0:00
You calling this a toy suggests to me that you do not want this to be a companion.
Search this video
0:04
>> I mean, it's definitely a companion in the same way that a cat is a companion.
0:08
Like as cool as it may look and as much cloud as it may get you, nobody really wants a Terminator robot.
0:15
>> Of all the things Silicon Valley could build with AI, Daniel Coons went the opposite direction.
0:19
No apps, no agents, no bicycle for the mind, just toy with some impressive tech built in.
0:25
Today on Access, we're diving into Starboy, the little alien that hangs off your bag, purrs like a cat, and might just break through to the elusive world of fashion and culture.
0:36
But first, Alex and I discuss his visit to Anthropic and my very failed image generation experiments with Claude. Welcome to Access.
0:50
Welcome to Access, or as Alex says, it's Access time.
0:54
>> I'm going to stop saying that.
0:55
>> has has debatable value, but it's it's actually growing on me. Alex, how's it going? >> I'm good, man. How are you?
1:02
>> Well, I sound better than yesterday.
1:05
I'm using this opportunity to serve you with notice that I may not be able to make drinks tonight because I'm almost positive that I have the human metapneumovirus. >> What is that?
1:16
>> Which is, according to the LA Times, not actually something new, but I just love how scary it sounds.
1:21
It's just like a new virus that all the children give to their parents for fun that is kind of like an infernal combo of like bronchitis and epic sinus headache.
1:33
>> You sound totally fine.
1:34
>> Well, I am a podcast host who can take a lot of pills and get over this stuff for the purpose of the next 15 minutes if need be.
1:43
>> Okay, but you're actually sick, so yeah, we will cancel drinks tonight. >> Yeah, we'll see.
1:46
Or you could just kind of shake my elbow if you really want to.
1:52
But um I've got I've got Japanese remedies over here.
1:54
I've got guaifenesin, which is way cheaper than Mucinex.
1:58
I've got Sudafed, which is the only thing that works with sinus headaches.
2:01
And the sponsor of this show, Simply Saline, of which I have bought 100,000 cans. I cannot, dude.
2:09
I cannot believe I pay $10 for a freaking can of salt water, but it works.
2:15
>> We need to get Vox Media on Simply Saline and Zbiotics for the show at this point.
2:21
>> Well, especially I think in the wake of COVID, where we've discovered just how disgusting the air is around us.
2:26
Um there was actually some good research, uh bullish long Simply Saline, long Arm & Hammer.
2:31
Uh there was some research the other day actually that that saline um spray up your nose actually like literally washes away viruses that get into your nose on the plane or otherwise.
2:43
>> Oh, well, I need to do that cuz I was on a plane a lot yesterday. >> Oh, yeah, dude. No, it's the best.
2:46
And I'm proud to say that I was the last member of my family to get it.
2:50
And I am getting it a little milder.
2:52
Um yeah, can we please get AI on this?
2:56
>> Well, Fiji's working on ChatGPT health, remember? So. >> Yeah.
3:00
Well, we'll see about that.
3:02
Um what else is on your mind, man?
3:04
>> I spent yesterday uh up in San Francisco, as I'm one to do, the Burbank day trip.
3:08
There for a few random things, but I think the most relevant thing to this show is that I spent a good chunk of the day at Anthropic headquarters, which was interesting in many ways.
3:18
This visit was set up before all of the Pentagon Department of War controversy.
3:27
Uh they had set me up with a bunch of product leaders from Claude Code, um and then people like on the economics side, on policy side, just to kind of get a lay of the land and meet people and talk to them about the product direction.
3:41
But I will say I got way more excited to go after the events of the last week and a half or so and see, I mean, it's so rare to get to see inside one of these companies, especially one of the AI labs, um uh at all, but to see inside at a time like this, you know, when a company's going through so much external perceived turmoil, uh was really interesting.
4:05
>> Did you take the front entrance to the building this time or your usual like cut the cut the barbed wire back door access?
4:11
>> No, uh front entrance, but it's an unmarked building as these AI labs are, um with a lot of security.
4:16
I think they have probably some crazies that show up.
4:20
It was gone by the time I got there, but last week they had a bunch of sidewalk chalk out uh in front of the office uh basically saying like, you know, we support you and standing up for your values and all of that.
4:30
And when you get inside, it looks like a normal a normal building.
4:35
It's the ex uh Slack building.
4:38
So, it's about 10 stories near Salesforce Tower.
4:41
And uh then you get inside at the elevator and it's like uh you're entering, I don't know, one of the most like pristine, intentional, most beautifully designed office spaces.
4:55
Uh and uh you know, there's lamb chops for lunch, and the office was absolutely buzzing, packed, and it was like there's a duality to it, right?
5:03
Because they've got all these business troubles because of what's happening with the government and the lawsuit, and they're suing Trump and all of that.
5:11
But then Claude downloads are going bananas, right?
5:15
Claude was number one in the App Store for the first time ever.
5:17
It shot from >> It shot from I think below 20 to number one in like a day.
5:24
Um and so to get to like meet with the head of platform uh while they are literally trying to figure out how to scale up infrastructure, uh and they're going through this moment that I'm sure you probably felt at Snap back in the day when Snap was kind of going parabolic.
5:38
Um it to get to see that up close was was pretty interesting.
5:44
>> Also, the part that reminded me of the Snap days was when you were talking about sidewalk chaos good or bad and the security team.
5:51
I don't know if you ever heard this, but the Snapchat security folks were kind of polarizing in the early days cuz they all wore black cowboy hats.
6:01
>> I do not remember this.
6:01
I remember going down to the Venice Boardwalk, but not seeing cowboy hats.
6:05
>> Yeah, I don't know why that was the thing, but that was the thing.
6:07
And so, of course, all the people who hang out every day and, you know, hawk their wares on the Venice Boardwalk knew that the black cowboy hats were snap security and were, you know, an invasive species and all this stuff, but um uh I want to return briefly to the lamb chops.
6:26
I'm surprised that there isn't like some narrative about all the meat lovers and the Hubermans and such being like, "In order to keep up with AI, we got to feed our brains with meat, the highest nutrient density of any uh object that you can eat."
6:42
You know what I'm saying?
6:44
I mean, I guess it's mostly about protein.
6:46
>> There were a lot of beautifully grilled vegetables and other things, farro, whatever your heart wanted.
6:52
>> Cuz the protein is all about saving time, which I think makes sense for engineers.
6:55
You know, they're trying to like drink Soylent so that they can like code for 8 more minutes.
6:59
But I think the whole meat narrative, where's the narrative?
7:05
And I mean, you could say this is coming from like big hamburger.
7:06
I don't think I have any shares.
7:09
I don't think I have any shares in any of the the meat producers, but >> It's almost a daily occurrence where people ask me if that's really your last name.
7:17
It is just really cool to see the culture of a of a company that is just at the center of everything in the in a very kind of inflection type moment.
7:24
It doesn't happen very often, even for me, where I'm, you know, very lucky I get to go into these spaces a lot.
7:32
You know, you really do get a feel for the place in a way you can't by just reading the tweets and reading the articles and watching the interviews when you're there.
7:40
I mean, last time I was at Anthropic, it was August 2025.
7:45
And I randomly was in the elevator.
7:45
This is my favorite thing to go to like a tech campus and randomly get in the elevator with someone.
7:50
Um and it was like their head of alignment, who's kind of a famous figure um in the AI community, Jan.
7:58
And he was wearing a I think it was a unicorn onesie. Um >> so?
8:04
>> unicorns on it and um and a backpack and like a pajama onesie.
8:10
I was just like, this is this is perfect.
8:11
This is this really kind of just tells you everything.
8:14
>> Was there an occasion or not really?
8:16
>> I think that's the kind of stuff he wears every day, is what I've heard.
8:19
>> Well, was his horn aligned?
8:22
Or was it off to one side? >> No comment.
8:24
Uh but being there yesterday, um I wanted to know like, yeah, how what is it like going through this moment?
8:31
Um and it does feel like the employees are really galvanized around the values that the company is standing up for.
8:40
And I think there's a lot of nuance to it.
8:42
I think um like for example, with like autonomous weapons, I think they are I think Dario will actually be fine with autonomous weapons when the models are good enough.
8:52
I don't think it was a moral issue with AI controlling weapons.
8:56
I think it was a these models are not ready to do this and we do not want to be the ones having to clean up the mess of the government um using bad models.
9:06
>> As soon as we can make Skynet a little more thoughtful and for problem solvers, then uh we're ready.
9:13
>> So, that one's less of a moral one, but I do think the more moral one was the surveillance problem that they feel, which is um AI will make it easier for governments to to mass surveil their citizens in a way that they, you know, it was never possible before.
9:27
And they don't feel like Congress uh has really internalized this and they don't think there's enough laws and protections around this.
9:33
And so, they exercise their right as a private company to not want to be involved in that, and even the the potential of that.
9:43
>> Wasn't this a pro- hasn't this been a problem since the beginning of AI that AI stereotypes things?
9:46
And that does not sound good for autonomous weapons.
9:50
Uh when you say point at the bad guy, and it's like, "Well, that looks like the bad guy in Die Hard, so uh I'm going to shoot him now." >> Right. Right.
10:00
Or like it hallucinates something. Yeah.
10:03
Um but the the surveillance thing was scary, you know, the way it was described to me, you know, like in China, if you send an image to someone over, you know, the popular social media apps, you know, if it's something that the government wants to censor, it just won't show up on the other person's device, right?
10:21
But the government can't do that to text because there's so much text that gets sent over the internet at every, you know, every minute.
10:30
And with AI, with Claude, it's actually possible to observe and intercept and censor at that scale. And um I don't know.
10:38
That was just interesting, and I appreciate that they're thinking hard about that, and that they're standing up for that.
10:43
And at the same time, man, like they're they're now like going, "Oh, I guess we're like a hit consumer company."
10:49
You know, like they were not prepared for Claude downloads to go crazy in response to this.
10:56
And uh there were a lot of war rooms and late nights uh getting the the servers and the GPUs lit up.
11:01
Um and now I think they're rethinking a lot about, you know, how much are we focused on consumer, like what opportunity do we have now with the growth continuing to just go like as- you know, skyrocket.
11:12
Uh it's a really interesting interesting time for them.
11:17
>> Well, I'll tell you one area where Claude is a little bit behind is on the image generation.
11:21
Uh >> This was a good one.
11:25
>> what I sent you yesterday.
11:25
I was trying to generate like a propaganda poster that said robots rule.
11:32
And chat GPT gave me literally exactly what I wanted.
11:36
This kind of blonde guy looking like he's from Fallout winking and giving the thumbs up and the robot in the background.
11:42
And I gave the exact same prompt to Claude and it literally generated a picture of like Mr.
11:48
Magoo out of like square shapes and lines.
11:51
Uh which is genuinely impressive in its own right, but this is like one of the most grotesque images I've ever seen in my life.
11:59
And I tweeted it and someone was like he's going to seek revenge.
12:01
And then some other reply guys uh informed me that Claude is actually not trying to be for this quite yet, but like >> Yeah.
12:11
>> I I don't know that, you know?
12:13
Uh I figure I figure they all do the same thing. >> No, no, not at all.
12:16
I talked to this about this yesterday on your behalf actually.
12:20
Um they are not focused on images on multimodal because they don't think that it's ready and good enough and that it's also >> request to somebody else.
12:31
>> Yeah, I I mean I think they don't want to do that.
12:33
That's kind of like uh Mac and Windows working together or something.
12:37
I don't think they're going to do that.
12:38
They're so singularly focused on getting to AGI and they think coding is the way they build the AI researcher who then basically makes everything so efficient and better that they can then get to AGI, build better multimodal uh and build world models and all this stuff.
12:55
Uh but they're just they have a much different focus whereas like Open >> There's also just the argument that like prompting is just I mean I've had a handful of clients in this world.
13:03
I think there's a strong argument that for most people prompting is never going to be the the best way to a great picture and there going to be tools for that.
13:12
And so maybe part of them is also saying that.
13:14
>> They think the visual aspect of AI is like pre-Napster and it's just not ready.
13:17
Uh and you can see that and like even Nano Banana which is good still gets a lot wrong.
13:22
It still is hard to kind of to to edit.
13:24
Um and so they're just not focused on it.
13:27
And I mean, they have all they have enough problems on their hands, which is like they can't keep the servers up, apparently.
13:34
Um, >> Yeah, there was like a little Claude outage the other day and uh >> Yeah.
13:38
>> my stomach dropped a little bit.
13:38
I was like, "What what am I going to do?"
13:42
>> Yeah, and I also uh >> used to go down and you're like, "I guess I'll go home."
13:46
>> I also saw the infamous vending machine Claudius. >> Oh.
13:50
>> Have you heard about this?
13:51
>> Yes, from the famous uh was it a New Yorker article?
13:54
>> There was that, there was Jo Jo our friend Joanna Stern did a video of it that was really good.
13:57
So, basically, they've had Claude, a version of it called Claudius, running a vending machine inside the office for the last like year with a budget and multiple agents that are acting as different parts of the company.
14:09
So, there's like a CEO agent that like checks in with the assistant agent.
14:12
And the employees interact with it via Slack and are basically constantly trying to game it and get it to do crazy things and agree to purchase crazy things to put in the vending machine.
14:23
So, they were telling me that it almost ordered a giant gold bar um, that they's they were then going to have to figure out how to like carry into the vending machine.
14:33
It would actually have just broken the vending machine cuz it weighed so much.
14:37
Um, it tried to uh >> I just like that Claudius is like, "Dear at security team, I'll meet you guys out front and I'll be wearing a blue shirt."
14:45
>> Literally No, literally.
14:45
And it like one of them uh almost got it to uh seriously reach out to a bunch of uh uh a bunch of places to try to get a Birkin bag um, and when I was there, there was a switch in it and some ketchup and a bunch of random stuff.
15:02
Um, and these vending machines are actually um going to be a product.
15:04
So, there's one at there's one in some of the other AI labs.
15:08
XAI has one, Google DeepMind has one.
15:12
>> Dude, this is not good. This is not good.
15:14
They're going to have cameras on them and they're going to say, "You look like this kind of person and so you have gotten dynamic pricing for your stickers bar, which is now $9.
15:24
99 because we have detected that that's the most you'll be willing to pay for it.
15:29
>> need to have the CEO of the company that is doing this vending machine roll out and in labs on the show.
15:33
I'm going to reach out because yeah, they're they're doing it and they're they're going to scale it.
15:38
But yeah, ran into Mike Krieger who's doing a lot of interesting stuff on the labs team there.
15:43
And yeah, man, just generally I think the thing that stood out to me in all of the conversations was Anthropic really emphasizes its culture in a way that I think is unique in AI.
15:58
And people kept telling me about this culture interview process they have where they screen, you know, new employees for their values and how they think about building AI.
16:10
And it sounds very unique and they ask questions that are like what are what is like a hard sacrifice you've had to make for your morals or would you be okay with Anthropic going to zero if we had to make a hard decision that was better for humanity?
16:26
>> Would you jump off a bridge if Dario did?
16:29
>> But it does seem like an you know, I think a lot of companies talk about that.
16:32
They talk about how unique the recruiting process is and you know, the values and how they weed out employees.
16:37
But I do think it's unusual at Anthropic.
16:41
Everyone was talking about how intense the culture interview was and I hadn't really heard that before. >> Absolutely.
16:47
And I think this is the whole value of you know, the mission, the vision, the manifestos these companies write is not just to create internal alignment to help you navigate, you know, downturns whether it's in the market or with your brand.
17:03
But to help people understand like what kind of folks we want to hire.
17:05
I think after all this time I'm a far bigger believer in just kind of like people are naturally going to hire people like themselves and that starts with the CEO.
17:15
Like the CEO and their behavior literally is the culture.
17:17
They hire people like them.
17:19
Those people hire people like them.
17:20
That's really how you do it.
17:22
There's not actually a way to kind of programmatically create a different culture from the founder, I would say.
17:27
But all these different tools, and of course your public decisions, uh go a long way.
17:31
And I think we saw some tweets the other day that people are turning down offers at OpenAI because of uh or or uh not even taking interviews with them if they're at Anthropic just cuz they're super super pumped.
17:45
And I mean there's, you know, obviously there's a money obsession in our industry, but I believe that that the most intelligent and experienced uh among us who often in tech already have a lot of money are are more motivated by by the mission than what what we're here to do than anything else, yeah. The researchers, right? >> Right.
18:05
And I was talking to one of the researchers, and I hadn't really heard it this way, but he was like, "Yeah, I mean, a year ago we were kind of on a knife's edge, like uh OpenAI still does have way more users than us.
18:14
We're worth like, to this day, we're worth less than half of what they are.
18:17
Uh we don't have the distribution.
18:20
We don't have hundreds of millions of users.
18:23
Um if Gemini had been really good like 2 years ago, we may have been killed in the cradle, right?
18:27
We had the benefit of being so focused on the models being really good, uh and really singularly focusing on that when everyone else got distracted by big consumer things, that now they have that edge, and it's leading to consumer wins.
18:40
Uh you know, with Claude Code and Claude Co-work, and now they've got this network effect.
18:44
So, they're in a really interesting spot, and you know, like I I think they have like eight co-founders, and they're all still there, which is super unique.
18:52
I mean, OpenAI alone has spawned like six other neo labs from like founders who have have left.
18:58
Um and >> Coming up next, octo founders from Bravo TV.
19:05
>> Um and yeah, I don't know.
19:05
Just a really interesting uh unique place uh that I I lucky to get to see it moment like this.
19:12
>> Well, Alex, I'm glad you had a good time.
19:16
And now, we will cut to our interview with Daniel, the creator of Starboy.
19:26
Daniel, welcome to the show. How the heck are you? >> I'm great. How are you?
19:31
>> For those who can't see, Daniel appears to be in a hot tub wearing some ear cans on a balcony, I believe, in Lake Tahoe. Is that correct? >> Yeah, that's right.
19:41
The whole background is AI generated, actually. I'm inside right now. >> Looks believable.
19:48
>> On the morning of your launch, is this the optimal place to launch a new tech product these days?
19:51
I hear the birds chirping now. It's lovely. >> I think it is.
19:55
Yeah, there's a lot of uh little guy inspiration around here.
19:59
A lot of little forest creatures.
20:02
>> I guess that makes perfect sense.
20:03
>> I was just in Tahoe last weekend.
20:03
It's beautiful there right now. So, yeah.
20:07
>> It's really beautiful.
20:07
And thank god the snow is gone.
20:08
I was going to come a few weeks ago, but then we saw the 9 in or the 9 ft or 90 ft or whatever it was, and I was in New York at the time, and I was like, "Damn, I don't do that right now, but it's beautiful now."
20:24
>> One thing that strikes me is that this is a very different environment than a war room where most launches take place.
20:31
I mean, I remember when we launched the Snapchat Spectacles gadget a long time ago, it was just the most chaotic room ever, and Evan himself was actually touching up the photos in Photoshop on someone's computer until the very last second, and your current situation could not be more opposite, I will say.
20:47
Are you bubbling with stress and intensity inside?
20:52
>> I mean, I'm feeling pretty good right now.
20:54
Uh it's kind of more war room vibes inside right now.
20:58
Uh we were touching up some stuff right before the launch, but um you know, we've done a lot of crap, and we've really we've really thought through it.
21:05
So, um I think it went about as well as it could have.
21:11
>> You've got the team in there grinding while you're in the hot tub. I like that. Founder mode.
21:15
>> Yeah, you're on his in there, my creative director.
21:18
Um most other people are remote right now.
21:20
So, industrial designer Rama is in Dubai and Randy the animator is back in New York right now. >> Oh, [ __ ] Dubai. >> Rama Rama? >> This is Rama Rama.
21:32
>> Oh my god, one of my idols. >> Yeah. >> Who is Rama Rama?
21:37
>> Uh Rama used to run a company called Rama Works uh and they made very, very nice mechanical keyboards, like solid stainless steel machined keyboards. Um Yeah, he's in Dubai.
21:50
Thankfully, he's still alive. >> Is he okay? Yeah, god.
21:53
He needs to get out of there. >> He's he's okay.
21:56
So, he sends me pictures occasionally of people like sitting outside and eating and drinking and he's like, "Bro, it's totally fine here.
22:02
The media is uh misleading you." So. >> Wow. >> I think it's fine.
22:09
>> You just can't fly in and out.
22:11
>> You can't fly in and out, but you can drive in and out. >> Oh, yeah. Wouldn't recommend that.
22:17
>> Um of all things, I have a Rama Works uh dish right here that I put things in on my desk because I don't I don't use a mechanical keyboard as much as I would want to uh and as much as would fit my type as like a design bro gadget nerd, but for some reason I never bought them, but I had to buy whatever I could find from him >> Of course. >> uh stick on my desk.
22:39
And so, yeah, mad mad respect for Rama.
22:41
I feel like uh his industrial design work and even his websites and his photography were absolutely top-tier, like worldwide.
22:48
You know, up there with like Apple, Teenage Engineering, Opal.
22:54
>> How how did you get him to come work on your little toy?
22:57
>> Yeah, I mean, I'm so honored to be working with him.
22:58
Um it was a friend intro.
23:00
All of the best people who I work with were completely random friend intros.
23:06
Yeah, it's almost like the universe conspires to send you the exact right person at the right time. It's kind of crazy.
23:14
I've I've learned to trust it, but Ramo Ramo was a completely random friend intro.
23:18
Uh we started working together about a year ago.
23:21
He liked the project so much that he just like kind of immediately dove in.
23:24
Um and yeah, we've been working together ever since then. It's [ __ ] incredible.
23:31
I think he's he's a genius.
23:33
He's He's one of the best out there right now.
23:36
>> Where are you based, Daniel? Are you in New York?
23:38
>> I'm mostly in New York.
23:38
I go back and forth between New York and SF, but I'm uh I've been more in New York the past 6 to 8 months.
23:47
>> I feel like it actually is genuinely helpful for you to be in New York given you're making kind of like a fashion uh tech crossover product.
23:53
What is going on with the keychain and lanyard and charm situation in New York City right now?
24:01
Is it still Labubu or or what are we talking about?
24:04
>> Yeah, I mean Labubu is kind of on the way out, but bag charms are definitely still really big. >> Oh, god damn it.
24:09
I was about to buy my first Labubu.
24:13
>> I mean, you could be you could be late to it, you know.
24:15
The cycles are getting shorter and shorter, so maybe in a couple years it'll come back, but bag charms are still really big.
24:20
Yura and I actually went around Soho the other day and we went in all of the major fashion houses and they told us that bag charms are their number one selling item.
24:28
Um they're a more accessible kind of like entry into the brand because they're cheaper than the other products.
24:37
Um so yeah, they're they're really really big right now.
24:39
Um bag charms have been around for a while.
24:41
Uh it started with like Kate Berkin back in the '80s, so it's definitely not a fad.
24:46
They're not going away anytime soon and they're really hot right now.
24:50
>> Was that the insight you had for the design of this to have it be like that?
24:54
>> Yeah, I mean the design went through a lot of a lot of phases.
24:55
I mean back in SF, uh I would wear it sort of like a like a Humane AI pin.
25:01
I would wear it like uh like right there.
25:04
And it eventually kind of evolved into like, okay, this makes more sense on the bag.
25:09
It's kind of bold to wear it right here.
25:11
You don't always want to do that.
25:13
Um but yeah, I think we think it makes a lot of sense on a bag, on a backpack.
25:18
I wear it like on a carabiner on my belt.
25:20
Um it's just enough for people to notice it if they're really paying attention to you, but it's not like screaming that you want to be seen, which I think is ideal.
25:32
>> Yeah, I mean how much of a barrier do you think there is?
25:34
I feel like a big issue is just kind of like the AI perception if you will, because the keychains and charms like Labubu are more, you know, out there and ostentatious and attention-grabbing than they've ever been.
25:48
And I love that you've kind of got Disney animators working on the eyes for Starboy and the shape doesn't feel, you know, like some uh surveillance object, if you will, like many like many out there do.
26:03
>> Yeah, I mean we just like endlessly toiled over how we make this feel friendly and approachable.
26:07
And there are so many product decisions that that play into that.
26:14
We don't really call it AI. It's not an AI wearable.
26:18
It doesn't connect to your phone in any way. That that's a big deal.
26:20
People are used to stuff connecting to your phone and especially when it has a camera and a microphone, um people are kind of nervous about the privacy angle.
26:29
So when we tell people, especially in New York, that it doesn't connect to your phone and it's completely standalone, um it kind of like de-escalates it a lot.
26:39
People see it as more of a of a toy, which is exactly what we want.
26:43
>> Also looks like a shuriken I could throw at someone.
26:45
>> Yeah, the brass one and the stainless one are super super [ __ ] heavy.
26:47
It's definitely a weapon. But yeah.
26:53
>> Uh I wanted to ask you, Daniel, uh was Ellis helpful with the launch on this or no?
26:58
I like to ask that for every guest that he works with.
27:00
>> Ellis and I first met like a year ago.
27:03
Um you know, we had a few chats recently.
27:06
Um I mean, it was good to get a take, but I mean, to be honest, we didn't really use any of the ideas that he came up with. Yeah. >> What was Ellis' idea?
27:17
>> I don't know how much I I mean, Ellis, is it okay if I get in and do it? >> Dude, of course. Yeah. Destroy me. Roast me.
27:24
>> I mean, we only had like we only had 2 hours with Ellis, so it wasn't enough time.
27:29
Um I think Ellis came up with this copy that was really from a first-person perspective, like I am the little guy.
27:34
Uh but, that didn't really make sense because little guy doesn't talk. He's an alien.
27:37
He doesn't really understand English, and he can't speak.
27:42
Uh he's kind of like a cat.
27:44
Uh we say it's cat-like intelligence.
27:46
So, the first-person first like copy didn't really make a whole lot of sense.
27:51
Uh we think the third-person, like he's just a little guy, you have to say hi to him, that kind of thing, um vibes a little bit better.
27:59
But, I mean, I'm of the opinion that you just have to sort of tweet from the brain stem, and whatever is your first intuition is probably right, and the more time you spend toiling over it, the worse it does because people can sense that you uh you know, you're a bit insecure about it, maybe.
28:17
So, you should just you should just you should just say [ __ ] >> Well, it's the same when we tweet clips from the show.
28:23
It's like the more you try and refine it, uh the worse it comes out.
28:28
Similar with headlines as well.
28:31
>> You didn't get the full Ellis experience cuz his best ideas are in the shower.
28:34
So, if you really want to have that, you need to let him have some shower time.
28:39
>> It's an expensive shower.
28:40
>> When Ellis is wrong, we say, "Well, you had to see all the other options to know what worked and what didn't." >> Very good.
28:46
I mean, yeah, you're a you're a master storyteller.
28:49
>> As long as as long as you don't call me a wordsmith, I'm I'm fine.
28:54
>> Can I call you a word cell? >> Oh, I like that.
28:56
>> Better than wordsmith, honestly.
28:56
There was one good line from our from our work together though.
29:00
I felt like I liked the line, I'm just here to hang out.
29:02
Which conveys that it's like not a utility object and it also kind of hangs off the backpack or whatever, but I think you're I think you're right.
29:12
>> Yeah, when people ask me what it does, I say he mainly just hangs out.
29:16
It's not meant to be serious.
29:16
I think people in tech especially have a tendency to say like what does it do?
29:22
Like what does it do for me? What's the utility? Is it an AI agent?
29:23
Is it like a pedometer or something?
29:29
And this is completely the opposite of that.
29:30
This is like it's like computing doesn't necessarily have to have utility.
29:35
I I think we've kind of like reached the pinnacle of utility with smartphones and laptops and all of that stuff.
29:41
And like these these devices are going to keep getting better, but there's this completely unexplored like branch where we just have purely aesthetic objects, right?
29:50
And that's that's kind of what we are exploring here.
29:56
It's purely an aesthetic object. It's a toy. It's a piece of jewelry.
29:57
It doesn't really promise to improve your life in any sort of way whatsoever. It's just a fun object.
30:07
>> Well, so how do investors take that?
30:08
They say, well, but how is this going to help me accelerate my meetings by 0. 018%?
30:14
If you're not doing that, what are you doing?
30:16
>> It's a love-hate thing. It's a love-hate thing.
30:17
You got to find the one excited guy or one excited girl who's really excited about it and uh you know, a lot of people just don't get it and you know, we end those conversations early, but the people that do get it don't really take a lot of convincing, which is really nice.
30:32
And I think our current investors do get it.
30:36
>> Yeah, I was going to ask who are your investors and how much money does it take to get this thing off the ground?
30:40
>> Uh we took we took money from Bungalow, Lakehouse, and Cocoa.
30:43
Uh, Bungalow and Lakehouse are based in New York. Cocoa is in London.
30:47
Uh, they're all really small funds.
30:49
Like Bungalow and Lakehouse are both like two guys, and Cocoa is just one woman.
30:54
So, really really small funds. Uh, we raised 1. 2 million last year.
31:01
Um, we've been very very frugal about it.
31:03
I think like Rama has helped a lot with that.
31:06
Um, but yeah, I I think we've done quite an exceptional amount of engineering for the amount of money that we raised.
31:15
>> You're saying you built this it with barely a million dollars?
31:18
>> Yeah, and we've spent under half the money.
31:21
>> Lakehouse is cool, too.
31:21
I know Isaac uh, from over there, and I know they've had experience working with things that are more in like the culture lifestyle category, whether it's like that air sign vacuum, uh, their base at-home blood testing, Bobby, the infant formula that I used for a few years until I ran out of money.
31:42
Um, I feel like that is the exact right move for this type of thing.
31:45
But, I will say, I mean, it's easy to say, "Oh, yeah, this thing is just for fun."
31:49
But, I think >> People spend a lot of money.
31:53
>> Well, true, but where does the motivation come from for you?
31:55
I feel like for people that want to make really fun like lovely little things, you know, it often comes from, "Oh, they have kids."
32:04
Or there's a history where, you know, they were a tinkerer as a child, and they like to make their own toys, or this or that.
32:09
You know, where does that motivation come from for you guys? >> I like having fun.
32:14
>> What type of fun do you like to have? >> Engineering fun.
32:18
Yeah, I I like um, like I tweeted recently, my whole life revolves around silly art projects that require serious engineering ball knowledge, and that's been true ever since I started engineering.
32:30
Uh, I would make iOS apps, and I would like, you know, make all of the icons, and like like try and make them as artistic as possible.
32:36
I wouldn't use any of the Apple like standard UI stuff.
32:38
I would like try and make it as custom as I could.
32:43
And those like had utility, but they were really they were really art projects.
32:47
It was more like like software as art.
32:49
And the transition to hardware as art has been really really like invigorating for me because building a physical thing is just like it feels like alchemy.
33:00
It feels like magic when you can hold something in your hand.
33:06
And you can also do a lot more with the software if you're really smart about it.
33:10
Like I think the the Steve Jobs quote of like if you're really serious about software, you should build your own hardware.
33:18
That's really really true. >> quote. >> That's really true.
33:21
And I I understand >> day.
33:24
>> Yeah, I I understand why that's true now because, you know, when you're building on top of somebody else's hardware or OS, you're always fundamentally limited in what you can do.
33:36
But when you build your own hardware, you can do absolutely anything you want.
33:41
I can clock the chip however I want. I can like >> Overclock? >> I I can overclock it.
33:45
I can set the core voltage too low and like permanently brick it if I want to.
33:49
I can do all kinds of like crazy [ __ ] to get the result that I want.
33:55
I can extend the battery life and make it like stupid for a device like that.
33:58
So, it's it's really cool.
34:01
It's it's a cool engineering challenge.
34:05
>> Well, so let's get into that briefly.
34:07
For those who don't know, I mean Starboy literally launched to the world today, which is Wednesday. Um what can it do?
34:13
Tell us about the sensors and and all the different capabilities.
34:18
>> Yeah, so it it perceives the environment with camera, microphone, accelerometer, and temperature sensor.
34:24
So, it kind of fuses those those sensors together to give the appearance of a living object.
34:34
Um Um it's really like watching a fish in a fishbowl.
34:36
Like I think a lot of people have a tendency to view it like, okay, is it a game?
34:40
Is there some kind of goal to this experience?
34:42
Do I have to take care of it?
34:44
Uh that's completely the wrong way of thinking about it.
34:46
Um think about it more like a piece of jewelry or a nice Swiss watch.
34:50
You just buy it because it's like >> Okay, there's the investor connection.
34:56
It's just like a Swiss watch, guys. I swear.
34:59
>> Yeah, it's just like you're like three $300,000 Cartier or whatever.
35:04
>> What can it do with the camera?
35:04
Is it going to watch me pee? What is it doing? >> It will.
35:08
Yeah, if you point it at your dick, it'll watch you pee.
35:11
Um but Yeah, I mean it recognizes people, it recognizes facial expressions, it recognizes gestures.
35:19
Uh we're intentionally a little bit coy about what all it does because we want some part of it to be a mystery.
35:23
So, you have to buy one to fully realize the capabilities of it.
35:28
Um So, yeah, that's the camera.
35:32
Sound, it reacts to sound volume. It's not recording you.
35:36
Uh it doesn't have any kind of audio model on it right now.
35:38
So, it can't really understand what you're saying.
35:42
Uh sort of like a cat, but it reacts to changes in volume.
35:47
>> Oh, and you said it purrs like a cat, too, sometimes, right? >> purr like a cat.
35:49
Yeah, it has a haptic motor, so it purrs like a cat.
35:51
Every one of them has a unique purr.
35:54
Uh we came up with this like procedural kind of purring engine.
35:58
Um so, it's a lot of fun.
36:01
It It feels It feels very lifelike.
36:04
Uh I think a lot of people like don't really get it until they hold it in person.
36:09
And then after 5 or 10 minutes, they're like, okay, I'm really attached to this thing.
36:13
I I want to give it back to you.
36:16
>> Did you have this fully formed in terms of what you wanted it to do when you all set out to build it, or was it a thing that kind of evolved based on the design?
36:24
>> Absolutely [ __ ] not.
36:24
Yeah, I mean like any any hardware project evolves based on the design.
36:31
Um Yeah, I mean like every hardware project out there is a series of compromises to get to where it is now.
36:36
Uh I had like vague ideas of of what I wanted.
36:39
I had some like less vague ideas and I I think we delivered on a lot of that, but um you know, at the end of the day you have to compromise and the next one is going to be even better.
36:51
>> So, speaking of kind of the things that might have crossed your desk as you were building this, I know there's been the friend AI pendant, uh which is also deliberately limited, but has a very very has a very strong reputation for uh being kind of surveillant.
37:07
There's also that Curio AI toy that Grimes was participating in.
37:12
There's the little rabbit AI device, the Humane device.
37:18
What do you think any of any of those did did right or wrong with how they were thinking about this type of space?
37:23
>> Um I mean, I think Rabbit had really enticing hardware.
37:26
Um I think they absolutely nailed the aesthetic.
37:29
They just created an object of desire.
37:31
Uh and that was really good.
37:34
Curio, I'm not so sure about.
37:36
I feel like I haven't heard from them in a while.
37:38
Um Friend, I've I've never heard of.
37:41
I'm not really sure what that is.
37:44
>> Maybe all the Curios got together and started their own death cult and that's why we haven't heard of them.
37:49
They've been trying to talk them off the bridge. >> Maybe.
37:51
That would That would make it a lot more interesting.
37:53
Yeah, if they like started a a religion, I think that would be pretty cool.
37:59
>> Yeah, it is interesting though. You said desire.
38:01
I mean, I feel like when you look at Apple, a lot of people talk about, let's say, the quality or the quote-unquote highbrow design that makes it so good and so desirable.
38:09
But at the end of the day, you know, they are kind of like status symbols in the same way that like a rare Labubu is or something like that.
38:18
And I mean, I remember as a kid like I was the very first kid that had a Tamagotchi.
38:24
And it was just fun and it created a little type of relationship that as a kid I could kind of nurture and cultivate.
38:30
And I remember like going to the school bathroom like every 15 minutes saying I had to pee so that I could like feed or clean up after my Tamagotchi and then I remember going wild going wild for Furby when that came out.
38:46
>> Yeah, Tamagotchi kind of sucks as an adult.
38:48
If you've ever like tried one as an adult, it's it's [ __ ] awful.
38:50
You have to like clean up its [ __ ] every 15 minutes.
38:55
>> I already clean up [ __ ] every 15 minutes with my two kids.
38:58
So yeah, Tamagotchi is not appealing to me in its current state. >> Right.
39:03
Yeah, I mean we wanted something more more ambient, I think.
39:05
I think as an adult this this sort of experience can be really tedious.
39:10
Yeah, this this is definitely it's more ambient like it has needs like it will get upset, it will get sad, but it's less of a like you have to take care of this thing or it's going to die and you'll lose your [ __ ] >> Wait, when does it get upset? You could piss it off?
39:25
>> You can you can torture it a little bit.
39:26
Yeah, you can shake it a lot.
39:26
If you bring it into a really loud environment, he'll get kind of anxious.
39:33
Sudden loud noises are not very fun.
39:36
Super extreme temperatures so like really hot, really cold are not fun for him. But he likes people.
39:42
I think people in general tend to calm it down.
39:47
>> Is this thing for adults or kids?
39:47
I mean I saw in the launch video it was more kids, but >> It's for kids at heart.
39:52
I mean it is an expensive device.
39:54
So we think like adults are probably going to be the first ones who are able to afford it.
40:00
You know, we say like cool 20-somethings in New York or maybe even a little bit younger than that.
40:06
But you know, we we want to make it a bit unobtainable to begin with.
40:07
Like all of the kids should want one, but only like maybe the slightly more wealthy kids should be able to get one at least to begin with.
40:20
And there there may or may not be a cheaper version.
40:23
There probably will be more expensive versions.
40:25
We would love to do like one of one machines sapphire versions that are like $50,000 or something like that.
40:34
So yeah, I mean we're we're kind of using the like fashion brands playbook here.
40:42
>> Well, I wonder I mean like what excites you about that?
40:43
Obviously creating scarcity and making money can be fun.
40:47
But >> Yeah, I mean I think there's some element of like culture contributing to culture that motivates you guys.
40:53
Like is there a thread there? >> Yeah, absolutely. Absolutely.
40:56
Yeah, I mean we've said that we want this company to be every other hardware company's reference, right?
41:02
I think part of my motivation for starting this company was seeing everybody do the same thing and being really afraid to branch out and do something entirely different.
41:14
Like nobody in Silicon Valley at least is just making a pure toy.
41:16
Besides Curio maybe, but Curio is like targeted to kids.
41:22
Nobody's really making a toy for adults.
41:24
Um but yet adults buy toys.
41:24
We don't really call them toys, but adults buy like like Hermes bags and Porsches and nice watches and they spend a lot of money on them and it's it's a huge market.
41:38
So we we think it it needed to happen.
41:41
This kind of like fusing of of technology and fashion and and and toys.
41:47
Uh so >> It's a huge market says the guy in a hot tub in Lake Tahoe. Is what you're saying.
41:54
And you're supposed to be an artsy guy.
41:57
>> It's a huge [ __ ] market and if you give money, you're going to make a lot of money.
42:01
You're going to make a lot of money if you hand me some money. I promise you.
42:05
>> I mean I feel like you don't hear a lot of Apple Design Award winners saying it's a huge market. Am I right?
42:13
>> Yeah, I mean apps are kind of over, I think.
42:15
So, I I I feel like a lot of Apple Design Award winning apps don't really make that much money, you know?
42:22
Like Like you can point to exceptions like Flighty or whatever, they probably make decent money, but um >> Alex is a big Flighty die hard, I think.
42:30
>> I'm a big Flighty die hard, too.
42:30
I wish every piece of software was like Flighty, but um you know, I've been making apps for 10 years and I I think I just got kind of uh kind of bored, you know?
42:41
Like most most things have have been done already.
42:43
Uh people don't download new apps anymore.
42:45
It's really really hard to get anyone's attention with a new app.
42:49
Uh so, yeah, I I think that ship is kind of sailed.
42:54
So, I'm really happy to be in hardware.
42:56
>> You calling this a toy and refusing to acknowledge friend's existence suggests to me that you do not want this to be a companion. Is that fair?
43:02
Do you not see this as like a companion for people?
43:06
>> I mean, it's definitely a companion.
43:08
It's definitely a companion in the same way that a cat is a companion, but your cat doesn't understand you. >> Right.
43:14
You don't want people to think of it as like a like a a being, almost like a human.
43:19
You want people to >> it is a being.
43:20
It's a being, but it's an alien being.
43:22
Um and it's not really there for emotional support.
43:26
Um I mean, it can't It can be.
43:29
It can be, but we're but we're not promising that.
43:33
>> But if it reacts to temperature and all that stuff and loud noises, that's going to elicit emotion in the owner, right? >> It will.
43:40
I mean, we hope that it elicits emotion, but it's less of a like therapist kind of vibe.
43:45
I I don't want to make I don't want to make a therapist.
43:50
>> that's a bad idea to do uh companion therapist type wearables?
43:55
>> It's just not a world I want to live in. >> Why?
43:57
>> So, that's why I'm not doing it.
43:57
I think there are perverse incentives at play with with with that kind of like motive interaction, right?
44:06
Like whoever creates an AI therapist sort of, you know, just just like any other digital product, social media has a very clear incentive to keep you engaged for as long as possible.
44:22
And that's not always like commensurate with like helping you thrive as a human.
44:27
And, you know, while you could argue that it's okay right now, I think based on the history of like consumer software in Silicon Valley, I I think it's very clear the way that this is going to go in 10 or 15 years, and I don't really want to be a part of that.
44:49
>> It's going to end badly.
44:51
>> I think I mean, I don't think it's going to end.
44:53
Uh I think, you know, the story will continue, but um I'm not sure it's really going to be a net positive for most people.
45:02
>> So, I have a feature request, quick pit stop.
45:04
I want uh Starboy to be able to see that I'm doing the live long and prosper gesture and then give me some like Spock eyebrows or something like that in return.
45:16
>> We'll we'll consider that, but definitely not at the top of our list. >> I love it. >> Well, damn. Um I mean, it's funny.
45:22
It's like It's like I pretend I'm a 20-something's urbanite. Uh I really want to be.
45:33
Um yeah, I mean, I feel like the the tone of your video was was very kind of like Gen Z, like Y2K.
45:37
I mean, it's funny that it it it has like Gorillaz vibes right at the time that the Gorillaz uh album and new like short films launching.
45:48
>> That was a hap- happy accident. Yeah, we love Gorillaz.
45:49
Gorillaz was a huge reference for the video. >> Yeah.
45:52
Um one of one of my favorites.
45:56
>> So, Daniel, could you have made this without LLMs, without like modern AI, or was that >> It doesn't have >> any LLM's. >> It has no AI in it.
46:04
>> Well, you made a local one, didn't you?
46:06
>> We made a local model, but it's an image model.
46:08
It's a conflict or all that.
46:08
You could have done it in 2009.
46:10
I mean, the chip that we use was not around in 2009, but um yeah, I mean, it's like you could call it AI in the sense that what people were doing in 2009 was AI.
46:20
Uh but it's it's it's not an LLM.
46:24
It doesn't connect to your phone.
46:26
We certainly can't run that locally.
46:28
>> So, the reactions it can make and like the way that the screen can react to you is no AI in it. That's so interesting. Huh. >> I mean, it's it's AI.
46:36
I think like we we've had a lot of term rot with like the word AI.
46:40
I think AI is kind of synonymous with like generative image models, generative video models, or language models.
46:46
But uh you know, back in 2009, people were calling this like ML.
46:51
They were maybe calling it AI.
46:51
Um yeah, there's been a lot of term meandering with that word.
46:59
>> I'm going to attempt to contribute to the project again.
47:01
And on the box, you could put a sticker that says no AI included.
47:06
>> Yeah, I mean, we actually say don't call him AI because in his culture, that's very offensive. >> What's his culture? Say more, please.
47:15
>> You have you have to respect his culture, right? >> Alien culture. >> Alien culture.
47:20
We're not really sure what it is though, but we know that they don't like that word.
47:24
>> Yes, so talk to us about that.
47:24
Where did the Disney animators come into play for the eyes?
47:29
>> I mean, another like another happy kind of like accident friend intro.
47:34
Um we we work with this guy Ren Deep Katari in New York who used to be at Disney Animation at one point.
47:40
Um he's currently at Google.
47:43
He was at Netflix at one point.
47:45
But he is a brilliant character designer character designer and he is kind of created the whole persona for this device in a really really beautiful way and I've super enjoyed working with him for the past like almost year now.
48:02
>> Yeah, it's a very deliberate choice, isn't it, to go for something like cartoon eyes as opposed to I don't know, realistic eyes.
48:09
I mean, everybody's experimenting with it.
48:12
Like whether I mean, you laugh at realistic eyes, but that's what Apple was trying to do on the outside of the Vision Pro, right?
48:19
And then you see uh >> a good choice.
48:23
>> And then you see industrial design work like with the Neo humanoid robot where the eyes are kind of those circles, but then you see the Mimo robot from Sunday Robotics that are more like Nintendo Kirby eyes.
48:35
>> Yeah, I like I like Neo.
48:35
I like Sunday the best, I think.
48:38
Sunday has like created the friendliest version of a humanoid robot so far, which I think is really good.
48:45
Um I think people people in general really undervalue the friendliness of the robot, especially especially in Silicon Valley.
48:56
Like nobody really wants a Terminator robot.
49:00
Like as cool as it may look and as much clout as it may get you in tech, um I think the average person is is kind of terrified of that and I think like using Terminator as a overt reference like that is just not a really good move PR wise.
49:15
So, I really I really like what Sunday is doing and I like what Neo is doing.
49:23
>> This really makes me wonder how Jony Ive and open AI's device is going to go. Like which direction.
49:29
>> think it's going to be an anthropomorphized device from what I'm hearing. No, I I don't think so.
49:34
I think it's probably going to be like headphones or something like that.
49:36
Um we'll we'll we'll see.
49:38
Um I mean, Jony hasn't he hasn't really done a character before.
49:43
So, I would I would like to see him do that, but um I'm I'm skeptical that's what it's It's to be.
49:50
>> I mean, it's also a debate not just with like the anthropomorphized elements, but even the tone of the thing, you know, the tone between Claude and ChatGPT.
49:58
There's some browsers like Strawberry browser that has little characters that represent the agents.
50:05
Where do you feel like it's appropriate or not?
50:07
>> Yeah, I mean like I think there's a reason that people like Claude so much more than ChatGPT.
50:10
The the persona is just so much better.
50:13
Uh they I think they spend a lot of time just like making it pleasant to talk to, but um you know, not letting it overreach and like enable you or lead you to psychosis.
50:29
Um so yeah, definitely a big fan of Claude.
50:31
I I find the ChatGPT persona very frustrating most of the time.
50:38
>> Is the persona on Starboy something that's locked or is it something you can update over time?
50:44
>> It's unique per device. It's unique per device.
50:46
So like the eye look and the eye color are unique per device.
50:48
It's kind of like going to adopt a cat or a dog from a shelter, right?
50:52
Like you can't really change their physical features.
50:55
Like I guess you could give them a haircut, but um you can't change their eye color, you can't change their personality.
51:00
Even if you clone a cat, uh the personality doesn't transfer over.
51:05
So uh we think there's really something special about every one of them being one of one.
51:12
>> And how many of unique ones are there going to be or is that TBD?
51:15
>> So there's there's about 4,000 eye combinations right now.
51:18
Um it may grow, it may shrink.
51:23
Um but yeah, they each have associated rarity.
51:27
>> You're in uncharted territory.
51:27
I mean it's probably hard to even market map and think about how many of these you could sell, right? >> It's really tough.
51:33
Yeah, I mean we have we've speculated endlessly about how many pre-orders we're going to sell and uh you know, I I think at the end of the day you just have to put it out into the market and see what people think.
51:45
>> You don't >> a laptop facing you in the hot tub with the orders as they come in?
51:49
>> No, I have you guys facing me.
51:52
>> I also I noticed on here that you say ships by September 2026.
51:54
I think a thing a lot of people feel with new hardware or startups is like frustration that they never ship on time.
52:02
>> Yeah, I mean we we we feel pretty confident that we can ship by September.
52:07
Um we would like to under promise and over deliver because I think there's an epidemic of hardware startups in SF that just ship complete vaporware.
52:15
>> And I think that's really [ __ ] lame.
52:16
>> Names, names, names, names.
52:18
>> I won't name any names on your podcast, but um you know, you'll you'll know it when you see it.
52:23
Um Yeah, I I think like we've we've waited to announce the pre-order for many months until I had a fully working prototype because I want to take actual pictures of it with my iPhone.
52:37
Um I think those like perform so much better than some random render even though we think our renders are really cool.
52:43
So, yeah, I mean the message that we're sending is that it it's not vaporware. It actually exists.
52:50
Um and we're going to ship it by September.
52:52
>> Well, so what are those other guys doing?
52:54
Is it that they are lying or that they >> Yeah, I mean I think a lot of them are lying.
53:02
>> Yeah, I mean what was the like the glasses project recently that um >> Oh, I lose track.
53:07
I feel like there's so many.
53:09
>> Well, there was some kind of like AR glasses project recently.
53:11
Um it's a um you know, it's it's like a like a pickled vegetable kind of product where, you know, some guy made a bet that they wouldn't actually ship any pre-orders. I think he he won that.
53:25
>> Oh, I know what you're talking about. Yeah.
53:27
>> Yeah, yeah, you yeah, you know what I'm talking about.
53:28
I mean like obviously that's not [ __ ] real.
53:29
And I think it's kind of like almost become acceptable to do that, which is which is ridiculous.
53:37
Like we should we should hold ourselves to higher standards than that.
53:42
>> Well, I think with hardware especially, you know, with software it's like whatever, you know, like it doesn't really matter at the end of the day.
53:48
It's just pixels and >> Kind of, kind of. Kind of.
53:49
I mean, I mean, I mean, I think it matters for the culture, right?
53:53
Like if you if you create a culture where you can ship vaporware, then um people stop paying attention because they can't tell what's vaporware and what's not.
54:05
>> Yeah, and I think like I experience that every day with the storytelling.
54:07
I mean, when you talk about what your product does, if you say what it does, the company is like, all right, we're saying the value, the benefit, we're saying, here's what it does for you. Great.
54:17
That may be true, but if their audience has been lied to 15 times over the last few years about the exact same value and benefit, you can't make that same value and benefit. >> Yeah.
54:29
Yeah, I mean, it's the same reason that it's so it was so difficult to start a biotech company after Theranos blew up, right?
54:36
And it was especially difficult for like women to start biotech companies.
54:40
Uh they kind of poisoned the well for everybody.
54:42
Um it's it's a profound like disservice to humanity, I think, because there are plenty of people who want to make real things that are really good, and they have a harder time because of some idiot who's pushing some vaporware.
54:59
And I I think that really [ __ ] sucks.
55:01
>> So, speaking of company building, Daniel, I know your last company was a really cool thing.
55:05
I mean, uh it was called Any Distance, a running app.
55:10
Beautiful design, good name, good swag.
55:13
It wins the Apple Design Award.
55:13
It aligns, you know, with exactly what you would want culturally, the rise of running and the rise of like high-performance, high-fashion running. And then, it's over.
55:23
And I saw in your post, you said, uh due to a fairly common combination of classic startup failure modes.
55:33
Um can you can you tell us any more about that?
55:36
And what you're doing differently this time?
55:38
>> Yeah, I mean like I I don't want to get into like super details there, but you know, it was a combination of like co-founder disagreements, which are the most common reason that startups fail.
55:47
Uh it was a combination of not quite finding product market fit.
55:49
Um I think we like leaned too far into the Twitter audience angle, which is inherently limited.
55:57
It's a very small set of people.
55:58
You know, we're launching this product on Twitter now, but I cannot [ __ ] wait to get on Instagram and start pushing it there because that's where the real audience is.
56:05
Um and I'm a solar solo founder this time, which is really great.
56:11
Um you know, I've I've done this a few times before.
56:12
I feel like I know how to start companies logistically.
56:17
So, it's been really refreshing to be the only one in the driver's seat on a project like this.
56:24
>> So, you're saying you're doing the Kanye I fired my manager because I cannot be managed. Is that what he said? >> Yes, sir. >> of the best words.
56:33
>> Um well, well, now that you're out of the running app game, are the running apps real? Is it a fantasy?
56:39
>> I mean, I think the running apps work.
56:41
Running apps are a very crowded category.
56:42
I think that was the other problem that we had is that um you know, there are tens of thousands of apps that have taken advantage of every possible sensor that you could use on the device.
56:52
So, you know, like short of making your own hardware, I I think that category is one of the most saturated app categories out there.
56:59
So, it really became a brand's play and I think we failed on posting content on platforms other than Twitter and creating a brand around that.
57:09
We were kind of just relying on Apple's to do our marketing for us and that's not the way to do it.
57:16
Um I also think, you know, like I mentioned earlier, people download fewer and fewer apps nowadays.
57:22
They're satisfied with the apps they already have.
57:24
So, it's a really, really, really tough pitch.
57:26
You know, it's not like 2011 where you launch a new app and everybody freaks out about it.
57:33
Like >> I mean a theme we've been talking about this on the show with like Eugenia from Lobby etc.
57:38
is like software is just getting super personal.
57:42
Uh I think that was it was one of the Collyer brothers that were saying recently it's software is going to be like pizza, you know, like made fresh and disposable and personal.
57:52
Um you you think maybe on that?
57:54
>> I I I don't know about that vision.
57:54
Um I think people said the same thing about C++ back in like 1980 whatever.
57:58
You know, C++ finally lets the average person like talk to computers and not write assembly anymore.
58:06
Uh I I think this is a a vision that's just been rehashed over and over and over again.
58:10
And I think the truth is most people want to get on Instagram and scroll until they die.
58:14
And I I think that's unfortunately like kind of true.
58:18
Um I I don't think the average person like they might have a vague idea of an app, but I I don't think it's like fully fleshed out it off like idea-wise, you know, assuming that the agent can make it perfectly and do whatever you want.
58:30
I I don't think the average person has ideas.
58:33
>> I think there's also just the idea that people like to use and watch really thoughtfully made things by artists, right?
58:43
Like there's actually um if you're only thinking of it from a utility point of view, you're like, "Yeah, I can make my own app."
58:49
And if it didn't exist then great, but if I had the choice >> There's a reason designers exist, right?
58:56
Like there are non-obvious insights about how to design software that the average person can't really apprehend, right?
59:04
>> Or brands for that matter. >> Yeah, or brands.
59:05
And so you like hand them something that's really really well designed and then like, "Okay, this is what I want."
59:11
People don't really know what you want until you show it to them. It's true. >> Yeah.
59:15
I'm also very happy to have your C++ reference on the show.
59:20
Uh in tech we do not have very many fresh references running around on a daily basis.
59:24
It's usually some combination of uh think different and Bicycle for the mind.
59:28
That's all we know in our world.
59:31
Or or at least at least people aren't talking about her anymore, right?
59:35
I feel like I haven't heard about her in a minute. >> Yeah.
59:37
Yeah, I mean I think I think Spike Jonze was not really okay with the tech industry interpretation of that.
59:42
I think he was more like trying to tell a good story and the AI was kind of a convenient plot element that let him tell that story and I think everybody in tech of course they like they they misinterpreted it.
59:56
Um I also think the the bicycle of the mind has become more of a flashlight of the minds and I kind of I kind of wonder what, you know, like I when I watched Johnny at at Figma Config, you know, last year or 2 years ago, I could detect a lot of shame in his voice and a lot of like, you know, desire to rectify the situation.
1:00:16
You know, I I wonder what Steve would think if he was still alive and I wonder how these people who created the technologies we use today really really interpret what has happened.
1:00:29
You know, not saying that it's entirely their fault, but um yeah, I I think it's it's not a bicycle.
1:00:38
It's not a bicycle anymore. That's very clear.
1:00:41
>> It's funny to think I think even while Steve was alive, Apple was experimenting with Facebook integration at the OS level for a couple of years there.
1:00:47
Do you guys remember that?
1:00:50
>> I Yeah, I don't know about that.
1:00:50
I've had people tell me that you know, Facebook had talks with Apple and >> Oh no, I mean it was in the contacts app.
1:00:59
It was in the contacts app.
1:01:00
>> Yeah, they integrated it a little bit, but Facebook had this angle where they wanted to have apps within apps and Steve was very not down for that for obvious reasons.
1:01:10
>> Guys, Zuck and Jobs were friends.
1:01:10
Like he was almost kind of like a mentor to Mark when Mark was young before they did ads. Um >> Wow, sources. news.
1:01:18
I did not know this at all.
1:01:21
>> Everything that Zuck wanted to do in early Facebook days, um, Steve did not want to do and that was mainly the apps within apps thing for obvious reasons.
1:01:29
They wanted to control distribution.
1:01:31
>> But they were they were doing uh iOS integration for a little bit uh and then Twitter came in. Um what could have been? Yeah.
1:01:39
>> What made that relationship work, Alex? Do you know?
1:01:41
I mean, I feel like they're both unbelievably curious people, which I think has emerged as one of Steve's most uh important virtues, including in the words of of a Johnny Ive letter I wrote the other day ju- or read the other day uh talking about Steve Jobs's insatiable curiosity.
1:01:57
But I feel like Zuck was curious about completely different things.
1:02:01
Like Zuck, I think was curious on like how to make what people want the most. >> Yeah. Yeah.
1:02:08
I I don't think Zuck has as good taste as Jobs.
1:02:13
>> No, I think where they probably saw eye to eye was like being platform builders and being ruthless business strategists. >> That's true.
1:02:21
>> Which they both both certainly are and were.
1:02:24
>> Yeah, abs- absolutely.
1:02:24
But I mean like I I think Apple is very intentionally like not made a social network because of the like, you know, there there's a lot of gray areas around making something like that and I'm I'm glad that they haven't done it, but >> Do not speak ill of iTunes Ping or am I dating myself?
1:02:43
>> even I don't even remember that, dude. I I wasn't even alive. >> iTunes Ping!
1:02:47
>> As a You weren't alive during Ping? Jesus, I feel so old.
1:02:48
Um as our youth uh hot tub correspondent, do you think Apple is not relevant in the youth culture anymore uh like they were?
1:02:59
>> I liked what I've what I've seen recently out of them.
1:03:01
I think they posted some cool TikToks recently. >> For the neo? >> Yeah, for the neo.
1:03:05
I like that they're making colorful devices again.
1:03:07
I think that's really really cool.
1:03:09
Um I'm excited for Steve Lemay's leadership.
1:03:14
I'm really thrilled that Alan Dye is gone.
1:03:16
Um so, you know, I I have high hopes for Apple.
1:03:19
I think I think they're going to make it.
1:03:21
I mean, every everybody has an iPhone, everybody has AirPods.
1:03:24
I think they're still very, very relevant, but they're relevant like as as a hegemon and not really so much as a tastemaker like they used to be.
1:03:33
>> Who is the tastemaker in tech? >> Good question. Good question.
1:03:35
I mean, that's what we would like to be.
1:03:37
Um, I I think there's like not an obvious tastemaker in hardware.
1:03:40
Like maybe I could point to Teenage Engineering, but they only make music devices, but they are a reference for everybody.
1:03:46
I could maybe point to Nothing, but their primary market is India, not the United States.
1:03:52
So, um, I I think there is a gap for a hardware tastemaker, and that's what Creature is.
1:03:59
>> Guys, the tastemaker is me.
1:04:03
And I have a lot to show it.
1:04:03
I have a lot to show and tell to close out the show.
1:04:08
>> I have so much to show.
1:04:08
Hand me lots of money and I will make the taste.
1:04:12
>> Okay, well, I'm going to show my show my show and tell really quick for uh for the Instagram Reels and shorts.
1:04:15
I have my Sony PocketStation. >> Yo, that's awesome.
1:04:19
I love the transparent vibe.
1:04:21
I'm so happy that's coming.
1:04:24
>> I have my Sega Dreamcast VMU. >> Wow.
1:04:28
>> Which you could have a little pet on, which is the inspiration uh you didn't know you needed.
1:04:34
>> Yeah, you brought out all the big guns.
1:04:35
>> I have my friend pendant that's been now going to put you uh under hypnosis. >> What is that?
1:04:43
>> I have my Nothing Ear buds, also transparent cuz I'm that guy. >> Classic.
1:04:48
>> And let's bring it back a little further. I've got I've got more. I've got more.
1:04:51
I've got my analog pocket. >> Let's go. >> Peak Design. I've got my Playdate. >> Hell yeah.
1:04:59
>> Which I can't see unless it is the absolute perfect lighting conditions.
1:05:04
I have my original Snapchat Spectacles.
1:05:07
>> Yo, I have Spectacles, too. I love Spectacles.
1:05:09
The packaging was so cool.
1:05:11
They did this like kind of soft thing at They did it too.
1:05:15
>> like a It was like a tennis ball cube. >> Yeah. So [ __ ] cool. I love Spectacles.
1:05:20
>> And um last but not least, I have the really slick 3D Spectacles that kind of got I don't know, forgotten about. >> Oh, I remember those. >> cool, too. >> Yeah. How are those?
1:05:35
>> Uh they hurt a little bit, but I have some good some good vision. >> good vision proud.
1:05:40
>> You know what's actually cool about these is that Snap partnered with a company to print little lenticular stickers of your 3D images. >> Cool.
1:05:50
>> Which is like ostensibly exactly what you'd want to do.
1:05:51
I mean, I think you know, Evan understands toys, I think, at a deep level, but >> Yeah. Snapchat is a toy. >> Right.
1:06:00
When you have 450 million DAU though, it's it's like real [ __ ] hard to try and create that scale with the hardware, as well, right?
1:06:07
>> Any kind of toy thing you make is going to be highly opinionated, and once you get to that scale, it's um it becomes increasingly difficult to be that opinionated.
1:06:16
>> Well, so why do you think, last question, Starboy is likely to work?
1:06:22
>> I mean, I think everybody's going to want one.
1:06:23
I think it doesn't currently exist.
1:06:26
Um I think like hanging out in New York and showing it to random people who are not in tech whatsoever, um they're like, "Oh my god, that's the cutest [ __ ] thing.
1:06:35
I want a um bag charms are really hot right now.
1:06:39
Uh I think collabs are going to be a huge angle for us.
1:06:43
Every fashion brand out there wants to have more of a tech kind of integration.
1:06:49
Like every fashion lab has a like some like hardware AI kind of lab, but they all don't really know what they're doing. >> Supreme Starboy?
1:06:59
>> Supreme Starboy would be great.
1:06:59
Our dream collaborator is Homer.
1:07:01
We would love to do a Homer Frank Ocean Starboy.
1:07:05
I think that would be super super viral.
1:07:08
>> There's got to be someone listening to this show that knows Frank.
1:07:09
So, we'll manifest a Frank connection for you.
1:07:13
>> Let's Let's manifest that.
1:07:14
>> Do you think the Snapchat Spectacles and the Humane AI Pin and Google Glass on the runway over the years helped or hurt them?
1:07:22
>> I think for Humane, it was I'm not really sure.
1:07:25
I didn't really like their rollout in general.
1:07:27
It felt a bit too inaccessible, I think.
1:07:29
Um like obviously we would like Starboy on the runway, but it needs to have more of a silly, down-to-earth kind of vibe.
1:07:38
Like we use uh the brand Kidsuper as a big reference for the kind of runway vibe we want.
1:07:45
It's just like incredibly whimsical.
1:07:47
It's not like hey, this is some unobtainable New York Paris thing that only models wear.
1:07:54
Uh we we don't really want to be like that.
1:07:57
I think the Humane was also it's just too much of a tech product.
1:08:00
Like it's it's an intimidating product to wear.
1:08:02
It's It's a $700 like camera right here that connects to your phone and and projects on your hand and it's like I I I don't I don't know.
1:08:13
I I think it um it's almost like the same vibe as like carrying around a big DSLR instead of carrying around like a vintage film camera.
1:08:22
Like photographers talk about this all the all the time how when they do street photography, you want to carry like a tiny little film camera because disarms people.
1:08:29
You don't want to carry like a huge [ __ ] like like Canon thing, right?
1:08:32
So, I And I think the Humane AI Pin is is the Canon thing, unfortunately.
1:08:37
>> "Please hold while my laser scans your retinas. Thank you."
1:08:41
>> while my AI laser camera scans you and talks out loud.
1:08:44
Like that's kind of terrifying, you know?
1:08:47
Like we we don't want to be terrified.
1:08:49
I get that terrifying is kind of like edgy and cool, but um I I think there's there's like a limit to it.
1:08:56
Like it gets a lot of attention, but I'm not really sure it converts into people actually liking it and buying it. It's just rage bait.
1:09:02
I >> Well, I'm a bit terrified that you've been in this hot tub this whole time. Are you feeling okay? >> I'm feeling fine.
1:09:08
It's actually really cold out here. >> Okay.
1:09:11
Yeah, I was going to say before you faint uh and then we have a problem on our hands.
1:09:15
Maybe we'll maybe we'll let you go back to your launch as well.
1:09:17
I'm sure you have a lot of DMs and emails to uh attend to.
1:09:21
But uh thank you so much for joining us.
1:09:24
>> Yeah, I I can't wait to get the dopamine rush.
1:09:26
I can't wait to get that dope of Twitter notifications, but uh yeah, thank you guys.
1:09:36
>> And that's it for this week's show.
1:09:37
Thanks to Daniel for coming on straight from the hot tub.
1:09:40
If you like this show, don't forget to like and subscribe everywhere you get your podcasts.
1:09:46
>> I could never stay in a hot tub that long. We are access. show online.
1:09:47
And you can find us in video at accesspod on YouTube.
1:09:54
You especially want to do that this week.
1:09:56
>> We would also love a five-star review or thumbs up and share with a friend or digital friend even. >> Or an AI friend.
1:10:04
You can find my newsletter at sources. news.
1:10:08
>> And you can find me at hamburger on Twitter and at meaning.
1:10:10
company for your startup storytelling needs.
1:10:14
>> Access is part of the Vox Media Podcast Network and this show is produced by Hooked Creators. >> Bye. >> Bye.