you meant to be here and then just alert like whoever's on like on guard at staff and and just call them and and ask for help.
1:32:35
Like that's how it should work, right?
1:32:37
Robots to help people, not not to replace them. >> Yeah.
1:32:40
>> Yeah. I I do think it's interesting that a lot of these humanoid companies are focused on the hardest possible thing, which is replacing like a house a housekeeper >> who is already not the highest comped person >> doing the most like intricate
1:32:55
specialized tasks where somebody that's a security guard, their prim primary job is to just stand there and look >> like they're paying attention and that's like the job and they make like >> the same amount as a housekeeper. >> Yeah. >> Yeah. >> Yeah.
1:33:08
We don't think we don't think the chat GPT moment for robotics is going to be the day that your robot will know how to fold your laundry.
1:33:14
>> We think it's going to be the day you start seeing robots everywhere in the street or like in shops, in coffee shops, in in events, like talking to people.
1:33:22
Um, >> and that's just really soon. >> It's going to be fun. >> Very cool.
1:33:27
How's uh how big's the team?
1:33:30
>> Uh, we're just a small team of three people.
1:33:32
Uh, we have a few people that we're working with that are helping out on top of it, obviously.
1:33:35
Um, but yeah, it's just a core team of three founders. >> Amazing.
1:33:39
And how's the round going? >> It's been very fun.
1:33:42
I mean, we we managed to close it like pretty early.
1:33:45
Uh, there's a lot of there's some interest now because, you know, like we're with the unitary deal.
1:33:50
We're pretty close to a series A milestone, so we're trying to like discuss that. We'll see. >> There we go. Series A time. Love it. >> Let's go.
1:33:58
>> Uh, really great to meet you. We'll talk to you soon.
1:34:00
>> Thank you for coming on and excited to meet you guys. Have a good one. >> Cheers. >> Bye. >> Yep.
1:34:04
Um, up next we have Dome, a unified API for prediction markets. This should be fun.
1:34:12
It's trying to sit on top of >> Pick a favorite. Pick a favorite market.
1:34:17
>> Well, there is a lot of arbitrage to be done.
1:34:18
Um, on on the uh on the topic of robots.
1:34:21
I I'm just I'm super excited about the lamps that are happening.
1:34:24
Have you seen that there's two robotic lamp companies now?
1:34:28
They're like they're >> One of them was just CGI, right? >> I I I don't know.
1:34:32
Maybe both of them were CGI.
1:34:33
Isn't Apple making their own robotic?
1:34:36
>> It just feels like something that can be done.
1:34:38
Whereas if it's, you know, full humanoid tomorrow for this much money like that feels like a taller order.
1:34:44
It's going to be a couple years away.
1:34:45
But the lamp I feel like we can do today.
1:34:47
The lamp can talk to you. It's going to be funny.
1:34:48
It's going to be awesome. I'm excited.
1:34:49
I'm really bullish on the lamps.
1:34:51
Um, but I'm also bullish on a unified API for uh for prediction markets.
1:34:56
So, we'll bring in the founder of Dome. Welcome to the show. What's going on?
1:35:01
Welcome to the TBP and Ultradome.
1:35:01
You're in the Ultradome and Company's Dome.
1:35:04
Uh, please introduce yourself and your company. What do you do? >> Hi, my name is Kru. We're basically Dome.
1:35:10
So, Dome is a unified API for prediction markets.
1:35:12
In a nutshell, what that means is we allow users and developers to trade an analyze across multiple platforms at once. >> Okay. Who's the customer?
1:35:18
Are you talking hedge funds or like the most advanced traders?
1:35:23
>> Yeah, honestly, it's it's all of the above.
1:35:24
A lot of our current customers are are folks building applications in prediction markets.
1:35:28
So these are folks building like prediction market skins or markets themselves or copy trading and agentic trading is like really popular right now.
1:35:35
Uh we talked to a lot of sports books and hedge funds as well.
1:35:38
They're they're getting interested in high frequency trading and also like platforms like you know things sweepstakes apps folks who are trying to like price internal parlays.
1:35:45
So there's a lot of applications currently being built right now. >> It's crazy. Uh very cool.
1:35:49
Uh who's your favorite Poly Market or Koshi?
1:35:54
>> I'm [laughter] just kidding. I'm just kidding.
1:35:54
I won't make you I won't make you answer that.
1:35:56
>> I was about to say that's the million dollar question. >> Yeah. Yeah. No, no.
1:35:58
I mean, it's uh it's unfort, you know, it's unfortunate that the timeline is just so incredibly uh toxic right now, but I feel like you're able to kind of like sit back and be hopefully like Switzerland and support a variety of different uh exchanges.
1:36:12
How do you think this market actually shapes out? Right.
1:36:19
Uh I think the big news from last week is that Robin Hood is uh getting into the game themselves.
1:36:25
they actually want to not just be a broker, they want to be the exchange.
1:36:28
But how does this how does this evolve?
1:36:34
>> Yeah, I mean we we're supporting currently Poly Market and Koshi. They're both great.
1:36:37
Obviously, we don't we don't pick a winner in the fight.
1:36:38
We want everyone to do well.
1:36:39
Um and what we're currently seeing is there are a bunch of new platforms launching different regions, different specific verticals.
1:36:46
Some folks are just like only sports, some are doing crypto mention markets.
1:36:50
So what we're actually seeing is there's going to be a lot of players coming in each trying to find their specific wedge, find their little market, their community there.
1:36:56
And so in addition, you have Kowi Poly Market, you'll have Robin Hood and a bunch of other big players that are are probably launching soon, but you'll also have a lot of these like smaller players in different specific regions and verticals.
1:37:06
And so we're excited to see like the whole world basically start adopting this.
1:37:09
Do you have a do you have a a reference point for how crossmarket uh transactions like like is there is there a public markets equivalent to you or or some sort of uh like like layer that's not necessarily a hedge fund but like like I I remember reading flash boys and in there they're talking about trading on the commodity markets in Chicago and then also the stock exchanges in New York and but it's done this is all done by the hedge funds.
1:37:39
there's not some sort of intermediary.
1:37:40
Why do we need an intermediary here in this markets particularly?
1:37:45
>> Yeah, I mean it's a great question for what it's flash is my co-founder's favorite book hit it on the nail but yeah absolutely.
1:37:51
So one as you get a lot more providers in right now a lot of the liquidity is fragmented.
1:37:55
So if you actually look at just calcium and poly market themselves about like 80% of their markets their underlying contracts are the same event.
1:38:03
>> So you actually have a good amount of overlap there.
1:38:04
But you also hit it on the nail as well is like there are other markets you can match against like sports books are obviously very very clear.
1:38:10
There's a lot of prediction market overlap there.
1:38:11
Crypto prices per and all these things.
1:38:14
So by kind of taking all this data in creating creating that centralized source it really helps out the hedge funds and those other professional traders who are trying to trade across multiple platforms because everything's in one spot.
1:38:25
spot. Are is some of your volume people just arbing markets on the different predict you know basically seeing like okay what are the odds on Kshi what are the odds on poly market and trying to find alpha through that >> yeah I mean arbitrage is a very like
1:38:40
common request from a lot of our customers right I we actually had a customer that like charted using our APIs like the different prices across the platforms and it's a really cool visual because you can see the gaps over
1:38:50
time of like free arbitrage and so arbitrage is a very common platform u one thing that we do really well is we make sure like when we're matching markets across platforms, we tell them like, "Hey, this is for sure one:1 market versus like a maybe one:1 market
1:39:03
because personally the way we got started was we were trading ourselves and and got burned as well when when two markets look similar but they're not perfectly similar and you lose a lot of money." And so that's something you
1:39:11
And so that's something you think your head think you're squeezing%.
1:39:15
Here's the here's the issue is you can have the same event but then like different criteria in the market based on the platform and where what exchange it's hosted on.
1:39:24
Um, a lot of people have been seeing the rounds coming together for the different uh prediction market platforms and and uh having flashbacks to like open C uh in 2021 and 2022.
1:39:33
Why do you think NFTTS which also saw explosive growth and volume are are are kind of not the right comp for uh this industry? >> Yeah, great question.
1:39:46
Uh biggest answer is like we've kind of seen this exact playbook before.
1:39:50
Both my co-founder and I, we were founding engineers at a company called Alchemy.
1:39:54
>> So they're the blockchain infrastructure layer for anything you're doing in web 3.
1:39:58
Uh they did extraordinarily well and prior to them really like the only really big businesses in crypto was exchanges.
1:40:04
after they came and solved the infrastructure problem.
1:40:06
You saw a bunch of companies build on top of them, including OpenC and Poly Market.
1:40:10
So, we've seen this wave, we've built a lot of like the similar technology, the infrastructure layer at these previous companies.
1:40:16
What you typically see is like there's a huge hype and boom cycle.
1:40:19
Everyone's excited and then like interest kind of fades away, but people keep building and then the next hype cycle you realize, wow, the floors raise.
1:40:26
And so with with prediction markets, you saw this during the 2024 election.
1:40:30
Everyone was super excited.
1:40:31
They thought this was the future.
1:40:31
the election ended, everyone's like, "Oh, this is fine. We'll see you in 2028."
1:40:36
But that but they people kept building.
1:40:38
And then the first week of the NFL Sunday, they did more volume than they did during the 2024 election.
1:40:41
And so that's just more proof to say like yes, there will be boom and bust as far as interest, but the overall market will continue to grow.
1:40:50
uh are you actually routing trades on behalf of uh on behalf of clients or just providing the data layer because I imagine it could get uh quite difficult when some exchanges are using digital asset you know stable coins others are
1:41:07
using traditional fiat rails I'm sure you would need to integrate both um what can you say there >> yeah so first things we start off with is you got to solve the read layer you got to give developers the tools they need to build Right? So that was the
1:41:18
So that was the first version of product is just give them data, give them prices, APIs, tools, whatever they need to display on their applications so that they can build applications. Right?
1:41:26
The next part of our plan was then okay, let's actually start doing order routing and and routing these requests to these different platforms.
1:41:33
And we actually just recently launched our order router as of last week.
1:41:36
And so we will be doing we first are starting off on the crypto angle like processing orders through uh onchain portions and then eventually we'll also do off-chain and and traditional fiat as well.
1:41:46
Uh do you think it's interesting that a lot of the sports books are uh funding lawsuits against the prediction markets while also starting prediction markets uh products themselves?
1:42:00
>> I think it's super interesting.
1:42:00
I think I think a lot of these sports books and sports companies are also very smart and aware.
1:42:05
They understand they kind of see the writing on the wall.
1:42:06
There's so many more advantages to having a pure prediction market, a P2B experience.
1:42:12
It's a lot better for the end consumer as well.
1:42:13
So I think they they they kind of see the writing on the wall.
1:42:15
I think while the the lawsuits are like the equivalent of like maybe the taxi industry suing Uber back in the day, I think eventually most of this industry will move towards prediction markets.
1:42:27
>> Uh how's the round going? >> Round's been good.
1:42:29
We actually closed up yesterday and so super super excited.
1:42:32
We're we're excited to get back to building. >> I had a feeling. I had a feeling. >> I appreciate that.
1:42:36
Yeah, I appreciate the excitement.
1:42:37
It's been it's been an exciting journey so far.
1:42:40
>> Well, thank you so much for coming on the show.
1:42:41
>> Yeah, great to meet you. >> Congratulations.
1:42:42
Appreciate you guys having fun. >> Celebrating domes. Yes, we appreciate it. We'll talk to you soon. Cheers. >> Have a good one.
1:42:48
Uh before bringing our next guest, let me tell you about none other than Turbo Puffer Serverless Vector and Fault Search built from first principles and object storage.
1:42:56
Fast, 10x cheaper, extremely scalable.
1:42:59
Um the Forbes 30 under 30 came out today.
1:43:03
Uh, and liquidity is having some fun because one of the guys who made it, he performed 150% equity growth since 2019, but the S&P is up over 172% over the same period.
1:43:18
>> So, seems >> he made money for his investors.
1:43:20
>> Well, yeah, this is the thing.
1:43:20
He might have taken less risk.
1:43:22
And so, if he took less risk and made almost the same amount of money, then that's good, you know.
1:43:27
So, there is a steel man for this particular person making the 4.
1:43:30
Man, >> but there's some there's some good folks on the 30 under 30.
1:43:34
We'll have to take you through them at some point, but until later, we will >> uh go head over to source and we're going to talk to David who's building Tinder for jobs. David, good to meet you. Welcome to the show.
1:43:47
Thanks so much for taking the time. Introduce yourself. Introduce the company.
1:43:52
>> Yeah, thanks for having me, guys. My name is David.
1:43:53
I am one of the founders of Source.
1:43:55
Source is like Tinder but for jobs.
1:43:58
So, you just upload your resume, swipe right, and AI will apply on the company's website for you. >> Okay.
1:44:04
How is AI actually helping there?
1:44:07
Because I'm still doing the swiping myself.
1:44:09
If I'm looking for a job, the AI is just doing the application. Is that correct? >> Yeah. Yeah.
1:44:14
So, you basically fill out one job application when you first set up the app and then when you swipe, then we have browser agents that will actually fill out the applications.
1:44:24
>> So, it just saves the filling out form time. How's the traction? >> What is Yeah.
1:44:29
So, talk about can you talk about the state of the hiring market because I feel like the >> the number one complaint that uh candidates and people that are applying for jobs have is that like seemingly nobody reads nobody actually looks at job applications and a lot of roles don't actually end up getting hired uh based on traditional job boards.
1:44:49
Um but yeah, what what what can you say about kind of what you're seeing in the market?
1:44:55
Yeah, I guess it very much depends on the company and the role in the sector, but in general, people definitely still get interviews from just inbound applications.
1:45:03
A lot of it is automated and recruiters do kind of like sift through the applicants applications, but I think the number one meta point is that it's definitely a field that's like ripe for disruption.
1:45:16
Like >> you are applying with many many other people and there's typically other ways to get in.
1:45:22
A lot of people email them themselves into a job or a lot of people um refer their way into a job, but >> the inbound is definitely still something that companies use because when you're hiring people at scale, there's just no other way to do it.
1:45:36
Like if you're a company that's hiring like 200, 300 people a month, it's impossible to do it through inbound. >> Yeah.
1:45:42
So where where where what kind of like jobs and and markets have you been focused on?
1:45:46
Because maybe it's not like you know uh other companies in a YC batch.
1:45:50
Maybe that's uh maybe that's incorrect, but where where's the focus been? >> Yeah. Yeah.
1:45:54
I guess a misconception about source is that we're not very directly comp working with these companies.
1:46:00
We're just a traditional job board like an Indeed or a LinkedIn.
1:46:01
So we directly scrape the ATS's.
1:46:03
So right now there's like uh like a million and a half jobs on the app and those are scraped from ATS's like Workday or Greenhouse or Ashb.
1:46:12
So if your company uses that system as an ATS then we've probably scraped your job and you're on source.
1:46:19
>> Are they okay with that? Is that fine?
1:46:20
just just scrape these because I know LinkedIn used to be amazing for scraping and then >> I'm assuming yes because they're like you're going to get more jobs as long as the the ATS's themselves aren't like advertising or marketing like they're just sass right so there there's kind of a contract in this industry to the ATS's
1:46:37
are there to be scraped like Indeed 80% of the jobs on Indeed are scraped most of the jobs on LinkedIn are scraped job boards themselves obviously don't want to be scraped like we wouldn't want to get scraped but the ATS themselves obviously they are just like sending out emails to candidates and managing that whole pipeline. So, So, >> got it.
1:46:53
>> Yeah, that that's completely fine.
1:46:53
And as for how the companies are reacting to it to answer someone's question, um >> like we've helped get over 25,000 interviews in the past year and those range from [laughter] those range from Thank you. >> That's fantastic.
1:47:13
from like >> I guess there's a very wide range of companies like we've helped somebody get a software engineering role at Anderil like a couple months ago but then very often you'll see someone get like a like a line cook job but it's really just >> the universal fact is that filling out the form is very very pointless.
1:47:32
>> How do you do top of funnel?
1:47:32
Like how do you get people to be aware of uh your app actually install it download it?
1:47:37
How are you driving attention on that side?
1:47:42
>> Yeah, we've gotten very good at going viral and getting views.
1:47:46
>> Oh, >> I think over the past year we've done over 100 million views on social media, mostly on Tik Tok and Instagram.
1:47:50
And that again is mostly just like me and my co-founder making >> videos on Tik Tok and Instagram.
1:47:56
We have like I think like 72K on Instagram right now.
1:48:00
And that's just from >> us pulling out the camera and telling people about what we're doing and people like it. So, >> makes sense. >> That's very cool.
1:48:05
How How are you going to make money?
1:48:06
Are you making money already?
1:48:09
So, we actually launched this while we were in school.
1:48:11
Like I I just graduated in May, but we launched this last like at the beginning of the fall semester.
1:48:17
And we used to make money from charging people for or by charging people for more swipes.
1:48:22
We recently have gone like very very free.
1:48:24
Like you really don't need to pay to apply to a lot of jobs anymore.
1:48:28
But yeah, we used to make money from that.
1:48:30
Since we've took taken that down, we don't really make money from that anymore.
1:48:33
And in the future, obviously, we plan to take the traditional job board route and work directly with employers, just faster matches, get more applicants, etc.
1:48:39
But right now, we're very much just like product focused and >> we're kind of wully ignoring revenue. Yeah. >> Yeah.
1:48:49
>> Uh how's how's the how's the round going?
1:48:52
>> Round is basically done.
1:48:52
I think my co-founder is >> talking to investors, but it's really just for fun.
1:48:58
Like we're we're not planning on raising any more [laughter] money.
1:49:00
Um, >> tell them to get back in the in the grind.
1:49:04
You don't need to be talking to investors if you close the R.
1:49:06
>> Uh, >> small recommend small recommend small I I I don't like Tinder 4X. >> Oh, sure.
1:49:13
>> Uh, I'm sure that that actually resonates really well with consumers.
1:49:18
>> But, uh, but, uh, the the the the product experience makes it makes a ton of sense.
1:49:22
Um, >> people think swiping. They know swiping. >> Yeah, they know.
1:49:25
They're not getting away for that. Makes sense.
1:49:28
>> But, uh, but anyways, very very very cool.
1:49:30
Well, congrats on on all the traction and uh uh and hopefully we find some people on source at that point.
1:49:38
>> Yeah, that'd be great. >> Yeah, absolutely. >> Thanks so much. We'll talk to you soon.
1:49:41
Have a good rest of your day.
1:49:43
>> Let me tell you about Gemini 3 Pro, Google's most intelligent model yet.
1:49:46
State-of-the-art reasoning, next level vibe coding, and deep multimodal understanding.
1:49:50
Before we go to the next uh >> to the next guest, uh >> Jurro in the chat says, "I don't know if anyone said it, but the Ryzen X3D is the only way to go for your racing sim." >> That's an AMD chip. That's an AMD chip. We might have to do AMD. I had email.
1:50:09
Give us give us some We've been talking uh >> I think we're dealing with an expert here. >> An expert.
1:50:13
I knew our friend Paul uh who's a a racing enthusiast.
1:50:16
Uh getting some recommendations there.
1:50:19
But uh putting together some rigs for the team. >> Yeah.
1:50:23
Uh well up next we have Mtorial with Kareem the integration layer for AI agents.
1:50:28
Welcome to the [music] show.
1:50:28
How are you doing Kareem?
1:50:31
Thanks for >> finally somebody that is integrating agents. Great. >> Wow. Almost almost correct. >> Okay. What are you doing?
1:50:41
>> So we basically give your AI agents so your LMS access to these apps and data sources.
1:50:45
So anything from your Gmail to your SAP to your Salesforce. >> Okay.
1:50:49
I was just we we were just talking to somebody.
1:50:51
Oh, uh Jason Frerieded, right?
1:50:53
He was saying that OpenAI just wound up building a base camp integration out of nowhere one day.
1:50:56
They just kind of told him, "Hey, it's live now.
1:51:00
You didn't have to do anything."
1:51:00
Uh is that not happening fast enough?
1:51:02
Like in what scenario would I need your service if all of the it feels like there's a massive war going on between the the the LLMs?
1:51:11
They all want to do the integrations as fast as possible.
1:51:15
How is this going to play out?
1:51:15
I mean actually one of the OpenAI uh member of technical staff reached out to us for our product.
1:51:21
So >> okay this makes sense. >> There's that.
1:51:24
But basically one way to think about this is right first of all open AI won't give you AI integrations for the other providers.
1:51:30
People still want to be using Gemini.
1:51:31
They want to be using Anthropic or any of these others.
1:51:36
So we basically provide you with the developer tooling to use any LLM model with any AI integration.
1:51:41
And it's not just the integrations.
1:51:43
It's also these things like access control, right?
1:51:46
Because these Fortune 500s can't just unleash an LLM with access to whatever your Salesforce, SAP, to all the members in their organization.
1:51:53
They need to think very concretely about who has secure access to which models and which data sources.
1:52:00
>> Yeah, that makes a ton of sense.
1:52:00
Um, >> what were you doing?
1:52:02
What were you doing before this?
1:52:04
>> So, I just graduated from NYU Abu Dhabi uh in May and before that I ran a different Abu Dhabi based ticketing startup for around three and a half years. >> Oh, that's cool. >> Very, very cool.
1:52:13
Um what uh what's traction been like?
1:52:15
You said a member of the technical staff at OpenAI [laughter] reached out.
1:52:19
>> That that's that's a development from yesterday.
1:52:21
So not too much not too many updates on that.
1:52:23
But we are open source with over 3,600 GitHub stars and we have uh close to thousand weekly active users uh just since launching around 5 weeks ago.
1:52:35
And then we're also in final stage discussions with some Fortune 500s and unicorns who will deploy this across the organization.
1:52:43
>> [laughter] >> Good good side effects regress organizations with 80,000 100,000 members >> is MCP complimentary competitive substitutive like how does MCP fit into this?
1:52:54
So here's so here's how we think about it, right?
1:52:55
So LLMs 10 years from now will still need access to apps and data sources with access control.
1:52:59
Yeah, right now the standard for that is MCP.
1:53:04
So we basically have this middleware layer translating between our platform and MCP.
1:53:09
>> But if the standard changes a year from now, we just switch to the new standard, right?
1:53:13
Cuz the long-term bet here is not an MCP.
1:53:14
I think that's what a lot of these other companies are getting wrong where they're building 100% on top of MCP but they don't actually think about what these companies need.
1:53:22
They're just kind of following the hype train of oh MCP is the next cool big thing which we are not fully in agreeing agree with.
1:53:30
>> Take me through sort of like the top five agent categories that are interesting to you.
1:53:35
I imagine coding agents are probably at the top.
1:53:37
Maybe knowledge retrieval, deep research agents, maybe >> Yeah, we don't actually think about that. >> Okay.
1:53:43
uh we we are completely unopinionated about how you build your agent.
1:53:47
We just provide you with the integrations. Sure. Right.
1:53:49
Because every agent will need to do read and write operations on these apps and data sources.
1:53:54
And if we can take a tax on that, you figure out how big the market is. >> Yeah.
1:53:59
Uh is there uh I I mean I guess to flip the question around just what uh what agentic capabilities are you excited to see out in the world in 2026?
1:54:11
Honestly, um I really like seeing all these new verticals where basically people just what what do what do they call them?
1:54:18
Those full stack AI native firms where free people go in there use these LM capabilities and these for example legal agents or healthcare agents to compete with >> unicorns or large established players.
1:54:28
I think that's really exciting.
1:54:30
You kind of got this Goliath story there. >> Okay.
1:54:34
So, so walk me through that.
1:54:34
If I'm a lawyer and I'm leaving my firm to start an AI native law firm, um, I might buy some AI legal SAS, but I also might need to integrate with some more niche tools or some more legacy tools.
1:54:48
Um, are you the firm that I would go to to do those integrations for me?
1:54:55
>> Yeah, we basically want to become the substrate for your integrations.
1:54:57
So really long term, we want to have a sort of Oracle story here.
1:55:01
of Oracle story here. how how similar to how Oracle became the substrate for enterprise databases >> and then sold those extra things like enterprise Java etc on top we want to be the substrate for the integration layer and the access control layer and then add these additional things like the workflow builders or also hosting your
1:55:18
agents right so that's kind of the long-term vision here >> I always like to take the temperature on YC folks on like what uh what's breaking out in their supply chain what's a what's a tool or company or service or technology technology that you are leveraging to build this company that uh you're you're particularly thankful for. >> See this might surprise you but kind of
1:55:40
>> See this might surprise you but kind of compared to a lot of other people we are very OG software engineers and what I mean by that is we my co-founder and I have been have had formal computer science education for over 11 years. >> Sure.
1:55:55
So we met in Austrian technical high school at 14 years old for basically computer science and that really allows us to think about first principles.
1:56:03
So in terms of building out our entire infrastructure ourselves thinking about the API designer from scratch and we don't really use that many tools that are available out of the market right now because what we find is that they speed up the process a little bit but we have been doing it for so long that we can just do it better ourselves.
1:56:21
ourselves. So really we invented a lot of new things here as well which kind of the other competitors who are mostly only VIP coding can't even do with their >> you need you need to get an organic certification on the website you know like [laughter] uh this is organic code zero
1:56:36
>> you can get the Austrian armor eagle [laughter] love it uh let me guess the round's already done >> yes very fast actually wrapped up in around 5 days >> 5 days I knew it knew it >> oh >> I knew it uh congratulations love uh yeah loved hearing how you're you know thinking about the opportunity and and how opinionated you are. So congrats on
1:56:57
So congrats on all all the progress. Excited to follow on.
1:57:02
Uh I'm I'm sure uh I'm sure you'll be back on the show soon.
1:57:07
>> We'll talk to you soon. >> Thank you so much.
1:57:08
>> Have a good rest of your day.
1:57:10
>> Before bringing our next guest, let me tell you about Finn.
1:57:12
AI, the number one AI agent for customer service.
1:57:13
Automate the most complex customer service queries on every channel with Finn. ai.
1:57:21
AI and we have Phillip from Crunched.
1:57:24
What a great name for an AI analyst for an AI Excel analyst for Excel power users. Uh just for power users.
1:57:32
>> Have you ever be Have you ever been an Excel power user?
1:57:34
>> You always had to have one hand on the mouse.
1:57:36
You were never just on the keyboards guy. >> Always had one. Very soft. Very soft.
1:57:41
>> I can I can I can hear Andrew Reed losing respect from you all the way from here.
1:57:45
[laughter] >> He's getting cooked.
1:57:47
>> All the way from uh the valley.
1:57:47
Um indeed well he is in the re room waiting room.
1:57:53
Let's bring in Phillip from Crunch to the TVP and Ultradom.
1:57:55
[music] Phillip welcome to the show. >> What's happening?
1:57:59
>> Thanks for joining us.
1:57:59
Please introduce yourself and the company.
1:58:02
>> Hey guys, pleasure to be on. Great to meet you.
1:58:04
Uh Michael actually from from Crunch.
1:58:06
We did the last minute last minute switch here. >> Oh, okay. Good to see you.
1:58:09
>> And the co-founder as well CO of Crunch. >> Fantastic.
1:58:13
>> Maybe I give you like a two second description of Crunch then.
1:58:14
Um, Crunch is your Excel AI analyst built by and for power users.
1:58:19
So, it's like this side panel chat in Excel basically cursor for the world's most popular programming language and then you just chat the natural language and it makes modeling for you. >> Makes a ton of sense. Very clear value prop.
1:58:32
I think everyone who uses Excel wants a co-pilot, but there is a company that's trying to build co-pilot and they happen to own Excel.
1:58:40
How are you imagining this plays out?
1:58:42
Is are you going to live in plug-in world?
1:58:44
Are you going to live at the OS level and be screen scraping?
1:58:48
Are you worried about sharp elbows for Microsoft?
1:58:50
How are you how are you dealing with all that?
1:58:54
>> Ah, it's a great question.
1:58:54
I think um Microsoft is for sure going to build a great product.
1:58:58
They're building a co-pilot for uh two billion Excel users and they're in competition with Google Sheets, right?
1:59:05
I think >> sure >> we're building a tool specifically for the top 1% finance professionals, investment bankers, private equity associates, management consultants of the world and who use Excel in a very specific way, right?
1:59:16
So this is more of the 5 million of the Excel users, the top 1%.
1:59:20
So that's a bit of the bit of the difference.
1:59:22
>> Uh lot of lot of big big market, big opportunity.
1:59:26
If you build a great product, there's tons of people that that will happily pay for it.
1:59:28
Uh there's also tons of startups as well going after this opportunity.
1:59:33
What what do you think uh what do you think they're getting wrong or like you know uh or or is this just going to come down to actual like product quality and and working super closely with these power users to make something that that actually integrates into their uh everyday Excel life. >> Yeah, absolutely.
1:59:52
So I think we have uh plenty of startups going after this opportunity.
1:59:56
We we don't think about competition too much, but out of the big ones with the most fraction, we're the only one with a team that has 10,000 plus real life Excel hours um in our previous jobs, me and me and my co-founder Philip in McKenzie and another finance gigs.
2:00:11
Um and I think that really shines through in the product.
2:00:15
Uh I think also um crunched is modeling more like a real life analyst and performing more of the real tasks that you do on the analyst floor versus like some of the bit artificial benchmarks you see around.
2:00:26
Um so for example crunch can detect mistakes in workbooks.
2:00:31
Um plenty of time is spent in like private equity firms and actually reviewing Excel and making sure they are correct.
2:00:36
As much time as uh modeling from scratch, right?
2:00:39
And then uh these professionals typically work with templates, right?
2:00:42
and they need Crunch to fill out and augment their templates, not build like basic analysis from scratch.
2:00:47
We can do that as well, but we're great at working with large models and and these sorts of things.
2:00:52
>> How do you think about uh the enterprise flywheel here?
2:00:54
It seems like one of Cursor's main advantages is that um they they have a really solid data flywheel.
2:01:04
Now um from open source developers and developers who are not in a you know enterprise level contract I imagine that the top 1% of of investment bankers consultants like on day one they're going to not want you to train on their
2:01:20
data because it's going to be not just some code that builds a front-end website but like extremely critical financial information, private information like it is probably a higher bar to not letting that leak into a training run. So, how do you get a data
2:01:37
So, how do you get a data flywheel going?
2:01:40
How do you improve the product iteratively?
2:01:44
>> It's a good question, right?
2:01:44
And and as you say, security is top concern, I think, for all of our customers.
2:01:48
We live with like global consulting firms um but also >> let's give it up for global consulting firms.
2:01:55
They don't get enough love.
2:01:58
[applause] >> They don't get enough love. [laughter] Except here. Except here. Except here.
2:02:03
>> They just don't >> I will I will defend. [laughter] >> Exactly. Exactly. Well, >> that's good.
2:02:11
Um, no, but but they're obviously super concerned about their security, right?
2:02:14
And do live public deals, right? All of this stuff.
2:02:16
Um, and so like >> uh in in in principle, we do not uh train uh the on the data of our customers and we cannot see what the prompter do, right? Yep.
2:02:26
>> At the same time, what we want to do now and just in record time closed our our fundra um we want to make sure that we tailor crunch to every single firm um as well. That's great.
2:02:37
Um um and then um when we tailor it, we have discussed with a few customers like the opportunity to like for some of the large organizations, they do enough modeling work on a global base that is possible to like tailor do some finetuning and tailor to their specific organization.
2:02:52
Um but as well we come in in a forward deployed manner, right?
2:02:54
and and do customization whether that is formatting or or solving for their specific workflows and linking into their templates.
2:03:01
um how like the SIMs that they get, how can we link that into their specific LBO template and then transfer that from like the simple LBO to the advanced LBO and and these sorts of custom.
2:03:11
>> What's the biggest deal uh Crunch has uh supported?
2:03:16
>> The biggest deal we have supported, that's a good question.
2:03:17
I can tell you about the the >> You don't need to name the company.
2:03:22
Yeah, you don't need to name the company.
2:03:23
>> It was a $500 billion company. They were doing a $1. 4 trillion deal.
2:03:25
They were doing about 20 billion in revenue.
2:03:28
Not going to say who it was. [laughter] >> Exactly. Exactly.
2:03:34
No, but I can tell you a real story about a mistake we caught though.
2:03:37
Crunch has this uh error uh detection system. Okay.
2:03:41
>> Um and on a live deal for an associate at one of our private equity clients uh in London um used the sort of crunch mistake detection system to uh to identify or like scan his one of his previous models on a real transaction and identified a mistake in the working capital that overvalued the deal by 10 million pounds. So, >> wow.
2:04:00
That one is uh >> You got You saved his job.
2:04:04
>> Send him an invoice for 5 million right now. He just saved him 10. I give me 50% of that.
2:04:09
That's your seed round right there. >> Exactly. Exactly.
2:04:12
>> Well, congratulations [clears throat] on a fantastic demo day. Thank you so much. >> Yeah. Great.
2:04:16
Great to meet you, Michael.
2:04:17
>> We'd love to talk to you tomorrow.
2:04:19
>> I live for Excel Asians.
2:04:19
I'm so excited about this category.
2:04:22
It just feels like >> I would love to. Thank you.
2:04:25
Um well, have a great rest of your day.
2:04:27
We will talk to you soon. Great hanging, Michael. >> Goodbye.
2:04:31
>> All right, thanks guys. >> Numeral. com.
2:04:35
>> What $500 billion company could that be? >> Numeral. com compliance handled.
2:04:38
Numeral worries about sales tax and VAT compliance so you can focus on growth.
2:04:47
Uh speaking of growth, uh there's some there's some folks putting Menllo Ventures in the truth zone.
2:04:52
uh enterprise large language model API market share has been falling for open AI.
2:04:59
It's been climbing uh amongst anthropic according to Menllo Ventures.
2:05:05
Ev Randall puts it in the truth zone over at Benchmark uh multi-time TVPN guest Ev Randall.
2:05:10
He says people are quoting this Menllo Ventures chart and extrapolating from it like it's official data from the Federal Reserve or something.
2:05:18
It's a small sample survey conducted by an investor inanthropic. Please calm down.
2:05:23
Uh I like that he's pouring some some cold water on this.
2:05:26
Um >> this was from uh November 3rd. Yes. Too.
2:05:31
So >> at the same time, does is it possible that OpenAI's enterprise large language model API market share is falling? Sure.
2:05:39
You know, they were the only game in town when they launched.
2:05:41
[laughter] Uh and so you would expect their market share to fall a little bit over time.
2:05:44
Uh will be interesting to see.
2:05:46
we will get more data on this.
2:05:48
All these companies are going to be public in a couple years and so we'll know exactly how it's breaking down.
2:05:54
But uh until uh until that happens, we will return to our coverage of YC Demo Day 2025.
2:05:59
Uh we have Sava, the AI powered trust company. Welcome to the show. How are you doing?
2:06:08
>> Well, what's going on? >> Hey, I'm doing great. How are you? >> We're fantastic.
2:06:11
Please introduce yourself. Introduce the company.
2:06:13
Tell us what you're building. >> Great.
2:06:15
Yeah, I'm uh I'm Nimit Maru. Um we're building Sava.
2:06:18
We're building a new modern agentic trust company. >> Okay.
2:06:23
>> Um that uh administers advanced trusts.
2:06:26
>> So is this specifically tr like will and trust? >> Yeah.
2:06:30
So it it's trust like will and trust. Yeah. Exactly. >> Yeah.
2:06:33
Not not because people would say a trust company could be somebody that makes sure your your password doesn't get leaked or something.
2:06:38
But uh this is >> How old were you when you realized you wanted to use AI to spin up trust?
2:06:45
[laughter] No, I'm just kidding.
2:06:45
Uh, what uh what were you doing?
2:06:46
What were you doing before this?
2:06:49
>> Uh, well, my my previous company was actually uh in John's um way batch uh summer 12 batch. >> No way. >> What company? >> Yeah.
2:06:57
And uh we we so at the time we were building Yelie, which was a um we like you know like the frontfacing camera had come out on the iPhone.
2:07:04
So we wanted to build like a tele medicine but we pivoted to being a early uh code education uh and like tech education company and that's how we kind of built that.
2:07:14
um and then sold it in uh like 10 years later.
2:07:20
>> If you if you hadn't pivoted, you could have been selling meth at scale like some of the other uh tele medicine [laughter] companies.
2:07:25
But no, I'm glad you did.
2:07:27
I'm glad you went the code. Very cool.
2:07:29
>> Did you say selling Did you say selling meth?
2:07:31
>> No, no, I'm just I'm just I'm just joking because there were some pills there were some kind of tele medicine companies that went a little bit too far and one of the founders is in jail now.
2:07:42
>> They're like check if the patient's breathing.
2:07:43
If they are, give them matter all.
2:07:45
[laughter] >> That's that was going >> um no uh more seriously talk about uh what's are these Nevada trusts?
2:07:50
Like what's what's what's the uh what's happening at at the actual like entity layer? >> Sure. Yeah.
2:07:58
So we so we're not drafting the trusts.
2:08:01
Um we we will uh basically like a uh an attorney or a uh like a fintech or legal tech that uses LLM to to draft trusts.
2:08:09
So they would create the trust document >> and then once they uh need someone to administer the trust to be an independent trustee um that's when we would take over.
2:08:19
Uh we're getting our charter in Nevada.
2:08:21
So we're going to be uh chartered in Nevada.
2:08:23
Um you know maybe eventually we'll go to other states but that's where we're going to be right now.
2:08:28
Um and yeah we work directly with attorneys, wealth managers, uh fintexs to serve as the trust administration there.
2:08:35
So, so would you uh do you have like no consumerf facing brand essentially?
2:08:40
It's like purely B2B at that point.
2:08:43
>> Well, I mean it is consumerf facing in the sense that the people who would be using it are also the families who uh have these trusts.
2:08:49
But the reason I say we work with attorneys is cuz generally the families are taking advice from the attorney or the wealth manager about which trust company to choose because I mean you know like how how would a family know even what a what a trust company is or so?
2:09:04
So, so we think of them as the ICP. >> Sure. Sure.
2:09:08
>> Are trusts underrated? >> Yeah.
2:09:11
I think they're underrated and they're underutilized and also right now they're very annoying and expensive to create and manage.
2:09:18
>> And so I think people don't use the power of trusts enough.
2:09:20
Um and that's not to say like you know every American or every person can be using them but definitely a big uh slab of people you know kind of um below where right now they are being utilized. >> Yeah.
2:09:35
Do you have a ballpark uh cost figure for you know uh doing a trust like at what at what scale does it start to make sense for customers to even participate in the market to even consider a trust?
2:09:49
So I think creating a a a revocable trust that you know owns your house or other assets that's applicable at you know at at like reasonably um you know like almost any uh level for at at like when someone would own some property >> just as soon as you make sense. >> Yeah.
2:10:05
But then using using something like Sava today, it's generally people who are trying to make um irrevocable trusts.
2:10:12
And so they would tend to have you know like some some millions like you know maybe like low single digits but or maybe mid single digits millions in assets before they start utilizing that.
2:10:23
Um I think that as as like tech makes it a like a lot um easier and cheaper to create trust in a good way and also um you know people like us can make it a lot more friendly and modern to administer trust like I think more people will be able to >> um use them.
2:10:43
[clears throat] Yeah, >> it should just get way cheaper.
2:10:44
I mean, if you think about just the YC story of how much it cost to set up a corporation and raise a seed round in 2005 or something, you were looking at like 20,000 maybe 50,000 in total like fees across everything.
2:11:05
Now it's like Stripe Atlas, one click, they charge you what 200 bucks or something and then 500 bucks.
2:11:12
500 bucks and then and then the safe is like 1 second and you know administered by a bunch of folks and like it's like really really low low cost and that's obviously led to just more entrepreneurship.
2:11:23
entrepreneurship. You imagine that something similar happens >> when the infrastructure gets better like usage goes up and even the safe like I was talking to my co-ounder the other day like the safe is an incredible invention that makes this like early stage of fundraising you know so much
2:11:37
smoother like back when we did it in the summer 12 batch it was like all convertible notes and you know even that a lot of investors wanted price rounds at this stage it's like a pretty >> um difficult thing so yeah I think when the infrastructure gets better like more people utilize it and um like more people can take advantage. >> Well, congratulations on the progress.
2:11:56
>> Well, congratulations on the progress.
2:11:58
Thanks so much for coming on the show.
2:11:59
>> Yeah, excited to check the product out >> and we'll talk to you soon.
2:12:01
Have a good rest of your day. >> Thank you. >> Cheers. >> Goodbye. >> Thank you, Pat.
2:12:05
>> Uh let me tell you about profound.
2:12:05
Get your brand mentioned in chat GBT.
2:12:07
Reach millions of consumers who use AI to discover new products and brands.
2:12:11
Um I want to pull up this chart of the day from CO2.
2:12:15
They say, "Hey, look, there's no code red here.
2:12:18
It's all Baja blast because Chad GBT traffic historically dips this time of year.
2:12:24
Um, and it's a fascinating chart.
2:12:27
If you actually zoom in on this Gemini 3 launch day, it looks like people stop using LLMs around Christmas.
2:12:36
The turkeyy's going around.
2:12:39
The tryptophan is coursing through their blood.
2:12:40
They're getting a little sleepy.
2:12:41
>> They're getting a little sleepy.
2:12:41
They're having an extra bottle of wine and they're taking time off from their chat app specifically from uh from Chat GBT.
2:12:50
Uh this is bizarre this that this chart tracks so much with when people do work.
2:12:55
You can see that Chad GBT grows in the in the uh in the spring every year up until summer.
2:13:00
Then it completely flatlines during summer.
2:13:02
Then it peaks when school year starts again and work starts back up.
2:13:05
Then it crashes on Black Friday.
2:13:09
Students and and students and workers, people with jobs. That's everyone. That's everyone. Come on. That's everyone.
2:13:16
>> What about the unemployed? >> Oh, yes.
2:13:17
I don't know what Well, they're they're the ones that are holding it up.
2:13:20
They're the They're holding it down during Black Friday.
2:13:22
They're like, "I'm still grinding."
2:13:23
Uh but clearly folks did not get the uh the great lockin memo because the whole point was that you were supposed to continue to use all the AI apps.
2:13:32
Uh anyway, it's a fascinating it's a fascinating chart.
2:13:34
I'm sure we'll be digging into it more, reading the tea leaves.
2:13:38
But up next we have Ben from SF Tensor. It's Versel for GPUs. Welcome to the show. Thank you so much.
2:13:44
Please introduce yourself and the company. >> Great to have you. >> Hi. Yeah, thanks. I'm I'm Ben.
2:13:47
Uh we're building the infrastructure layer for AI researchers.
2:13:52
So basically from you know training models from like small experiments all the way up to large scale frontier training runs.
2:13:57
We basically deal with the infrastructure to allow you to do all your training runs. >> Okay.
2:14:04
So there's a bunch of different layers going down to somebody that owns the ground, somebody that builds the data center, somebody that racks the GPUs, and there's then there's the Neoclouds.
2:14:13
Are you interfacing with multiple Neoclouds? Are you a Neocloud?
2:14:18
How are you positioning yourself?
2:14:19
>> Yeah, so we we work with all sorts of Neoclouds and hyperscalers and we basically just say we're building above all of them.
2:14:25
And so our customers should only be worrying about what they want to be researching or training and not like how the actual technology like the underlying stuff works.
2:14:33
And so we deal with you know finding GPU allocations for different GPUs.
2:14:38
So we also allow you to work with TPUs or AMD GPUs or any of this stuff to allow you to train your models. >> Okay.
2:14:44
So this is specifically for research and training runs and you're and and less uh focus on on like actually inferencing on the product side. >> Yeah.
2:14:54
So we focus exclusively on the on the training side.
2:14:56
the training side. There's great companies even you know from last batch for example there's luminal they do great things for inference uh we focus just on training because we think training is a problem that's not been solved by anyone and there needs to be way more training happening
2:15:12
>> uh what are your clients uh like what's the shape of them I guess there's a lot of focus when when people think training they think open AI anthropic Google deep mind right but take me through the variety the landscape ape of folks that you talk to who are actually doing training runs. Who are these folks? You Who are these folks?
2:15:30
You don't have to give exact names, but tell me the the shape of their workloads, how they're what problems they're trying to solve, the scale of their training runs.
2:15:40
Take me on a little tour. >> Yeah.
2:15:41
So, there's there's a huge variety.
2:15:43
I mean, you have on the one hand, you obviously have like the academic, you know, or home researchers at home who are training like small models and then you have, you know, larger scale academic research happening.
2:15:53
But then you also have startups that have raised maybe you know call it $10 million.
2:15:56
You know there's some companies from from YC as well who are training models for super niche use cases.
2:16:03
And then there's also you know companies that have raised hundreds of millions or you know up to a billion dollars.
2:16:08
There's a bunch of labs actually in that like area who are doing who are training their own models.
2:16:12
You don't just have anthropic.
2:16:13
I mean like >> the the textbased models like LLMs there's not an awful lot of competition going on there anymore.
2:16:19
Things have sort of converged at the top there.
2:16:21
But for everything else like you know drug discovery or um you know protein folding all of these things are still problems that have not been solved by anyone.
2:16:31
>> Is it correct to say that SF Tensor is a bet that there will be millions of of of uh smaller models for specific use cases or or one day billions?
2:16:44
>> I wouldn't say billions but definitely a lot more than there is today especially just in the modalities that haven't been explored today.
2:16:51
I mean, we're all focusing on text and text is great for a lot of things, but I can't really use a textbased model to do things like, you know, um text to speech, for example, is another type of model or we have protein folding models or all of these things can't really be solved with text.
2:17:03
We need models that are specialized in those pockets.
2:17:07
What what about uh I mean we were talking to the CEO of AWS yesterday and he was saying that uh AWS launched a product that that is actually a checkpoint uh 80% of the way done on an actual foundation model and then a company can come in and add their own data to the pre-train and then they can do everything else with it.
2:17:27
Uh, and that felt like an interesting proposition when you think about if you do want a textbased model and you want it to be to really know your company's data at the core in the pre-train, really know it, not just drop it in the prompt, not just fine-tune on it, actually bake it in.
2:17:45
Uh, that feels like we're going to see a Cambrian explosion of every company wants their own trained model earlier.
2:17:52
They're going to want training workloads for that.
2:17:53
Is that something that you think you can play in?
2:17:54
is are there already other companies that are working there?
2:17:57
How do you think about that?
2:18:00
>> So, it's a very unexplored area so far.
2:18:04
The idea of basically saying you have like you know 80% of the way the model can already form coherent sentences have basic reasoning abilities and then I add my own information.
2:18:12
my own information. I think that's going to be very important in the future just because it allows me to take a base model and then not just do like post-training but sort of you know continuous pre-training almost you know continuing the pre-training I think
2:18:29
there's going to be a lot of use cases that come out of that and I think we can we can help there I mean we don't really care what you're training on the hardware we you know if it's if it's an AI training um we can we can help with that so you know that's definitely something we're looking into. >> You want to ask about progress?
2:18:45
>> You want to ask about progress? >> Yeah.
2:18:47
What kind of metrics were were you sharing today uh during demo day? >> Yeah.
2:18:51
So, the metric we're sharing is we launched like two weeks ago and we did $41,000 in usage based revenue since then. >> There we go. Love it.
2:18:59
Uh and how's the round going?
2:19:05
>> We we closed the first day of of fundraising.
2:19:07
So, >> first day of fundraising. >> There we go. >> There you go.
2:19:11
I'm not going to I'm not going to dox but a a friend of ours.
2:19:15
>> We got a text message about you.
2:19:16
>> We got a text message about you.
2:19:16
A friend of ours just backed one company this batch and he's known for backing great companies and he just backed you.
2:19:24
Uh so I'm excited for for you guys to announce the round soon.
2:19:26
Uh and come back on and do it on TBPN. >> Thank you so much. >> Awesome. Great to meet you.
2:19:32
>> We'll talk to you soon. >> Cheers.
2:19:33
>> Have a good to meet you. >> Good to meet you.
2:19:36
>> Let me tell you about getbasel. com.
2:19:38
shop over 26,500 luxury watches.
2:19:42
>> Super intelligence for your wrist.
2:19:43
>> Fully authenticated in-house by Bezel's team of experts.
2:19:46
Um Brad Gersonner on Trump accounts pus was elected on main on a main street agenda to get the rest of America into the game.
2:19:53
And that's exactly what this does.
2:19:55
Bill Gurley showing him some respect.
2:19:57
Uh and we didn't cover it yesterday, but Michael Dell uh donated $6.
2:20:01
5 billion to these Trump accounts.
2:20:05
the uh the accounts where children get them, they can't be touched.
2:20:10
They're invested and they compound over >> $50 for a bunch of >> Yes. >> individuals.
2:20:17
>> And and and there was some push back.
2:20:18
Some people were saying, "Well, you if you compound at the S&P, even if you compounded 10% for 20 years, like it's only a thousand bucks or a couple thousand bucks.
2:20:25
It's not that much money.
2:20:26
It's not life-changing, but you know, it's like a piece of it's one that's that's just Dell's contribution.
2:20:32
Like there's going to be other people that are contributing corporations 000 from America >> and and Yeah. Yeah.
2:20:37
And there's a whole bunch of other ways to add money to the account over time at birthdays and Christmas and stuff targeted donations >> and the most important thing is that it's a lock box.
2:20:47
So it's psychologically a lock box.
2:20:48
So I still stand with the uh with the Gerson accounts. >> It's incredible.
2:20:53
>> But we have our next guest in the ream waiting room uh from Locus payment infrastructure for agents. How are you doing?
2:21:00
Please introduce yourself and tell us what you're building.
2:21:04
>> Tie-dye shirt on TVPN.
2:21:04
We've done over a thousand interviews.
2:21:08
>> I don't think we've ever seen one. This is unique. I like it. >> It's a first. It's a first. Thank you. >> Yeah.
2:21:12
And and they uh you know, thank you.
2:21:13
And so they actually switched me up uh with the other guy. >> I sorry. >> Great.
2:21:19
Well, Henry Tie, welcome to the show. >> Yes, sir. >> Henry from Icarus.
2:21:22
Uh please introduce yourself and tell us what you're building.
2:21:27
>> Yeah, so I'm Henry, founder and CEO at Icarus.
2:21:29
my background, aerospace engineer at Georgia Tech.
2:21:31
Built drones for NASA and satellites at orbital.
2:21:35
>> Icarus were building solar powered autonomous drones that fly at 60,000 ft >> for weeks at a time. >> Close to the sun. >> How close to the sun?
2:21:42
Not the closest, but >> not the closest.
2:21:45
And and in fact, if if we flew any higher, we'd actually fall out of the sky.
2:21:48
So, we want to stay at 60,000 ft.
2:21:51
>> You're like You're like, but but but we're we're gonna try flying a little higher. >> Okay.
2:21:55
How many how many hard tech How many hard- tech companies were were in this batch?
2:21:58
M >> uh I believe like 5 to 10.
2:22:02
>> Yeah, that seems about right.
2:22:02
And so I feel like it's been like steadily at >> 5 to 10 for >> forever basically.
2:22:10
So >> uh take me through uh the bare case for Stratospherics drones.
2:22:14
Um yeah, what I've heard is, you know, people always refer to the SR71.
2:22:20
It's such an amazing plane.
2:22:23
It flies, I think, around 60,000 feet. Uh the SR71 Blackbird.
2:22:25
It's this amazing Loheed Martin plane built at Skunk Works. Flies super fast.
2:22:30
Uh we can't build planes like that anymore. We don't have it in us.
2:22:35
And when I talked to folks who were like, "Yeah, it kind of sucks we can't build that because it was really cool, but we have satellites now and satellites go way higher and way faster.
2:22:44
And so if you need to put a camera over something, uh we usually just use a satellite.
2:22:48
So why not satellites for this use case?" >> Yeah.
2:22:52
From first principles, you're 20 times closer than lower Earth orbit. >> Mhm.
2:22:57
and you can say fix an area.
2:23:00
>> So just from an engineering perspective, it makes a lot of sense.
2:23:04
>> The bare case is pretty much like none of this is new.
2:23:07
Even what I'm doing, the solar powered version, >> um it's all been done.
2:23:11
It's just been too expensive. >> Sure.
2:23:14
>> So the question is like can you get the cost down?
2:23:19
>> Can you how are you doing that?
2:23:19
Is it just being a startup?
2:23:21
Like are you using cheaper materials?
2:23:24
Are you standing on the shoulders of giants?
2:23:25
It's like what are you leveraging to actually make it? >> Yeah.
2:23:28
Well, we'll talk about the form factor first because I'm on the website and this thing just looks like a massive really skinny bird.
2:23:34
[laughter] >> So, it's very unique. Very unique. It's Icarus1. com.
2:23:39
Icarus >> or sorry, Icarus. 1. >> Icarus. 1. Correct. >> Yeah.
2:23:44
The to your point, John, it's about getting the right product specifications. >> Okay.
2:23:50
>> For the first go to market. >> Oh, wow.
2:23:52
And so, yeah, our first product, it's a 20 foot solar powered bird.
2:23:56
Uh, flies for weeks at a time. >> Yep.
2:24:01
>> The bird the bird noise is perfect.
2:24:04
>> Uh, so so it's effectively like a loitering drone that's just sitting at 60,000 ft.
2:24:08
And it's I'm assuming it's it's incredibly light.
2:24:10
You're you're uh it's it's solar.
2:24:13
It has a battery, but it can it can uh generate uh solar power on the fly to to increase the battery life effectively.
2:24:21
Like it's not efficient to hold it in the air forever yet, but >> uh but it can stay up over a specific area.
2:24:29
So is this primar primarily like defense applications early on?
2:24:32
Who who are you trying to sell this to?
2:24:38
>> Yeah, act one is all defense.
2:24:38
I do think this is much bigger than a defense company.
2:24:42
I [snorts] do see the stratosphere as a category and um and once you kind of are able to uh make the stratosphere affordable then there's many things you can do.
2:24:54
So one easy example like yeah today you can't really carry very heavy payloads.
2:24:58
You can't carry and deliver a lot of power but what the future looks like and there's like no laws of physics that says you can't do this.
2:25:05
You can essentially take like a Starlink satellite and have that in the stratosphere and imagine if you had this Starlink satellite that's 20 times closer and fixed over an area.
2:25:14
closer and fixed over an area. Uh so then that's like that's the future and what you can do from that it's I don't [snorts] know it's anyone's imagination near-term there's a lot of clear direct uh line of sight towards defense and a market there again it's like really
2:25:30
difficult it's not it's not like a a category yet today there's there's no real markets but uh with defense there's there's a clear need >> uh very cool how do you actually get the drone up is this something that you launch uh like a rocket hit and then it and then it sort of spreads its wings at some point. Like how do you actually get
2:25:48
Like how do you actually get a 20ft drone 60,000 feet in the air? >> Into the air. >> You use a balloon. >> You eat it.
2:25:56
>> Oh, you use a balloon. Okay. >> Okay.
2:25:57
That's [laughter] That seems less violent than yeeting a 20 foot drone.
2:26:01
>> Some Some drones are yeetated.
2:26:01
I believe this is a real thing.
2:26:06
>> So you use effectively like a weather balloon to take it up.
2:26:08
>> Are you a beneficiary of Starlink?
2:26:11
>> Are are we a competitor? >> No, no, no.
2:26:13
A beneficiary like like beneficiary >> like like can you use Starlink effectively as like the backbone for communications?
2:26:21
>> That is our beyond line of sight method. >> Sure. >> Yeah.
2:26:24
So we have Starlink on it as an option.
2:26:26
>> Yeah, that's very cool. Um >> yeah. Yeah. Fascinating.
2:26:28
So uh how how close are you to actually getting this up in the air? Have you flown?
2:26:32
Uh is it just test at this point?
2:26:35
Are you actually going to sell these things?
2:26:36
How how >> uh we are selling them today to the Army. Okay.
2:26:40
And yeah, we've done uh over 30 successful stratospheric flights, successful demos with Special Ops Command, SOCOM, and the Army as well.
2:26:49
And we have Oh, there you go. >> There you go. >> There we go.
2:26:54
>> Yeah, super impressive traction.
2:26:54
I noticed uh is it Ronic on your team?
2:26:55
Was that Red Bull Racing before this? How cracked is Ron? >> That's awesome.
2:27:04
>> He is uh he is very very hardcore.
2:27:04
Um, >> like I imagine if you want to make something that's ultra light, ultra durable, he's your guy. >> That's right. That's correct. Yeah.
2:27:13
So, a third of our team is SpaceX Tesla.
2:27:15
Uh, Ronx worked at Tesla before Red Bull Racing and also SpaceX as well, but he's definitely a character. Um, >> yeah. >> Awesome.
2:27:24
Uh, well, great to meet you.
2:27:26
I'm excited to uh to follow along.
2:27:26
How the round already done? How's it going? >> Yes.
2:27:32
Uh, raised a lot of money.
2:27:32
Um, [laughter] >> there you go. There we go.
2:27:37
>> Hit the gong again, John. >> There we go. >> Yeah, buddy. Yeah, buddy.
2:27:43
>> Yeah, >> just coming on. >> Absolute legend. You're a TVPN legend.
2:27:46
U We might have to uh we might have to make a TVN tie-dye shirt in your honor.
2:27:54
>> Y [laughter] I love it. Thank you so much. >> Very cool. Very cool.
2:27:58
>> Well, have a good uh rest of demo day.
2:28:01
Congratulations on all the progress.
2:28:01
Uh very excited to see these up in the stratosphere.
2:28:05
Just uh don't fly them too high. >> Yep. Exactly. Perfect.
2:28:08
All right, Jordy, John, thanks so much.
2:28:11
>> Have a good rest of your day. Goodbye. >> Um >> what a legend.
2:28:15
>> Uh I need to know from you if we have some breaking news that we can share right now.
2:28:20
It sounds like we might have some surprise guests joining the stream. So stay with us.
2:28:24
Uh I will also tell you about adquick. com.
2:28:28
Out of home advertising made easy and measurable.
2:28:31
plan, buy and measure out of home with precision.
2:28:34
Um, also uh people are calling for Google to make glasses now because uh Google Glass they did this 20 years ago practically Google Glass.
2:28:44
Uh and >> but they're work they're still working. >> Yes. Yes. Yes. Yes.
2:28:49
But through partners through partners.
2:28:50
through partners. So, so, uh, they they've done the Google Pixel, they've done a they've done a variety of of hardware devices and they're and they are working on, uh, some augmented reality glasses again, but they're certainly not making as much of a, you
2:29:04
know, big push, media push as they did with the original Google Glass, which was like it dominated the news and it was like the future is here and then the product didn't really get to escape velocity and is sort of remembered as a failure, but it wasn't a [snorts] failure. They were just early. they were They were just early. they were just early.
2:29:20
Uh, and that's the important thing to remember.
2:29:22
But we have our next guest here in the Reream waiting room.
2:29:26
Let's bring him in from Locus. Welcome to the show.
2:29:29
Thank you so much for taking the time to join us.
2:29:31
Please introduce yourself and tell us what you're building. >> Yeah, for sure. So, I'm Cole Dermott.
2:29:37
I'm the CEO and co-founder of Locus.
2:29:37
We build payment infrastructure for AI agents. >> Okay.
2:29:41
MCP currently doesn't have payment infrastructure. That's why you exist.
2:29:44
Is that's what's going on? >> Yeah, basically. Plus trust. Okay. Okay.
2:29:48
Trust is a huge part of it. >> Okay. Interesting.
2:29:50
How are people are people actually like solving this manually right now?
2:29:53
Are there are there payment are there like are there like agentto agent payments that are happening right now or is this something where we're thinking like in the future they will all be flowing stable coins to each other in the future?
2:30:06
>> I think agentto agent isn't really adopted yet.
2:30:08
What we're looking at right now is more so developer use cases of um if if you're familiar with X42 paying for API endpoints on a pay-per-use basis potentially doing payouts to people.
2:30:18
Y >> uh the way I like to explain it is historically payment automation has been deeply rooted in in conditional automation a series of ifs, ends, ors etc.
2:30:29
>> Now with Aentic payments you open up this new frontier of contextual automation right and that's a pretty huge evolution. Mhm.
2:30:35
Um what uh how do you imagine the first adoption of agentto agent payments or or even just payments for agents broadly uh playing out?
2:30:48
I I was me and Jordy have been talking about this with the agent commerce stuff. We're using Chat GPT. We're using Gemini.
2:30:53
Uh there's all these times when I run into a payw wall and I can tell it's running into a payw wall.
2:30:58
It's like a actually I can't tell you about uh you know what's going on on that website.
2:31:02
And I'm like no you actually could if I gave you my credit card.
2:31:04
I know you could uh but they can't and it seems like that's something you could potentially help with.
2:31:08
But how do you see the first early adopters using your service?
2:31:13
>> I see the first ones as really developers building these um autonomous agents, right?
2:31:17
Being able to essentially pay for services as they do their workload in the wild and discover those services autonomously, right?
2:31:23
Um, in terms of like the more commerce side, I think that'll be an industry that evolves over the next few years as trust is really developed because frankly on a on a widescale consumer basis.
2:31:33
That's really the biggest barrier right now is trust rather than tech. >> Yeah.
2:31:38
Uh, what kind of numbers did you share uh during your pitch or are you planning to share?
2:31:44
>> Yeah, so we processed around 3,500 transactions and have around 80 projects built using Locus so far. >> Amazing. >> That's amazing.
2:31:50
>> Uh, what were you doing before this?
2:31:52
John's got the gong for you. Hit it. Hit it.
2:31:55
Then what are you going to do?
2:31:55
Uh what what were you doing before this?
2:31:58
>> Yeah, so I interned at Coinbase.
2:31:58
Uh I was one of the people who helped build Coinbase business over there.
2:32:02
My co-founder was one of the six software engineering interns at Scale AI.
2:32:07
>> Uh studied CS at Waterlue, business at Wilfford Laurier, was the financial lead at Waterl Blockchain.
2:32:11
So >> Waterlue mentioned >> fantastic.
2:32:14
Well, thank you so much for coming on the show.
2:32:16
Congratulations >> and I'm sure we'll be seeing you soon.
2:32:22
Have a good rest of your day.
2:32:24
>> We'll talk to you soon.
2:32:24
Uh, let me tell you about wander. com.
2:32:26
Book a wander with inspiring views, hotel, great amenities, dreamy beds, top tier cleaning, and 247 concier service.
2:32:31
It's a vacation home, but better.
2:32:34
Uh, and we have some surprise guests, I believe, um, joining in just a second.
2:32:39
We will have them in >> Jessica and Paul. You may know them.
2:32:44
They started a small startup accelerator called Y Cominator. >> Yes, that's right.
2:32:49
And it's Jessica's second time on the show.
2:32:51
We had a fantastic conversation with her the last time she was on the show.
2:32:54
We talked about the get your bag culture and the carpet baggers and just all the cultural uh es and flows of Silicon Valley uh and where we are culturally.
2:33:04
So I'm very excited to bring in Jessica and Paul uh the founders of Y Combinator.
2:33:11
They are living situated living legend.
2:33:13
Oh >> you you're live now. Welcome to the show.
2:33:19
Thank you so much for taking the time to talk to us, >> guys. >> Hi. Good to see you again. >> Thank you.
2:33:24
>> This is so fun with you guys here at Demo Day. >> It is. It is. It's always great.
2:33:27
This is our fourth demo day live stream talking to tons of founders.
2:33:32
Uh it's always fun picking out.
2:33:35
>> I can't wait for the 400th.
2:33:36
>> I got a ways to go, but I'm excited. >> 100 more years. We'll make it.
2:33:40
>> Uh great to have you guys on.
2:33:40
How what's what's uh what's it been like today? >> It's been crowded. It's buzzing.
2:33:45
And by the way, this is our first demo day that we've been to in a few years cuz we're in England and can't manage to come back for it. It is just buzzing.
2:33:55
Um the energy here is just kind of like what I remember in the early days of YC and the investors are all excited to be here. It's magical. I'm on a high. >> Incredible.
2:34:08
>> There's a lot of stuff happening. >> Yeah.
2:34:10
>> Yeah. uh how are you thinking about uh there was this moment a few years ago where I think in tech maybe we were afraid to admit it but it felt like a lot of founders and a lot of entrepreneurs were were sort of grappling with this idea that uh that
2:34:25
open AI might just build every startup and there might be no more ideas and people were a little bit nervous about that of course they went and build companies but it feels like now things have calmed down a little bit and the founders that we talked to are building with more confidence. Have you noticed
2:34:39
Have you noticed anything in the founders that you talked to in an eb and flow of just the confidence with which they view the future right now? >> No.
2:34:48
No, they weren't founders really weren't really worried that OpenAI was going to eat them.
2:34:54
>> I mean, maybe they were in denial for whatever reason, they weren't worried about it.
2:34:57
They're too busy working on their companies.
2:34:59
They're making their thing.
2:35:00
They're trying to get users Open AI eating them in some theoretical future three years from now.
2:35:04
Like, they're not thinking about anything three years from now.
2:35:07
So thinking about that >> we had another we we were talking to Har about this this idea that uh potentially I don't know we're just in a new era where um where it has become easier for a small team of scrappy entrepreneurs to sell to Fortune 500 companies to sell to the government even do you feel like something has materially changed in go to market for YC companies >> well if you're an AI company.
2:35:36
All these big organizations now have some bureaucrat who's been told you're supposed to AI our organization, right?
2:35:47
And he's thinking, damn, I have no idea what to do.
2:35:49
And so some startup shows up and says we'll AI organization.
2:35:52
Like great, come in here, [laughter] right?
2:35:56
Very different from the way it used to be.
2:35:57
I mean, if you show up with other other products for the big company, they'll still tell you to talk to the hand, but um nobody's coming to them with AI things except startups.
2:36:06
So, they have no choice but to talk to startups.
2:36:10
>> What about uh this this tweet that you put out just recently?
2:36:13
Uh we were we were sort of debating it earlier, this idea of of the the the circular economy, selling to other startups.
2:36:21
Uh there are a ton of benefits.
2:36:24
Obviously, uh, startups are very discerning.
2:36:26
If you mess up and don't deliver the product that they're buying from you, you might hear about it publicly, they'll churn. They'll talk to you.
2:36:34
They'll talk to their friends.
2:36:35
Um, but are there any risks from that that you caution entrepreneurs on if they are going to be selling to a lot of startups?
2:36:40
Do they have to message anything differently?
2:36:42
Is there anything that they need to be doing?
2:36:45
>> Well, you have to not suck because startups are discerning.
2:36:47
You can't like have some product and sell it based on a bunch of hype.
2:36:50
Yeah, >> it's got to actually work because they don't have time to mess around with things that don't work and they're very sharp observers of technology.
2:36:56
They're run by the founders themselves usually at that point.
2:37:01
So, you got to actually be good.
2:37:04
>> Uh, I'd love to reflect on how marketing and launching startups has changed over the last few decades.
2:37:10
Uh, Jordy, we we we had Clad Labs on, uh, which we we had a really fun time talking to them, but they they sort of went viral for the wrong reasons.
2:37:20
They uh they they were in their view is the right reason. In their in their Yeah.
2:37:24
In their view is the right reasons.
2:37:25
They were offending people by putting uh gambling in your IDE.
2:37:29
So the software engineer can be gambling while they're coding, I guess. >> Yeah.
2:37:34
And it felt like it felt like this year ra like the concept of using rage bait both at the marketing level and the product level like kind of exploded.
2:37:39
I guess the question is like has has intentionally pissing people off been something that YC founders have have utilized across the eras to get to get attention?
2:37:52
Is it is it really is it really?
2:37:55
>> That's that sort of technique sounds like the technique that would be popular with someone you'd describe as a bit of a scammer.
2:38:04
>> And the thing about these scammers is they don't make the giant companies.
2:38:08
They don't have a long-term focus.
2:38:10
They're not earnestly doing engineering.
2:38:12
They're thinking about what's some gimmick I can use to get ahead, right?
2:38:15
And so long term, they don't matter.
2:38:19
>> You can skip the companies that do random like that because, you know, they're never going to be that big.
2:38:24
>> And of course, I haven't heard of the ex the term rage baiting um either.
2:38:26
Of course, >> it's the Oxford it's the Oxford word of the year.
2:38:33
So, you can go look at their uh definition.
2:38:36
>> It's it it's so interesting.
2:38:36
It's getting attention by making people know what it means, but >> yeah.
2:38:41
And we and and I had written an article and and Gary and I had a nice back and forth uh where uh I basically said like in startups if you in startups you need to build a coalition of people that want you to win.
2:38:53
This is like talent, the media, uh investors, customers.
2:38:58
>> You don't even have to do that actually.
2:39:00
All you have to do is make something really good and find the people who want it.
2:39:05
You don't even need a coalition.
2:39:05
You think like when Facebook was taking off at Harvard, there was some coalition of investors in the media that [laughter] wanted it to take off.
2:39:12
All that mattered was that Zuck had this thing and everybody at Harvard wanted to use it. That's all that matters.
2:39:17
That small intense fire, right?
2:39:18
Or when Apple was getting started and the users were like the people at the homebrew computer club, right?
2:39:25
The media didn't know about that.
2:39:28
>> There was no coalition do a little phenomenon.
2:39:32
>> Zach kind of did a little rage bait.
2:39:32
did rage bait with the hot or not app that definitely enraged a lot of people who didn't want to >> Yeah.
2:39:38
But he didn't do it deliberately. >> No. Exactly. Exactly.
2:39:40
And I was thinking about the Airbnb example like the whole Obama O's and McCain Captain McCain's uh crunch like the cereals uh that they made that was sort of a side quest for them.
2:39:51
But >> that was simply to get attention from the press. >> Interesting.
2:39:57
And make No, actually it was to make money. It >> was to make money.
2:40:00
>> That was [clears throat] before YC.
2:40:00
They didn't have any money.
2:40:02
remember they were dying.
2:40:03
They needed to make money.
2:40:05
They went and got these like off-brand Cheerios and they glued together the boxes themselves to make money.
2:40:10
>> I don't think they knew they were going to make money.
2:40:14
>> We're going to have to consult.
2:40:15
>> I'm pretty sure that was mainly to make money.
2:40:18
>> Uh what's it like being back in San Francisco? [snorts] >> Sunny. >> It's fabulous.
2:40:23
The energy is so great here.
2:40:26
Um, I'm just so h I'm so happy to be back and so happy to be around startups right now.
2:40:32
I'm having a great day if you can't tell.
2:40:35
>> It gets better every time we come back.
2:40:37
Like Daniel Luri is really cleaning up the city.
2:40:40
>> Every time we show up, it's like a little better. >> That's great.
2:40:43
>> I was asking like how far back have we gone?
2:40:45
Have we gone all the way back to when Edley died? Not yet.
2:40:47
We're like, >> but we've turned the clock back to maybe two years into London Breed. >> Oh, that's good. Okay. Yeah, that's great.
2:40:56
great. Um I have one uh I yeah I have one more I want to I want to um think through this concept that's been sort of lightly banded about in the startup discussion you know ecosystem this idea of the deals guy era that you can
2:41:14
actually build a business now by being more of the business person the more of the deals guy and less of the of what I remember about the the Y combinator promise which was uh just the earnest hacker, the earnest hacker, the earnest hacker. And it feels like there's a lot
2:41:28
And it feels like there's a lot of people that are saying, "Yeah, but there's actually a way to go and get this person just marshall the capital and, you know, do something that's just been forgotten, not necessarily discover something new."
2:41:40
And I was wondering if you have any any reactions to this this idea that that increasingly there are entrepreneurs that sort of get really big.
2:41:50
Who knows if they win, but they but they seem to win on the back of just raw dealmaking talent as opposed to raw engineering leadership.
2:42:01
>> Maybe in enterprise more, >> you know, um the like enterprise you like >> sell crap to CTO instead of selling good stuff to programmers.
2:42:12
>> Um >> so salesmanship has always mattered more in enterprise. >> Yeah.
2:42:16
I have one observation from this morning's session of demo day >> there.
2:42:22
Everyone most everyone that presented this morning is an earnest hacker.
2:42:26
I said to the person next to me, they're all nerds this time like [laughter] 100%. And I love it. >> Yeah.
2:42:33
You know, if anything, YC drifted too far away from funding earnest hackers. >> Interesting.
2:42:39
>> And so YC for the last few years has been focusing more on like getting back to the essentials, back to the roots.
2:42:45
And so if anything, I would say YC batches are more like a higher percentage earnest hackers now. >> Yeah.
2:42:53
>> You know, honestly, I would still bet on earnest hackers. >> Yeah. Yeah. I I I agree with you.
2:42:55
Do you think that that's uh that that is what the essential skill set of YC leadership needs to be?
2:43:02
Because uh I don't want to discredit all the hard work you did in the early days, but you didn't have to fight the fact that there were people out there writing blog posts of how to reverse engineer and make it look like you're an earnest hacker when in fact you are the, you know, the the carpet bagger.
2:43:17
And now it's there there's a whole industrial complex for how to fake your way and make it appear that you're an earnest hacker when in fact you're not.
2:43:25
If the YC partners are themselves hackers, you can sniff out a faker like that, it's not even a problem. >> Yeah. Yeah. Yeah.
2:43:33
But that seems like the main the main way that uh YC creates value these days.
2:43:37
We'll just be continuing to to hold that line essentially.
2:43:42
>> What do you think is your most >> Yeah.
2:43:43
I think I think you know here's something that will reassure you if you think okay is the earnest hacker thing was that did that just work for a while and now maybe it's over?
2:43:51
Isaac Newton was an earnest hacker, right?
2:43:53
[laughter] It's way older than startups.
2:43:57
>> Yeah, >> this is this is what wins.
2:44:00
>> Uh what do you think is your most underappreciated essay >> because a lot of them are sufficiently appreciated.
2:44:06
[laughter] >> The thing is I don't know how much people appreciate them.
2:44:10
I don't know how much people appreciate different ones. So it's hard to say.
2:44:17
>> Um how to do great work is pretty good, >> but I think people like that one. >> Yeah. >> Right.
2:44:22
I read Life is Short at least once a year, but people like that one, too. >> Yeah. >> I don't know. I don't know. That's a weird question.
2:44:30
>> You'd have to You probably have to look at inverse page views.
2:44:32
Which gets the least page views historically over the past year.
2:44:36
Let's say >> if I was looking at a list of page views, I could tell you. >> Okay.
2:44:39
Well, maybe breaking news.
2:44:42
Um, yeah, that's uh that's very funny.
2:44:46
Um, do you have anything else to bring? What? What else?
2:44:50
Uh Paul, are we in a bubble? >> No. No.
2:44:55
Everybody is always saying we're in a bubble, you know.
2:44:57
Um like every year people say we're in a bubble.
2:44:59
Every year people say like the valuations at Demo Die, they're too high now, right?
2:45:06
I mean they [laughter] were saying this back in like 2010 when the valuations were like $4 million.
2:45:13
>> Um and now they're like what 30 or something typically.
2:45:16
So, um, people are always saying stuff like that and I don't No, I don't think so. I think I'll tell you.
2:45:25
I think like AI is very highly priced, >> but it might not be overpriced.
2:45:30
That's the interesting thing.
2:45:32
Is it is it as big a deal as prices seem to suggest?
2:45:35
It could be, maybe even bigger. It's definitely real. It's not hype. The AI is real. Mhm.
2:45:45
>> Are foundation models good at writing lisp?
2:45:49
>> You know, I've never I think they would be good at writing lisp. Yes. Yes.
2:45:53
Because they're good at writing things that have a lot of um a lot of a lot of training data out there, right?
2:45:57
And there's a lot of lisp source code.
2:45:59
So, I think they'd be fine at writing lisp.
2:46:03
>> How are you using AI in your life?
2:46:06
>> I just use it like ordinary people do.
2:46:06
I ask I ask you questions. >> Sure. Sure. Very boring answer.
2:46:14
[laughter] >> It's a good answer.
2:46:16
It's not like, oh, I string I I'm training my own model to do a better Google search. >> What? Uh, >> no, no, no.
2:46:22
I haven't actually written anything using AI. >> Yeah. >> You know, I feel bad.
2:46:26
I really should write an LLM because you can't really understand this stuff unless you've written one.
2:46:31
I should write an LLM, but I haven't done it. >> Yeah.
2:46:33
Didn't Carpathy publish a whole uh stat first principles type of thing and it seemed really fun to teach himself, you know. That's why he did it.
2:46:42
>> Well, he has a new company that's education technology company and I believe that the the main course will be teaching yourself to build an LLM, teaching yourself to build a chatbot effectively, which would be very >> that's what I tell high school kids.
2:46:52
I get all these emails from high school kids say I'm working on a startup, >> you know, to introduce like founders to VCs or some crap like that.
2:46:59
And I say don't start a startup. Get good at technology. Write an LLM. Yeah.
2:47:06
>> Then you can start a startup.
2:47:06
Do you do you think uh reflecting on the history of YC, do you think it's it's fair to uh try and create a concept of eras around um like what the key insight was?
2:47:18
I I remember a lot of people saying like one of the first key insights was just this idea that you you could take someone fresh out of college and actually give them money and they could go and build a business.
2:47:31
They didn't need $10 million.
2:47:33
they didn't need 10 years of experience in the enterprise.
2:47:36
Um and then >> or an MBA >> or an MBA and then maybe the second era was thinking that maybe the same rules applied internationally and that was like a second wave of of entrepreneurial energy that was unlocked by the YC.
2:47:47
We had international I mean >> we always had international >> we understand countries aren't all that [laughter] >> but but do do you think there are any other like underappreciated aspects of like the YC strategy or or is it really just as simple as uh you know >> well there were things we didn't appreciate in the beginning. >> Yes.
2:48:07
>> So for example we didn't understand that as a byproduct of funding all these companies we would create this alumni network.
2:48:14
We had no idea but the alumni network is enormously important. It's out there now.
2:48:18
All these alumni are investors. >> Yes.
2:48:21
It's staggering how many are investors now, actually. >> Yeah. >> It's amazing.
2:48:26
>> It's like taking over Silicon Valley.
2:48:28
And we never had any idea that was going to happen. >> Okay.
2:48:30
On the alumni network, uh do is it fair to uh to characterize YC as a bit of a union against venture capitalists?
2:48:41
Yeah, it's a lot like a union >> because you mess because Yeah, because if you attack one individual, one one founder, if you fire the founder after investing, you you get you get board control from them and you oust them, uh that might make it sway into the rest of the YC community and it overall raises the level of founder friendliness. Is that correct? You know what though?
2:49:00
It's not simply one-sided because if founders screw over investors, if they like do a handshake deal and then and then refuse to go through with it, we would tell them not to do that, too.
2:49:13
>> We want everybody to like play by the rules. >> Yeah.
2:49:16
And behave well >> because the big wins don't come from breaking the rules.
2:49:20
The big wins don't come from little cheats that get you 2x multiples in a world of like thousandx returns, right?
2:49:26
It's for the same reason like people in Silicon Valley don't focus a lot on tax evasion >> because what's tax evasion going to get you like 2x returns in a world where getting the right startups will get you a thousandx returns. Yeah.
2:49:37
you know, >> uh do you think that uh the process of founding a company, raising money is is at its uh the end of history in terms of uh efficiency like the safe is the most efficient document we will ever have or do we need to speed things up even further?
2:49:59
>> Well, C Levy, Carolyn Levy invented the [clears throat] safe and she also invented the convertible note that everybody used before it. Yeah.
2:50:07
>> So, she has twice rewritten the rules.
2:50:10
She has twice recreated the chess board that the game is played on.
2:50:12
Um, if she thought there was a better thing than the safe, she probably would have created >> Maybe she has a third one in her.
2:50:19
>> We should ask very popular. >> Oh, yeah. Okay. Yeah.
2:50:23
And you could ask her that, John, when you come on our podcast. >> We'd love to. I'd love to. I I [laughter] I can't.
2:50:31
>> Is there anything wrong with the safe?
2:50:32
And if And like if there is, why hasn't she fixed it already? Yeah. >> You know? >> Yeah.
2:50:36
So probably not because C Lev is not Slack.
2:50:38
If there was anything missing, she would have she would have like made a new version. >> Yeah.
2:50:42
I mean, from my perspective, it seems like it's worked.
2:50:44
>> What What problem in the world did you think a YC startup would have uh uh fixed by now?
2:50:48
Think like uh housing affordability or any of these sort of major, >> you know, we don't have any grand strategic vision for what the startups do because the founders know that, not us, right?
2:51:01
That would be like asking a publisher what what novel do you think you know would you have expected someone to write good publishers they just like they let the they let the novelists write the novels.
2:51:13
So we would just we just try and find good people. What do they do?
2:51:17
Whatever these good people are interested in.
2:51:19
Anything any preconceptions we had about what they should do would just be adding noise to that.
2:51:26
>> How do you think about coaching folks through pivots?
2:51:28
It feels like we're in an era where uh there's a lot of companies that are still finding product market fit.
2:51:34
Pivots are probably just as common as they always have been, but everyone has an order of magnitude more money, >> if anything more common. >> Yeah. >> Yeah.
2:51:42
I think that it's more common.
2:51:44
You're talk you talk about new ideas with startups all the time in your office hours.
2:51:48
>> This is one of my specialties. Yeah.
2:51:48
Um when people are just dead in the water and they need to get a new idea, they often get sent to talk to me and we cook up something.
2:51:57
Um, >> has the advice changed if someone comes in and says, "Hey, I have uh $200,000 raised and I have me and my co-founder are living in a apartment together and we need to pivot versus I come in and I say, "Hey, look, I got 5 million bucks and I got 20 employees already or something like that." >> 20 employees. >> I don't know. It's happening, right?
2:52:18
It's you you do see this, right? >> Well, no.
2:52:21
Usually have 20 employees.
2:52:21
Um, usually usually I mean that would be that would be alarming.
2:52:29
That would be very alarming because there's so many companies >> those 20 employees constrain the idea you're going to have.
2:52:35
If you just have the founders, you could do anything.
2:52:36
If you already have 20 people, you either have to fire them or do something that those 20 people can do, right? >> Yeah. Yeah.
2:52:43
>> Um, which really [clears throat] constrains your options. >> Yeah.
2:52:46
>> So, it's the problem with the 20 employees is not the cost.
2:52:47
It's that they change what they limit what you can think of, you know, >> which is why you shouldn't hire. Just don't hire. Just don't hire.
2:52:58
>> What kind of guidance do you give to founders around that are feeling a pressure to go from zero to 100 million in ARR in like three years or whatever like the new gold standard is?
2:53:08
>> What I tell startups over and over and over is all that matters is growth rate, not the absolute numbers.
2:53:13
because mathematically you'll see if you try simulating it if if your growth rate is high enough doesn't matter what the absolute numbers are you'll get [laughter] there. >> Yeah.
2:53:23
>> You know and so you just get a really good growth rate.
2:53:25
And so the great thing about focusing on growth rate means you can like focus on startups.
2:53:29
You can sell stuff to startups for cheap instead of having to go and do these big deals with big companies that take a long time and make your product stupider, right?
2:53:36
you can sell things to these quick quick deciding early adopters and then you just get more and more of them and your your company grows by several percent a week eventually it's going to be huge.
2:53:50
>> Are you still recommending to folks who ask for advice for kids uh that they should learn to code? >> Yeah. Oh yeah. Yeah. Yeah.
2:53:57
I still tell people that or at least learn technology.
2:54:01
It doesn't have to be coding specifically.
2:54:03
You could learn how to make rockets >> um or drones or work with lasers or gene editing or something like that, but you should do the stuff and not just like play house pretending to start fake startups in some business plan competition, you know?
2:54:22
>> I tell everyone who says they might want to start a startup someday to learn to code because it's the most important thing you could do that and save your money. Mhm. Yeah.
2:54:30
No, that that's really good advice.
2:54:34
>> And no one likes to hear that, by the way, but I tell them anyway. >> Yeah.
2:54:37
No, no, >> you give a lot of advice.
2:54:38
People come and they like want advice.
2:54:40
It's like if you went to the doctor and you said, "Doctor, what can I do to be healthier?"
2:54:44
And the doctor says, "Eat less, and exercise more."
2:54:46
And you're like, "Oh, I was hoping you'd say something else, right?"
2:54:48
Well, that's what it's like when they come to me.
2:54:51
>> They come to me for advice.
2:54:51
And I say the startup equivalent of, "Eat less, exercise more."
2:54:55
And they're like, "Oh, isn't there some trick I could use to be get virality?
2:54:59
Could I couldn't I get firality instead?
2:55:03
>> Just like do the startup equivalent of eat less and get more exercise, which is build stuff and talk to users, understand your users, and be good at building. That's the recipe.
2:55:11
It was in 2005, and it's just as much the recipe now. >> Yeah.
2:55:17
>> How many startups do you think >> How many startups do you think YC will have uh per batch uh a decade from now?
2:55:27
M >> because I think in a perfect world we have a lot more earnest hackers >> and uh they can apply to YC and and if they meet them I know I know you're not setting targets and there's not like a specific you know acceptance rate that you're trying to track but uh we we feel
2:55:47
like YC is one of the most important institutions in the world and ideally it can be bigger but but maybe there's some >> no no they will be bigger they they will inevitably bigger because there's this secular trend of more people starting startups. >> Yeah. Do you think we're early? Do do >> Yeah.
2:56:01
Do you think we're early?
2:56:01
Do do you think we're we're early in this in this trend?
2:56:04
I mean, it feels like there's so much so much it's now you can create a startup, you can create a CC Corp in a few minutes, right?
2:56:10
It's like there's all this sort of like underlying uh infrastructure that's been built that is reducing friction to starting companies.
2:56:19
is you can ask chatbt, "How do I start a business?"
2:56:20
And it'll give you a good playbook.
2:56:22
And that maybe helps somebody that uh hasn't found the YC blog yet uh figure out how to get going.
2:56:30
>> We're the training data, even if they don't know it. >> Yeah.
2:56:34
>> So, will more people startups? Yes.
2:56:34
If you talk to like ambitious 15y olds, they all want to go startups.
2:56:40
Nobody wants to go work for some company and work their way up the corporate ladder anymore.
2:56:46
whole idea sounds so like sounds so like 1980s.
2:56:48
Um, and there's a lot of earnest hackers.
2:56:52
The limit and the limit you think like what's the limit.
2:56:56
So the limit is what people want, right? That's what startups do.
2:56:59
They make something people want.
2:57:00
What are what are people's people's wants? They're limitless.
2:57:04
[laughter] Not literally limitless because eventually you run out of atoms in the universe.
2:57:10
But for all practical purposes, in the near term, people's people's wants are infinite.
2:57:14
And so there's infinite demand for good stuff you could make. >> Yeah.
2:57:20
Well, that's a great place to end it.
2:57:22
We have to catch a flight.
2:57:22
Thank you so much for taking the time to talk. >> Yeah.
2:57:25
Thank you for everything uh you guys have done for the industry and the and the world uh through YC. It's an honor.
2:57:31
It's an honor to >> cover every batch and uh it's been great having you guys on. >> Yeah. Nice to meet you. >> Thanks for having us. I love you guys. >> Yes, we love you too. Thank you so much. >> Have fun in SF.
2:57:43
>> Have a great rest of your trip. We'll talk to you soon. >> Goodbye.
2:57:48
>> We have to hop on flight. But you hear that, John? >> I hear it.
2:57:51
Yes, I hear the goat noise, the sound cue.
2:57:54
That one's a little bit subtle.
2:57:55
I think that there's a lot of people that might not pick up on why they're hearing this random goat noise that low in the back.
2:58:00
But if you know, you know.
2:58:02
And also, if you want exceptional sleep without exception, you go to eight. com. You fall asleep faster. You sleep deeper. You wake up energized. And we should close out.
2:58:11
There's a lot of stuff going on, dealbook summits going on.
2:58:16
There are debates raging on the timeline, but we will have to cover them tomorrow.
2:58:20
Uh we will close out with a congratulations to Eden, uh the co-host of the Prof Markets podcast.
2:58:29
Uh, I love his bio because he says he's not Prof's son, even though they look somewhat similar.
2:58:35
Post yesterday because he got into Forbes 30 under 30 and he said, "I'll see you guys in prison."
2:58:42
>> He said, "Woke up to learn I made Forbes 30 under 30.
2:58:45
Congrats to the other winner winners.
2:58:48
>> Can we play this before we can we before we jump?
2:58:50
Can we uh can we play this?"
2:58:53
Oh, Gary Tan's in the chat. >> Ali's in the chat.
2:58:56
>> Gary, we hope you feel better. >> Hannah's in the chat.
2:58:58
Gary feel better the show.
2:59:00
Thank you so much for making this happen.
2:59:02
Uh we're very sorry we couldn't be there in person, but we had a blast.
2:59:05
We went on a whirlwind tour.
2:59:05
We talked to tons of uh YC founders and uh Y the state of YC is healthier than ever, stronger than ever.
2:59:14
>> He's got a Gary's got elementary school or preschool.
2:59:18
>> It's [laughter] so rough. I've been there, man. I've been there. It's virus. Yeah.
2:59:23
Well, uh we hope you get well soon.
2:59:25
uh team, we need to definitely send some soup or some flowers to Gary Tan as soon as possible and uh and we will see you all tomorrow. Thank you.
2:59:34
>> Thank you for tuning in.
2:59:34
Thank you to Y Combinator for hosting us and all the founders.
2:59:37
It was a it was a it was a whirlwind tour and I'm very excited about a lot of these companies.
2:59:42
>> Yes, we will talk to you later. Chair.