0:00
I mean, didn't you just tell Trump you were gonna spend like 600 billion? I mean, that's >> I did. Yeah.
Search this video
0:03
Through 2028, which is uh >> that's a lot of money. >> It is.
0:06
And and if and if um and if we end up misspending a couple of hundred billion dollars, I think that that is going to be very unfortunate obviously, but what I'd say is I actually think the risk is higher on the other side.
0:22
[Music] Welcome to Access from the Vox Media Podcast Network.
0:28
I'm Alex Heath and Ellis, you are I am Ellis Hamburger, not your favorite sandwich, but your new favorite podcast host.
0:38
>> I have had uh a lot of people since I've been saying that I'm doing a show with you ask me if your actual last name is Hamburger. >> It is uh verified. >> Yeah.
0:49
And you had hamburger on X, which is a flex.
0:51
Uh, >> that's why I'm so scared to leave.
0:56
Please don't make me leave.
0:57
>> Ellis, why are we doing a podcast?
0:57
I feel like there's so many podcasts, but you know, uh, I I've been getting that question a lot, too.
1:05
>> Yeah, Alex, I think we have great chemistry.
1:09
>> We've known each other for a long time.
1:12
We both, I think, see a different side of tech as it is today.
1:16
>> I feel like you're so well connected in big tech.
1:19
uh you love to smoo with all the biggest founders.
1:22
I've got a good strangle hold. >> Yeah, you do.
1:26
I've got a good uh good strangle hold on on the AI startup arena with the work that I've been doing at meaning.
1:32
And I feel like we could just really bring something different.
1:34
I mean, you've been in media forever.
1:35
I was starting in media at the Verge, then went into startup world for a while at Snapchat and the browser company.
1:42
And so, I think we have something interesting.
1:45
I think we want to talk about >> the inside conversation, what people are really thinking and talking about as opposed to just what's in the headlines.
1:54
So, >> inside >> uh hopefully we could do that.
1:57
>> Yeah, I think we both just wanted to make a show we wanted to listen to and didn't feel like a show like that existed.
2:05
>> And I hope we're going to make that. I think we will.
2:07
And it feels really fun to be doing it with you.
2:09
And the way we're going to structure these episodes is, you know, it's a talk show and interview show. Pretty standard.
2:16
You and I are going to rap about some things happening in our world, things we think that uh you're going to want to know about, even if you're pretty plugged in that's that's coming soon or stuff that just hit.
2:28
And then we're going to go usually to an interview either with a big name.
2:34
This week we've got the one and only Mark Zuckerberg.
2:36
Uh next week we've got uh Dylan Field, the CEO of Figma for his first pod since the biggest tech IPO of the year.
2:44
We're going to have an interesting mix of I think big names and then also early stage founders, some of which you work with directly at meaning um that we just think people should know about that are going to be the companies that you're going to be hearing about uh in the next few years or maybe already.
3:01
And uh we want to make something that feels feels good if you're tapped in, but also relevant if you just want to know more about this this crazy tech world we're in.
3:12
Did I did I get all that?
3:14
>> Yeah, I think you got it all right.
3:14
I think the one other thing that's on my mind is I just want to have fun with this tech has been a part of my life for so long and while this industry is so often mired in, you know, often valid skepticism, pessimism, uncertainty, I think there is still so much brightness and optimism and fun uh just to have in building the future together.
3:39
Obviously, we all want to be honest and hold each other accountable, but I think tech is culture these days and I want to cover the whole thing holistically uh versus just the last earnings call and uh what the latest RPO and DAU numbers are.
3:55
>> Yeah, I can ask that stuff.
3:55
I mean, I'm I'm pretty interested in the business and the strategy of these companies.
4:01
It's what I cover a lot at Sources, my new uh publication about the tech industry and Silicon Valley and AI.
4:07
I was previously deputy editor at The Verge where I had a pretty successful newsletter there called Command Line.
4:18
And now I'm uh I'm an entrepreneur, I guess, like you, Ellis.
4:20
Uh I'm I'm on the other side.
4:22
And this pod is part of that.
4:25
I I hope it feels part of the same cinematic universe, I guess.
4:27
But, uh, Sources is going to be, I guess, maybe where I also play a little more bad cop to the the good cop of of the mood we're trying to make on this show.
4:37
We found out that Zuck wanted to do the first episode.
4:41
That gave us a deadline I think neither of us were planning for, but it's been great.
4:46
It's been a good uh, kick in the pants to get this thing going.
4:49
And, you know, usually we're going to do these interviews together.
4:53
This week is a little unusual just because of the timing and so it's just Zach and I.
4:57
Uh but you know I think there's also an element of the different perspectives we can bring to these conversations right like you've got this really interesting perspective working with a lot of these startups directly and then I've got my kind of more journalist POV of having met a lot of these leaders in big tech especially in the big AI labs over the years. >> Yeah.
5:22
Well, first I want to get at the real distinction between how you would interview Zuck and how I would interview Zuck.
5:26
I think you're going to be in one of those fireside chat chairs.
5:27
I want to go t-shirt shopping with him.
5:31
>> Maybe head to the jewelry store after, look at some chains.
5:35
>> That's the interesting combination I think that we represent.
5:36
Alex, >> support for access comes from Warp.
5:39
Warp is a topranked coding agent and one of the fastest growing AI companies.
5:46
Trusted by over 700,000 developers and 56% Fortune 500, Warp is redefining how software gets built.
5:52
It understands your machine stack and codebase.
5:55
Combining the power of a terminal with the interactivity of an IDE, giving you a tight feedback loop with agents to prompt, review, and ship productionready code.
6:05
Find out why developers save an average of 5 hours a week.
6:07
Try Warp Free or Unlock Pro for just $5 at warp. dev/ac.
6:19
Support for access comes from Notion.
6:22
Ellis, you're a Notion power user, right? >> Oh, yeah.
6:24
I've been a Notion power user for years.
6:27
And when we had one month to bring access into reality, Notion helped us track our guest database, to-dos, and notes all in one place. And that is the truth. >> That is the truth.
6:37
We run this show on Notion. And you know what?
6:38
They just introduced Notion agents.
6:40
They're almost like teammates that can use Notion with you.
6:45
They can draft, organize, and execute across your workspace and keep your work moving.
6:50
We'll talk more about how we use Notion later, but for now, try Notion and agents for free at notion. com/fox.
7:00
So, I got an early sneak peek of the conversation with Zuck, and I'll say this, he seemed very confident, very comfortable, having fun.
7:08
What was the vibe you got from him during the conversation? What was he wearing? How was he feeling?
7:15
Um, it's pretty pretty crazy, you know, to go from sitting next to Trump to uh sitting next to Alex Heath.
7:22
It seems like he's he's got some swag these days.
7:26
>> I mean, yeah, the the swag has been the the Zuck arc for the last couple years, I would say.
7:31
I mean, you notice the chain is tucked in this year.
7:32
He was still wearing it, but it's tucked in, which I don't know if that means, you know, we've got business to do, which is maybe this new AI stuff.
7:39
Uh, I do think they feel pressure on these glasses.
7:44
They really want these glasses to be wellreceived.
7:46
They're Rayban branded and they're pretty wild. They can do a lot.
7:53
Uh it's not full augmented reality, but it's a pretty good heads-up display that can do texting, heads-up navigation, a bunch of a bunch of other stuff.
8:02
And they have this neural band that controls the glasses that feels like legitimate sci-fi.
8:08
It's one of the coolest demos you will do.
8:10
Um, and they announced it this week at Connect and people are going to start seeing that seeing them out in the wild.
8:17
They're kind of expensive, you know, they're they're $800.
8:20
I think these are very clearly a early adopter uh kind of product um proumer so to speak and I got to demo them last week and they are really cool.
8:34
You cannot be a fan of tech like you and I are Ellis and not think these are cool.
8:39
Now, whether I'm going to wear them all day or they're going to start replacing the smartphone, I think that's very much TBD, but uh yeah, man, I was I was pretty impressed.
8:49
>> So, what was the magic moment across all these different use cases that you tried with the glasses?
8:53
Because I feel like there have been so many ambitions for what they could be.
8:57
I mean, I was back at Snapchat in the early days when we launched Spectacles, and all you could do is just take like a 10-second video or a photo with them.
9:03
We've seen people iterate on them over time.
9:05
I know you mentioned to me that it did a lot more than you expected.
9:10
Like what was just first principles?
9:12
What was the best thing you tried?
9:15
>> There was this moment where I felt like I had super hearing.
9:18
It was a thing that only something like this form factor could do where they're calling it, I think, live captions.
9:26
So, I was in this room with a bunch of people and um you could look at someone and everyone was talking very loud and if you were looking at them, it would live caption what they were saying even if they were six to eight feet away and it was super loud and you couldn't hear them on your own.
9:41
And then they've added language translation to this where it can do live language translation back and forth.
9:45
Um that was pretty magical.
9:48
Uh the display itself is is honestly pretty good.
9:52
It sits to the side of your eye, which is kind of unusual when you first try it, but it lets you see the picture you're about to take or the video you're about to take, which is honestly something that feels simple, but in that form factor of a camera on your face, it actually matters a lot.
10:11
And the band has this little gesture where you can like twist in a knob like in in the air to like zoom in and zoom out.
10:18
So, it feels like you're Tom Cruz and Minority Minority Report a little bit.
10:22
The crazy thing is is with the band, and I really can't describe how much of a game changer the band is for input because you don't have to just talk to it or wave in front of it to get it to do something.
10:33
You can just do this very light tap gesture.
10:36
>> You do a little pinch, >> little pinch, and the display just melts away, comes back, melts away as you pinch.
10:42
>> Um, and I started getting that pretty fast.
10:44
I wore the glasses for about an hour, took them outside, took the demo off rails a little bit, asked the AI things that weren't verbatim what they told me to ask. It still worked.
10:53
You know, like place uh you know, some glasses and um plates on this table, make it look like a table setting, it did it.
11:01
You know, stuff like that.
11:01
The display is like 5,000 nits or something.
11:05
It's super bright and it's very clearly designed to just be worn everywhere you go.
11:10
And the battery lasts around 8 hours.
11:12
I bet it's less than that with the display going the whole time and it really get to push that to the limits.
11:16
But um overall >> I love a good nits conversation. >> Goodness.
11:20
>> But how many nits is it? >> Yeah.
11:22
>> Is it 8K quantum dot or is it ocular occlusion? >> Yeah.
11:27
>> No, this is Gorilla Glass version 3 vestibule.
11:31
>> I think you made some stuff up there, but that's good.
11:33
>> Well, so setting aside the display, I feel like so much of AI is, you know, obviously the hardware has to be there, but how does their AI work?
11:39
You know, I feel like especially in the age of AI and MCP and people trying to do things agentically, so much of it is like what sources is it using? What can it do?
11:48
Can it plug into your other services or was it just kind of like better Siri at least at the outset?
11:56
>> I would say better visual Siri.
11:56
It's still meta AI which obviously is not the leading AI.
12:00
AI definitely took a backseat in the demo.
12:02
I think they probably wished that they had more AI features, but Zuck is doing this massive AI reboot that everyone's been reading about.
12:10
I asked him about that actually in the interview.
12:12
We talked about it and he dropped some new stuff, I think, about the lab.
12:14
That was pretty interesting.
12:16
I don't want to give it away, but um I think they know they're behind on AI, but they have the bare minimum.
12:23
And no one has a product anywhere close to this in terms of the form factor, the price, the display, the band, and the ability to like do texting with the band.
12:33
Zuck said he types 30 words per minute with the band, which I >> Yeah. So, how does that work?
12:39
>> It's like auto scribbling with your finger.
12:42
>> Yeah, it's like autocomplete with slight wrist finger gestures.
12:44
Like you can write almost like on your leg or whatever and it autocompletes it.
12:48
Um, and he's doing he says he's doing 30 words per minute. Uh, which is impressive.
12:55
I think it's really more for just shooting off like quick texts, which I did and it worked.
12:58
Um, but yeah, it's uh it's it's a wild device.
13:02
Like it's definitely something if you are listening to a show like this, you're going to want to try.
13:10
So, watching Zuck for a great many years, I feel like he's been trying to build and own the next platform forever.
13:17
You know, he's tried phones, he's tried VR, trying to build a metaverse, now AI glasses.
13:27
Alex, you are a betting man.
13:27
Do you bet that this is the one where the platform works?
13:32
Cuz it is such a it is such a fine line, right?
13:35
Like Apple is does appear to be uniquely good at combining the hardware and the software cuz >> everybody's competing on the hardware.
13:42
I do think at the end of the day, the software is going to be a pretty big deciding factor.
13:47
or does it have the apps you want?
13:48
How does, for example, vision OS feel to you compared to the OS that they're kind of building across AR and VR and this and that?
13:56
Is this is this the platform where they win?
14:01
>> I think it's going to take a few years for this to really get mature and become something that is compelling to a lot of people, not just early adopters.
14:07
You can totally see the path to it, though, when you try this, I think.
14:12
Um, and I felt that way when I tried Orion, their first pair of like full AR glasses that are not a consumer release last year, which that when I say Orion in the interview, that's what I'm talking about.
14:23
Um, yeah, I think compared to the Vision Pro, it's definitely not as fullfeatured, but it's a different use case. It's not a headset.
14:32
It's not a pair of goggles that fully blocks you off from the real world.
14:36
I mean, these are chunky, but like in the right lighting could pass almost as normal glasses, and that's their main job.
14:45
I mean, their main job is to be something you can wear around, and the tech is supposed to be supplementary to that.
14:53
And I think there's a lot of work to do to make the tech actually feel like it fades into the background appropriately.
14:59
Um, I think Meta is very motivated to figure it out. You're right.
15:04
They really want another platform and they're sinking billions of dollars into all of this because they, like pretty much every big tech company I know of, thinks that this combo of glasses with a display and AI is maybe going to be the next smartphone.
15:22
>> So, I have one more question for you before we get to the feature presentation, an evening with Alex and Zuck.
15:29
Um, we like to talk about people here on the Access podcast, even though this is the very first episode.
15:36
So, every time that I feel like Meta goes through a different transformation, we hear about how that team has been moved next to Zach's office. >> Yeah.
15:47
>> How does that work out over time?
15:47
Do you just kind of find yourself in the inner orbit and then with each new tech trend, you get a couple inches away kind of like tectonic plates?
15:55
Who's in the inner circle right now?
15:58
Who has been pushed to the outer orbit?
15:59
Yeah, the inner circle right now is the new AI lab, which I think I'm the first outsider to see physically.
16:08
Uh when I was there for my demos, they walked me in and I was getting a lot of side eye, I would say, from the researchers in there who are like, "Who is this guy that's looking at our llama algorithm written out on these whiteboards?"
16:20
Luckily, I don't know math, so their secret is safe with me.
16:26
Um, but yeah, there the lab is is sitting there in this kind of special area with Zach and uh, as I say in the interview, they were they were cranking.
16:35
I think I saw some shoes off uh, a lot of code happening and uh, it's very clear that Meta is is rebooting this stuff and I think this device is part of the reason why.
16:47
I think they know that AI is the killer feature for glasses like this and they want to be on the frontier of that.
16:55
Most importantly, what are they drinking?
16:57
We got Viples, Monster Energy, Diet Coke is making a comeback. What was on the tables?
17:03
>> I didn't catch the drinks.
17:03
That would have been a good catch.
17:05
But, uh, you know, these big tech campuses, they have just about everything you need.
17:08
You know, you never need to leave.
17:10
It's like Hotel California. All right, man.
17:15
Well, uh, I guess we'll we'll get into the convo with Zuck here.
17:19
>> Support for access comes from Notion.
17:22
Notion is the connected workspace where teams write, plan, track, and ship together.
17:27
And with Notion agents, those teammates aren't just human.
17:30
Agents are AI teammates that can complete multi-step workflows end to end, building pages, updating trackers, and moving work forward while you focus on the hard decisions.
17:41
We even use Notion ourselves.
17:43
I use notion to run my business, meaning from using AI notes during meetings with very verbose founders, creating beautiful clientready docs, and I even generate my invoices on there.
17:54
So nice to have it all in one place.
17:55
We actually run the show on Notion, too.
17:57
It's how we track our guests, to-dos, and everything we need to get you access every week.
18:01
The result of all this, less chasing, more reporting, and better conversations on mic.
18:08
Notion is where the newsroom and the studio meet and agents keep the work moving.
18:13
Try Notion and agents for free at notion. com/vox.
18:18
[Music] Support for access comes from Warp.
18:23
Warp is changing how developers code.
18:26
As a topranked coding agent and one of the fastest growing AI companies, not to mention trusted by over 700,000 developers and 56% of the Fortune 500, Warp is redefining how software gets built.
18:41
Warp understands your machine stack and codebase and is built at the terminal level, meaning it can write code, respond to system events, and even deploy to prod.
18:50
And their mixed model approach outperforms single model approaches.
18:55
It's built for the entire software life cycle from prompt to production.
18:59
With the powers of a terminal and the interactivity of an IDE, Warp gives you a tight feedback loop with agents so you can prompt, review, edit, and ship productionready code.
19:11
Find out why developers save an average of 5 hours a week and see how Warp transforms your coding timelines.
19:19
Try Warp free or unlock Pro for just $5 at warp. dev/acess. dev/acess. That's work. devac.
19:33
Mark, I don't think you can be into technology and the cutting edge like I am and not try these here in the middle, these new display glasses, and think they're just not really cool.
19:42
Uh, and I want to get into what they do and why you're building them, but can you kind of just initially set the stage for us and explain why you all are doing a display in this form factor?
19:55
Because you've had the AR glasses and you have the glasses that don't have displays.
20:01
So, why do something in the middle here?
20:04
>> I mean, we're working on all kinds of glasses.
20:05
I mean, my my theory is that, you know, at a high level, glasses, I think, are going to be the next computing platform device.
20:12
Um, I think that they're great for a number of reasons.
20:17
One is that they don't take you away from the moment, so you can stay present in the moment, unlike with phones.
20:21
So, I think that that's a big deal.
20:23
Now, they're also basically the best device for AI because it's the only device where you can basically let an AI see what you see, hear what you hear, talk to you throughout the day, and then once you get the display, it can just generate a UI in the display for you. So, that's great.
20:39
And then the other thing that glasses can do is it's really the only form factor that can put holograms in the world to help you seamlessly blend the physical world around you in the digital world.
20:52
Um, which I mean I just think it's it's a little crazy that we're here. It's 2025.
20:56
We have this incredibly rich digital world and you access it through this like 5 in screen in your pocket.
21:01
So that I think like there's going to get these things blended together.
21:05
So that's glasses at a high level.
21:06
But then you get into okay well what do people want with glasses?
21:10
And glasses are very personal.
21:13
It's not just like a phone where everyone kind of is okay with something that's pretty similar.
21:16
Maybe you get a different color case.
21:18
Um, people are going to want a lot of different styles.
21:23
Um, people are going to want different amounts of technology depending on whether they want a frame that is on the thinner side or bulkier side or whether they can afford more technology or less.
21:36
So, I think there's just going to be this whole range of different things.
21:38
um from, you know, simple glasses that don't have that much technology in them, maybe the ability to talk to AI and have it see what's going on around you, all the way up to full augmented reality glasses that have kind of wide field of view like the Orion prototype and and everything in between and different styles, right?
21:58
So, we did uh we started with Rayban, which is probably um the most single most iconic and popular glasses design in history.
22:06
And now this year uh we're adding Oakley.
22:09
So we did the um Oakley Meta Housins this summer and then at Connect we announced u these guys the the Oakley Meta Vanguard which we'll get to in a bit.
22:21
>> Um this is I think what people kind of had in mind when they thought when they heard that we were doing something with Oakley.
22:25
I think it was it was more this >> but >> these are dope. >> Yeah.
22:28
I mean it's I mean they look great.
22:30
They're great for performance and we'll we'll talk about that in a minute.
22:33
But the the but the deal is I mean people are going to want all kinds of different things.
22:36
So um there's going to be this whole spectrum and one important technology is obviously going to be getting um getting a holographic display and then within that there's a whole world of options too.
22:47
You could have a small holographic display that can just display a little bit of information.
22:51
You could have wide field of view that can basically overlay kind of avatars and deliver a sense of presence >> which was Orion last year.
23:00
>> That's Orion and that's kind of what we're building towards in the consumer version of that.
23:02
Um, so there's a number of different points on the spectrum and what we're doing here with the Meta Rayban display, I think, is a good kind of starting point where it's not a tiny display.
23:16
It's actually it's it's quite meaningful.
23:17
You can read a whole text thread, you can, you know, watch videos, you can do a video chat, you can watch videos that you've taken.
23:23
Um, I guess you could even watch reals on it if if you want.
23:27
Um, but so it's it's a meaningful size display.
23:30
Um, but this one isn't really meant for putting objects in the world.
23:36
It's more meant for just showing information.
23:38
And um, so anyway, we've been working on this for like I mean all the glasses at Meta we've been working on for more than 10 years at this point.
23:46
So, you know, when we get when we have these moments along the way where we get to like show a new technology that I think is pretty different from what others are working on, which the the display in the glasses is one thing, but the the meta neural band as the way to interact with it with you just get these like micro gestures in your hand and you're controlling what you're see it's it's just wild. >> The band is wild.
24:08
We we got to talk about the band.
24:12
that surprised me the most in my demo of the glasses last week.
24:17
Frankly, I mean, I've been reporting on these build, you know, coming for them and I I thought they would have a little bit of a more limited use case to start, but they can do quite a bit.
24:27
And I'm I'm curious, what was the goal of their overall functionality?
24:32
What What are you trying to achieve with this?
24:34
Are you trying to replace the phone or just get people to use it less?
24:39
I mean, what's what's the big picture idea of like this is what it can do?
24:43
>> Well, I always think about everything from a communication and connection angle first, right?
24:49
Because if that's that's kind of the the legacy and the the DNA of the company.
24:53
Um, so probably the most important thing that I've focused on wanting to get them to do is be able to wear them, get a text message, respond really quickly and subtly with your hand if you want.
25:09
Um, with like like we're having this conversation now and I could I mean like we're talking about like this level of hand motion, right?
25:16
Like I'm making it's like not >> I thought you might wear them in the interview and then I wouldn't >> I mean I could I could I could Yeah, I'll put them on.
25:22
Yeah, that's um um another thing about them is you can't you can't tell they have a display like even when and they're transitions and even when they're sunglasses.
25:32
>> So that's actually an important part of the technology is um is light leakage is a feature of some types of waveguides.
25:41
And so basically you get these trade-offs where you want them to be very efficient.
25:47
Um, there are waveguides where you have to pump just a ton of light through them in order to get anything to show up.
25:52
Um, but then some waveguides have um just different artifacts.
25:59
Um, and usually in a bad way.
26:02
It's like the light will catch them and you'll see all kind of like rainbowing or something.
26:05
Um, another artifact that we think is pretty bad because it's a privacy issue is if the person who you're looking at can basically >> see it. >> Yeah.
26:14
the the very worst version of it would be if they could see what you're seeing.
26:18
Um, but I think >> another version that I think is still not that socially acceptable is if they can see that you're looking at something at all. Right.
26:28
Um, so I I think that that that's one of the things that we're really proud of in the design here and that that we put a lot of energy into is, you know, the the uh the displays are super bright to you and to the person that you're looking at, they can not even really tell that you're doing anything.
26:48
And that that's an important thing for having it to be socially acceptable, right?
26:51
I mean, we we also we designed them so that way when when the display comes up, it's offset slightly to the side.
26:58
We don't want to block what you're doing.
27:00
An important principle for the glasses is the technology needs to get out of the way, right?
27:04
It's I mean, fundamentally, it's like, you know, this is something that you're going to be wearing on your face for a lot of the day.
27:11
I mean, we designed these specifically to be um, you know, both indoor and outdoor that with the transition lenses, they work really well sunglasses outdoors.
27:18
Um, but the reality is is that most of the day you're not going to be using technology or at least visual yet, right?
27:25
So like maybe uh maybe you'll be listening to music or something, but um but we want, you know, when you are interacting with something, it should show up.
27:33
It should kind of be off to the side.
27:35
If you don't interact with it, it needs to get out of the way really quickly.
27:37
That's like a really important principle of the of the whole technology.
27:41
>> You've got this wake gesture with the band where you can just tap quickly to make the display go away. >> Very very subtle.
27:46
>> And again, I want to get into the band.
27:48
The band is wild in its own right.
27:50
>> Um, >> the thing that really stood out to me from my demo from these was some things that you only could do in a form factor like this.
27:57
Cuz I mean, the texting is cool, but like there's this live captions thing where I was in a room with a bunch of people and they all started talking really loudly and I if I just looked at someone, it would live caption what they were saying and tune out everything else. It's like super hearing.
28:12
>> Um, and then you're also doing that with language translation.
28:14
So, you can do real time.
28:16
I mean, ideally, both people are wearing the glasses to get the full experience, but you don't actually need to have the other.
28:21
You could just hear what they're saying in your language or see it.
28:24
>> Yeah, >> that's pretty wild.
28:25
And that speaks to like what just this form factor could do.
28:28
I'm curious like there's that.
28:28
Are there other things that this form factor uniquely can do in your mind that a smartphone can't?
28:37
>> Um, I mean the all the things around AI. Yeah.
28:42
where you have an AI that basically you want to have context around what's going on with you, right?
28:48
So like if you want an AI that can >> see what you see, hear what you hear, can just kind of talk to you passively throughout the day, um, and then can show you information contextually.
28:58
That's just not something that a phone can do.
29:01
I mean, I guess technically you could walk around holding a phone like this, but you can't really do that. That's no one does it.
29:05
Those demos have existed forever and I'm always like I don't want to hold my phone up and >> so so I think that's actually going to be the main one.
29:11
Um and I think all the live AI stuff. It's interesting.
29:15
It takes on a different um like a different feel.
29:20
So we have live AI in the Ray-B bands without a display too, the the kind of classic Ray-B bands.
29:25
And for that it's audio only live AI.
29:28
So, it's really helpful for when you're doing something kind of um by yourself.
29:33
If you're cooking or something, then you know it can just give you it can it's it's watching what you're doing with the video >> and and you can ask it questions about what you should be doing or it can give you tips and that's all that's all great, but it's not really useful when you're in another kind of conversation.
29:54
And the thing that the thing that I've observed with um kind of the thought experiment I've run but also just wearing these is you know so we go through the world we have dozens of conversations a day um and in every conversation I usually have like five things I want to follow up on.
30:12
Maybe there's like some it reminded me that I should, you know, go do a thing or or it reminded me of a person who I wanted to talk to.
30:21
Or maybe I'm I'm talking to someone and um maybe they like assert some assumption that doesn't quite sound right and I want to fact check it or like gut check it.
30:33
Um, these are all things that I think with live AI, you can have this AI that's sort of running in the background and that goes and often does work for you and um and then can bring that context back whether it's asynchronously kind of offline when um when you're done with a conversation or sometimes like when you're in the middle of a conversation it's just useful useful to have more context right then. Yeah.
30:56
>> How have you been using these?
30:56
Someone on your team was saying you text a lot through them.
31:00
>> Yeah, I um well I'm a texter.
31:00
I like run the company through text messages.
31:03
So, um, so and when you were asking about what, you know, what can you do that you can't do on a phone?
31:13
I mean, I guess you can you can obviously text on a phone and we all do it dozens of times a day, but I think that there's like a lot of times where it's just not socially acceptable to send a text message.
31:24
And so, like, let's say you're in the middle of a conversation, you want to like ask someone a question or get some information.
31:30
I have this like all the time.
31:31
I'm like having a one-on-one conversation with someone and I like a I like wanted to ask someone this or like I I I wanted to ask someone else a question to like pull some context so I can ask this person what they think about it but I'm not going to like pull out my phone in the middle of a conversation.
31:46
Um with this it's actually just super quick.
31:49
You you can just like send a message in like five seconds, get the context back.
31:54
Like it actually just really improves the conversations that you're having.
31:57
I find it's um to me this is the one thing that I think is basically better about Zoom than inerson conversations is that you can sort of multitask a little bit, right?
32:10
It's like a lot >> it's it's worse in basically like every other way than than kind of an inerson physical conversation.
32:16
But the one thing that I that I think is useful is you can go from having a conversation to basically asking someone else a a question.
32:25
It's not necessarily distracting. It's additive, right?
32:27
because otherwise your option is like all right um you have a conversation then you go check in with someone else then you have to go back and call the other person back and have a whole second conversation so it just short circuits these things all the time and now I think this kind of brings the best part of that into physical conversations
32:44
while you basically feel present in in the conversation you can pull in whatever information you need it's super super helpful >> yeah a real like holy thing about this product is the band I I thought that cuz it was with Orion the demo you guys had last year too and I thought it At the time I was like, there's something special with this band. And
32:58
And you're calling it's a neural band. Is that right?
33:02
That's >> Yeah, neural band.
33:03
It's a cuz it's a neural interface.
33:04
It picks up nerve activity.
33:06
>> So like it feels like it's reading your mind.
33:08
It's not doing that, but >> it's not reading your mind.
33:09
What you're doing is you're >> sending signals through your muscular nervous system that it actually picks up before you even make movements.
33:19
But um you it basically picks up these micro gestures and that allows you to control your interface um no matter where your hand is.
33:32
So it's not doing like hand tracking visually or anything like that.
33:35
Like you could have your hand by your side, you could have your hand behind your back like whatever in your jacket pocket um and it's fine.
33:40
And um and the gestures are really subtle, right?
33:44
It's like it's like this is all I need to do to bring it up.
33:48
I mean this brings up Met AI.
33:49
So, and I really like the music one.
33:52
Did Did you um >> I didn't try that one. No.
33:53
>> Oh, so the way when you're listening to music that you adjust the volume >> is you just kind of >> Oh, the dial thing. >> Yeah.
33:59
You pretend that there's a dial and you just turn the dial.
34:02
>> I did that with the zoom in on something.
34:03
>> Yeah, you could do you could do it on on photos, too. >> Yeah.
34:06
>> Um >> that's like It feels like Minority Report when you do that.
34:09
>> It's like in real life.
34:10
>> It's a It's a good interface.
34:10
Yeah, I think is what you mean by Yeah.
34:12
It's like not not not the weird part of Minority Report.
34:15
Yeah, the good part where Tom Cruz is doing the hands like it's it feels like sci-fi >> and I'm wondering >> why this band like why did you land on this band as the input for this cuz people have been trying to figure out input for glasses like these forever and it's usually voice or hand gestures or something but it's like I'm not going to be in the subway like gesturing out into space.
34:37
>> So, okay, I think that those are going to be useful too, but I don't think that they're I don't think that they're complete. Right.
34:45
So voice is is obviously going to be important.
34:47
People talk to MEAI, they do voice calls, they, you know, you do video chats.
34:52
Um, so voice is going to be a big thing, but I think the reality is that a lot of the time we're around other people.
34:58
And for the use case that we really wanted to nail, which which I actually think is the most frequent and most important use case that we do on our phones, is messaging. Mhm.
35:05
So, if you want to nail that, what you need is the ability for a message to come in to not be distracting, not be like center in your field of view, but just be there.
35:16
Um, and then you need a way in whatever situation you're in to be able to quickly respond in like five or 10 seconds um in a way that is not interruptive to your your current interaction and is socially acceptable and doesn't feel rude.
35:31
And so, then you get to okay, hand gestures.
35:34
I mean, yeah, I I think that there are going to be useful things for Yeah.
35:37
I mean, in Minority Report, he's doing a fair amount of that, but for gaming and things like that, I think you'll do that, but like you said, you're not going to walk down the street like that, right?
35:45
I mean, that's that's kind of um >> pretty goofy. >> Yeah. It looks It looks weird. Your arms get tired.
35:49
You know, it's um uh much more the former than the latter in terms of reason why it doesn't work.
35:54
But the the latter is also true.
35:56
Um, so we needed something that was basically silent and and subtle.
36:06
>> So the questions are there there are a few options for that.
36:09
One that people are working on is basically whispering, right?
36:14
So you can like >> subaudibly even um pick up on the sound or you could have some camera that can like look at your mouth and do like lip reading.
36:25
That's still pretty weird in a meeting. >> I agree. I agree.
36:27
So, it didn't pass my bar for kind of subtlety even though it is silent in theory.
36:37
>> So, I I think like so I think you need to go for the neural interface.
36:43
>> And the other thing that's nice about the neural interface is you can get really high bandwidth input too.
36:46
Um, so it's not like, you know, with like, you know, smart watches today, you basically like, you know, you can move your arm and it can pick up like a gesture or two, but it's but it's a very low bandwidth.
36:59
There aren't that many things that it can do.
37:00
Um, you need something that can basically be reading the muscle signal.
37:04
Um, so that way you can just control it very subtly.
37:08
>> And um, >> and this can do that.
37:09
I mean, I can I can I mean, already I I haven't like, you know, we're not that optimized.
37:13
I think we're going to get the auto correct a lot better, but I'm already I think around like 30 words a minute typing. You can Yeah. No, it's Yeah. Yeah, >> man.
37:21
I mean, how how advanced do you think the band gets in its current form factor in terms of what it can do?
37:27
>> Um, I think quite a bit.
37:27
I mean, basically today, well, I guess you have the sensors which can pick up the signals from your muscles.
37:38
Um, but then on top of that, it's basically just an AI machine learning problem to be able to pick up what you mean by the thing.
37:49
And right now, um, it's not particularly personalized.
37:54
So, um, >> so you get it out of the box and it needs to work with certain gestures and you've never used it before.
38:03
So it works with these like like you're you're doing kind of it's like I mean this isn't a big gesture but I mean this is much bigger than what it needs to be in the future.
38:11
And then >> in the and then for for um kind of the neural text entry you're basically you can kind of think about it as if you have a mini pencil >> and you're just like writing out the letter.
38:24
But over time, what should happen is that the AI learns your pattern for how you, you know, write each letter.
38:34
And you should be able to write make like increasingly subtle and invisible motions that it basically learns are your way of doing that letter or that input or whatever.
38:49
And I think the future version of this is that the motions just get really subtle and you're effectively just firing muscles in opposition to each other and making no visible movement at all. Um and it picks it up.
39:02
>> So personalized autocomplete via your wrist basically. >> Yeah.
39:05
So super fast because if you're not moving there's no latency from actually having to actually physically move and then retract after after making a motion.
39:13
>> Um so I think the upper bound of that is very high.
39:16
very high. The the other thing is um and that's just for typing which is one kind of modality but >> I think you can kind of there are all these different dimensions where you can use it to input into things and you could control a a hand in space that is like operating a UI right it's like
39:35
there's um all kinds of different things that it can do that that I think will just be really interesting to get into over time um and you know we basically We invented the neural band to work with the glasses, but I actually think that the neural band could end up being a platform on its own. I
39:55
I >> to basically just interact with all of your electronics and and devices and and do all kinds of different things once we kind of get it to be a little bit more mature.
40:05
>> So, you could have an API for it that could theoretically plug into a smart home or something like that. >> That'd be wild. >> Yeah. Yeah. >> Um Yeah.
40:11
>> The price point for these is also lower than I expected.
40:13
It's only like 800 bucks. Yeah. >> Like this is for you.
40:22
You're not making a ton of these, right?
40:23
I assume this is not like a mass going to be a massive thing for you.
40:27
It's more to to see how people use this technology.
40:29
What's >> I mean, I think that this is going to be a big part of the future.
40:33
I mean, my my view is that, you know, there's between a billion and two billion people who wear glasses on a daily basis today to for vision correction.
40:45
Like is there a world where in five or seven years the vast majority of those glasses aren't AI glasses in some capacity?
40:50
Like I I think that that's it's kind of like when the iPhone came out and everyone had flip phones and it's like just a matter of time before they all become smartphones.
40:56
I think that these are all going to become AI glasses and the question is all right well there are 8 billion people in the world not you know one to two.
41:03
So are there going to be a bunch of other people who also wear glasses? I would guess yes.
41:08
I mean there are a lot more people who wear sunglasses some of the time.
41:10
Um, so yeah, I mean I think it's a big category.
41:14
Um, and there's a lot going on here.
41:17
I mean there's I think what you see when you're building products is that V1 um you build what you think is going to be great and then you get a lot of feedback, but you also didn't get everything exactly perfect in V1.
41:36
So, you know, V2 and V3 just end up a lot better, right?
41:39
It's not a it's not a coincidence, I think, that, you know, the first version of the Ray-B bands, Rayban stories, um, we thought it was good.
41:50
Then when we did the the second version of Rayban meta, I think it sold five times more.
41:55
Um, and it was just refined, right?
41:58
It was it was like So, I think that there's going to be some dynamic here where it's like you have the first version, you learn from it, the second version is just a lot more kind of polished and take and and it's and the software gets polished too, not just the hardware and that just kind of compounds and gets better and better.
42:13
>> Um, >> and full AR that fills your vision that's still coming too. >> Yeah.
42:18
I mean, we're working on all this stuff and we want to get it all to be as affordable as possible.
42:22
Um, the reality is that the more technology you want to cram into the glasses, the more expensive that is because you're putting more components in.
42:30
Um, we also want the glasses to be as thin as possible.
42:32
And um, that's a process of miniaturaturization that happens.
42:37
And similarly, the more technology that you cram in, the harder it is to make smaller.
42:41
So, >> as much as we can miniaturaturize this technology, >> it will always be true that if you put half the technology in, you'll be able to make even thinner glasses.
42:49
>> Um, and then some people have different aesthetic preferences where they'll want, >> I mean, like fortunately, you know, thick glasses are kind of in style.
42:55
Uh, but, you know, some people want thinner ones like yours.
42:58
You can't fit many electronics. Yeah.
43:00
Well, it's >> Yeah, you may have to rethink your aesthetic choices in the future.
43:06
>> Uh >> but >> um >> see on pricing we um I mean we did we work on on getting it to be as affordable as possible.
43:15
And um you know I mean our view is that our profit margin isn't going to come from a large device profit margin.
43:24
It's going to come from people using AI and the other services over time.
43:28
um >> because you'll pay a subscription for the AI or something >> or yeah or use it and do commerce through it or whatever whatever the different things are that that people do.
43:37
So um so we're not like you know a company like Apple that is that whose primary or a large part of their margin comes from from having a large margin on the hardware.
43:48
Um but in general yeah we try to make it as affordable as possible.
43:51
And my hope is that, you know, if we build another one of these, hopefully it's even more affordable or with the other choice that we can make is put even more technology in it and keep the price point there.
43:59
But I think you're going to have >> a few different price points.
44:02
There's going to be the >> kind of standard AI glasses that don't have a display.
44:09
And I think those will sort of range between, you know, 300 to$500 or $600 depending on, you know, the aesthetic and and kind of how high fashion it is.
44:19
um maybe even more than 600 if you get something that's really high fashion.
44:23
But that's kind of the range that we've seen so far from the the early Ray-B bands to to some of the the uh Oakley Meta Vanguards with like all the kind of you know custom stuff in it like you know optical lenses that can um with a prescription. >> Yeah.
44:38
Um, then there's a category like this with the kind of a display that isn't kind of a full field of view AR display.
44:46
And I think that that's going to be >> yeah on the order of $1,000 like maybe you get it a little better, maybe it's a little more, but I but you can call it in that area.
44:57
And then I think when you get to the full AR glasses, um, >> that'll be somewhat more to start.
45:02
And and I think like people will just want to have the whole portfolio.
45:07
And then the goal over time will be to get as much of that technology into as affordable and as thin of a form factor so you can just like have as many styles as possible.
45:19
>> Support for access comes from Warp.
45:19
Warp is changing how developers code.
45:22
As a topranked coding agent and one of the fastest growing AI companies, not to mention trusted by over 700,000 developers and 56% of the Fortune 500, Warp is redefining how software gets built.
45:37
Warp understands your machine stack and codebase and is built at the terminal level, meaning it can write code, respond to system events, and even deploy to prod.
45:46
And their mixed model approach outperforms single model approaches.
45:51
It's built for the entire software life cycle from prompt to production.
45:55
With the powers of a terminal and the interactivity of an IDE, Warp gives you a tight feedback loop with agents so you can prompt, review, edit, and ship productionready code.
46:07
Find out why developers save an average of 5 hours a week and see how Warp transforms your coding timelines.
46:15
Try Warp free or unlock Pro for just $5 at warp. dev/ac. dev/ac. That's work. dev/access.
46:30
Support for access comes from notion.
46:32
Most work these days feels like busy work.
46:34
It's hunting, updating, and coordinating.
46:36
With Notion AI, you get your own AI team powered by your work and all in one place.
46:41
It knows your work context, your meeting notes, tasks, calendars, and plans.
46:46
And now with agents, it actually does the work for you and your team.
46:49
Anything you can do with notion, your notion AI agent can do, too.
46:54
Agent can build pages, structure data, search across your tools, and execute multi-step workflows at your direction.
47:00
Ellis, how are you going to use a notion AI agent?
47:04
>> Honestly, I'm excited to use an agent to turn all my notes into nice, easy interview questions ranked by juiciness.
47:12
Of course, >> we got to have it juicy.
47:13
At the end of the day, Notion is AI that understands the way a team works because it lives inside the same workspace you do. Notion plus agents.
47:22
Fewer tabs, fewer handoffs, better work.
47:25
Start for free at notion. com/vox.
47:32
>> And so you've got these new Oakleys and your deal with Elsor Lexodica means you can do other smart glasses with all their other brands, right?
47:38
So I think that implies, right, that there will there will be future brands that have met to do it. Yeah. >> Yeah.
47:45
>> So, you see it as like building out a kind of constellation of all these different form factors and price points and >> that's the goal. Yeah.
47:52
>> What What about all these other AI wearables that aren't glasses that are happening?
47:56
Like there's the friend pendant.
47:57
I'm sure you've seen there's all these like displayless non-glasses devices.
48:01
Sam and Johnny I apparently working on something.
48:04
There's a lot of interest in this and I'm curious is this something that since AI has really taken off in the last few years you see opportunity there in addition to glasses or are you still just the main focus is is glasses?
48:18
>> Well, our main focus is glasses because I think glasses are the best form factor for this for all the reasons that we talked about before.
48:24
Um, I I think that anything else that you have to kind of fiddle with takes your attention away from the the physical world around you.
48:33
Um, I don't think that there's any other form factor that can see what you see, hear what you hear, talk to you throughout the day, and generate a UI and your vision.
48:40
Um, >> and then there's the whole augmented reality part about blending the physical and digital world. But I don't know.
48:47
I mean, people use different electronics.
48:51
So, I mean, I I'm not going to I mean, I I I certainly don't think that it's like in the future all 8 billion people in the world are doing the exact same thing.
49:00
I mean, some people use their phone more, some people use a computer more, some people use an iPad instead of a computer, right?
49:05
Some people, you know, primarily um, you know, watch videos on a TV, some people watch videos primarily on a on phone.
49:12
So, I do think they're going to be different things.
49:16
Um, my guess is that glasses will be the most important.
49:19
Um, I think something like earbuds is kind of interesting, too.
49:24
I mean, Apple clearly is, yeah, like by far the leader on that with AirPods.
49:30
I think partially because they did a good job and partially because I think they gave themselves some um kind of unfair advantages with how they bundle it and couple it and and have technology that works with the um with the phone that I guess they're now just starting to open up which is great but um but for a while I think it just made it impossible for anyone else to build anything like the AirPods.
49:52
Um, watches I think are interesting. >> Mhm. >> Um, in some ways too.
49:56
And I kind >> You're not a fan of the like the pendant thing, the pendant trend that's kind of starting right now.
50:04
>> Um, I mean, is it a trend?
50:04
Is that >> there's a lot of startups doing this stuff?
50:08
I mean, >> I think it's an interesting idea.
50:09
I I don't want to be I don't want to be too dismissive.
50:11
I mean, my my point is that >> I actually think that they're >> My guess is that different people are going to like different things, but that glasses are going to be the most popular.
50:22
There was no new quest this year at Connect and I'm curious how you're feeling about the quests these days in VR mixed reality generally as a category.
50:29
It seems like glasses are really taken off.
50:31
There's obviously a lot more quests out in the world. They've sold a lot more.
50:36
>> Um, but I'm curious, you know, not there not being a new one this year and also just how you're feeling about it these days like the category. >> Yeah.
50:43
No, I mean I think we're we're making progress on it.
50:45
I mean this year what we focused on was the Meta Horizon creation tools.
50:50
So, we announced um Meta Horizon Studio and Meta Horizon Engine, which are are these basically foundational tools for creating worlds and content using AI.
51:01
And um >> and that's going to go towards making it so that people can create a lot more content um in VR, but I think that that's also going to be the case in um all that stuff I think should have should translate over to AR too.
51:15
I think a lot of this content you'll be able to to have there.
51:18
Um, I mean, glasses that are that are see-through may not be quite as immersive as VR, but you can deliver a lot of the same kind of holographic experience.
51:27
And then, um, I think a lot of these things will also um, end up showing up on phones, right?
51:40
I mean, and you know, it's I I think that there's this huge opportunity with AI >> to, you know, you're browsing your feed on Instagram and Facebook and like it's like each story um should be its own world that you can jump into.
51:54
But like the the and you're starting to see some of this with some of the AI models, some of the stuff that Google has put out recently, for example, are like interesting glimpses of of where that could go.
52:03
Um, but I think that there's this real sense that the whole stack for how you create those kinds of immersive experiences um, needs to get rethought.
52:17
It's not just going to be people doing things in the same way that they've created 3D video games historically.
52:19
I mean, that's a very very intensive um, process where the tools are like very >> hybrid. Yeah. >> Yeah. I mean, >> yeah.
52:32
I mean, so my my kids are kind of into are into programming and into making things and, you know, we try to, you know, build different 3D worlds and I think some of the stuff is just like intractable for them, right?
52:43
It's like it's it I mean, they're still kids, so it's fun to make >> extracted to like a prompt.
52:47
I mean, then it becomes >> I think by the Yeah.
52:49
So, that's but with with the um Meta Horizon Studio, which I've been playing with with with them and and you know, obviously, you know, it's this isn't primarily designed for like an 8-year-old to be able to use, but my my bar is if it's if it's if it's enough that like I can kind of make something good with my with my 8-year-old, then um that's pretty cool.
53:08
Um you really can create all these interesting things, right?
53:12
It's like it you can you can define what the world dynamic is, like what what kind of you want the world to be.
53:17
If you want to put stuff in the world, you can do it.
53:19
If you want to texture things differently, if you want to change the skybox, you can do that.
53:23
Um, you know, if you like, so I I think it's um it's I think it'll be a very different way of creating things that's fundamentally a lot more accessible, which will then unlock a lot more creativity and there will just be a lot more interesting worlds and things to do.
53:39
And that I think is going to be important not just for VR and AR, but I think it's gonna unlock all these experiences that billions of people will probably first see on their phones at some point.
53:50
>> So, um, that's what we're doing with the Meta Horizon studio work.
53:53
It's this kind of agentic AI flow where people at different levels of sophistication can go in and create really interesting worlds and immersive environments.
54:05
And um, I think that's neat.
54:10
Then we paired that with Meta Horizon Engine, which is basically this this custom graphical rendering engine that we've been creating for two years now.
54:19
Um, it's this project that we've had for to build it from the ground up because, you know, previously we were using Unity, >> which is great, but it's not really built for this use case.
54:28
I mean, most games, you know, you you load a game, it takes, you know, 20 seconds to um to kind of get into the game, which kind of makes sense because you're loading this whole 3D world that you now need to be able to interact with.
54:41
But we want the worlds that you can interact with to feel more like jumping between um two web pages or jumping between like two screens within a a native app that's like really fast.
54:58
So the whole like okay it has to take 20 seconds to like page this whole new world into memory was not going to cut it.
55:05
So we rebuilt we we basically built Meta Horizon Engine from scratch um to be this graphics engine that can support rendering these kinds of worlds um with high concurrency and the avatar system and the photorealistic avatars and all this um with just a few seconds of load time.
55:24
So it's more like a website or just a transition in an app.
55:27
And that's the kind of thing that'll make it so that when you're in VR you can jump between worlds easily.
55:34
It's not like some big commitment or some big decision.
55:35
You can just feel free to explore because um because you know it's not like you're going to have to wait 20 or 30 seconds for the next thing to load.
55:43
You walk through the portal, you don't like it, you walk back through the portal the other direction.
55:47
Um and similarly for things like within Facebook or Instagram, having the ability to kind of see a post and jump into a world, that's something that needs to have very low friction to do.
55:58
So the the Meta Horizon engine is this kind of core piece of technology.
56:00
So yeah, I'd say on the um on the metaverse side, this year's announcements at Connect were more about kind of the software foundations than hardware.
56:10
Um >> but you're still committed to the hardware. >> Yeah.
56:15
I mean, the way that we do the hardware is we we don't have it plans that there's a new device every single year. Sure.
56:21
>> It's um we have multiple device lines.
56:24
There's sort of the higherend one where we um introduce some new technology and then we try to get it to be as affordable as possible.
56:30
So, we did Quest 3, then we did Quest 3S, but it's not like every year there's one.
56:35
It's basically um you know, most years there will be a new one of one of those two and then sometimes there's like an off year where we're pretty much just tuning the the software to get ready for the new paradigm. >> Got it. Okay.
56:49
So, Quest is still going.
56:52
>> Yeah, I we're we're we're focused. Okay.
56:54
I I've I've been been saving this.
56:54
I've got to ask you about this.
56:57
You know, there is a tremendous amount of interest in your AI strategy right now, unlike anything I've seen in the tech industry, honestly. And >> AI overall. >> AI overall.
57:06
But I think like what you've done over the summer with the hiring and the the super intelligence, uh, you know, mission that you put out and all of that. >> Yeah.
57:15
>> Um, and we've been talking about AI as it relates to the hardware throughout this conversation.
57:19
And, you know, AI was a part of my demo.
57:21
It wasn't, I would say, like front and center.
57:23
And it seems like a lot of the work you're doing now is to get ready for when it will be.
57:26
And I'm curious, you know, when I was here actually for the for the demo, I got to see the pod of the the new the new lab and see them at work.
57:35
And they're in there cranking like you can tell.
57:37
And um I I would love to know like maybe we can start here like when you decided I need to I need to change things and and why you decided to go about it the way that you did cuz I think that's the thing that people were like whoa like this is this is crazy.
57:52
I think if you're on the inside it doesn't feel as crazy cuz you know the talent market is very small.
57:58
You know it it's kind of rational if you look at the numbers but just the strategy like walk me through like when you were like okay I want to make a change.
58:05
this is what I want to do. Walk me through that.
58:10
>> Yeah, I mean this is an area that I just think it I I think AI and super intelligence are going to be the most important technologies in our lifetime.
58:20
Um, I think it's so important that it sort of demands its own hardware platform, which is a big part of why I'm so excited about glasses because I think glasses are going to be the best um kind of hardware device category um to provide personal super intelligence to people.
58:37
But I think AI is just this incredibly profound thing.
58:41
It's going to change how we run the company.
58:42
It's going to change how all companies run.
58:44
It's going to change how we build products.
58:46
It's going to change what products are possible.
58:50
Um, change how creators do their work, right?
58:54
So, change the content that that's possible.
58:55
Um, the mix of content, all these different things.
58:58
So, um, I think being on the frontier there is really critical if you want to continue just doing interesting work and pushing the world forward.
59:09
I think know just like with mobile you didn't you know if you didn't invent the mobile phone you could still do interesting work building apps >> but I I I do think it's some level you can do even more interesting work if you can both pair the software with the hardware experience.
59:27
So >> um so we are definitely committed to being at the frontier and building super intelligence.
59:38
I think it's going to be the most important technology like I just said and um and because of that you know I I just think we we're basic we're we're we're very focused on making sure we build a leading effort.
59:49
So over the last few years we stood up um an effort that was improving very quickly. Right.
59:56
So Llama was a good initial academic project.
59:58
Um Llama 2 was a kind of good initial version of that as an open source release.
1:00:05
Llama 3 was a big improvement over Llama 2.
1:00:07
Um, and then Llama 4 introduced some important improvements over Llama 3, too.
1:00:12
But, um, I didn't feel like we were on the trajectory that we needed to be on to to basically like be at the frontier and be pushing the field forward.
1:00:22
And that was so I I you know I like I think every company at some point um goes through periods where you know you're not on the trajectory that you want to be on something and these are decisions that you get to make in your life or in building a company where the real question is not like it's not like is there going to be a moment where you feel like you're not kind of on on the track that you want to be on.
1:00:47
It's what do you do in that moment?
1:00:48
Um and so I I just decided that you know we should take a step back and build a new lab and um you know I think part of that was informed by um by the shape that I thought the effort should be.
1:01:04
We have this real focus on talent density, right?
1:01:06
And the idea is that you really want to have this is like a group science project, right?
1:01:15
So you want to have the smallest group of people who can fit the whole thing in their heads at once. >> Mhm.
1:01:22
>> And >> and there's not many people who can do that.
1:01:25
>> No, but you also want the group to be as small as possible. Right.
1:01:26
So there are some problems that we're working on around the company where like you can just have more people work on them and even if the marginal productivity per person declines, you can just keep on scaling the net productivity of the effort.
1:01:39
Um uh you know our our feed and ads recommendation is an interesting example of this where we have a lot of people are just testing different improvements to the systems and if one guy sitting next to you um if if if that guy's experiments don't work that well doesn't necessarily slow you down that much.
1:01:55
Um but I think building these language models um is not that way right it's like it's a it's a small group effort.
1:02:05
um you want the smallest group of people that can keep the whole thing in their head and do the best work that they can.
1:02:11
So each seat on that boat is incredibly precious and in high demand.
1:02:15
Um you also don't want a lot of layers of hierarchy.
1:02:21
Um because when someone gets into management, their technical skills kind of start decaying pretty quickly.
1:02:25
of start decaying pretty quickly. even if they were an IC um you know a researcher a few months ago you know now if they're spending all their time kind of helping to manage then you know okay after >> after 6 months a year they they might be less in the technical details than they
1:02:39
were before so you kind of I think that there's this huge premium on just having a relatively small extremely talent dense effort that is organized to be quite flat um >> and you're very hands-on with this team >> well yeah I mean in the sense that I I mean, I'm not like an AI scientist, but the thing that >> they're sitting near you. I mean, it's
1:02:59
I mean, it's clear that this is like the priority.
1:03:02
>> So, the the thing that that I'm focused on is one getting the very best people in the world to join the team.
1:03:07
So, I've spent a lot of time just meeting all of the top researchers and folks around the field and getting a sense for who I think would be good here and um who might be at a point in their career where we can give them a better opportunity. Um, that's one piece.
1:03:24
Another thing that I'm very focused on is making sure that we have um significantly higher compute per researcher than any other lab, which I think we are just way higher on compute per researcher than than any other lab today.
1:03:38
And you know, as the founder and CEO and because we have a strong business model that can support this, I mean, we make like, you know, a lot of profit. Um >> a decent amount. Yeah. >> Yeah.
1:03:48
It's a it's a it's a reasonable amount.
1:03:49
Um >> you can just call up Jensen and be like, "More GPUs, please."
1:03:52
Um, >> it's not that simple. >> It's not that simple.
1:03:54
Um, and I normally text him with my glasses, but no, but it's um >> but uh there's there's a whole supply chain that goes into it.
1:04:03
I mean, the GPUs are one part of it, but then you also need to build data centers and get energy and get the other pieces and get the networking and like all so um and yeah, but the bottom line is we're very committed to that and doing what we need to do to make sure that we have leading levels of compute.
1:04:19
So we talked about recently how we're building um this Prometheus cluster which I think is going to be the first um kind of single contiguous gigawatt plus cluster for training that I think has been built um in the world.
1:04:33
We're building this Hyperion data center in Louisiana that I think is going to scale to 5 gigawatts um over the coming years and um and several other of these we call them Titan data centers.
1:04:47
they all have different Titan code names and um that are each going to be you know one to multiple gigawatts and um that's a significant investment.
1:04:59
I think it took a fair amount of conviction.
1:05:01
So I think unless you're um I think like a bunch of conditions need to be met like basically you need to have a business model that can support it.
1:05:12
you need to have a CEO who like believes in this very deeply, right?
1:05:15
It's that they're just willing to like make that kind of investment for that.
1:05:19
Um, and then you need to have the technical ability to actually go build the things and bring them online.
1:05:26
>> And I think we're one of, if not the only company in the world that meets all of those criteria.
1:05:34
So, um, so yeah, um, I mean other people will do other interesting things, too, but no, I think that this is going to be very interesting.
1:05:42
The the other principle though that we have for the lab is um you know it's it's split into different efforts right there's the the lab that we call TBD which um it >> that's what I saw.
1:05:54
>> Yeah that's the that's the research lab.
1:05:57
Um TBD was it was the placeholder name but then it's stuck because it's kind of a good vibe right it's like all right it's like a work in progress type type vibe.
1:06:04
Um then we also have applied research and product um that in Nat Friedman's group and that team is working on a lot of research that goes directly into the products.
1:06:14
Um so things that may not necessarily directly be on the path to super intelligence like um you know speech that passes the touring test and things like that but but are important for the products nonetheless.
1:06:26
Um so we're working on all those things.
1:06:30
Um and the research effort, the TBD effort is truly a long-term research effort.
1:06:36
So one of the principles that we have for the lab is um is just is no deadlines, right?
1:06:42
So when so people are always asking, okay, when is it going to when are we going to ship the model?
1:06:46
Like my my um this is very kind of it's a strategy and it's also the values that we're trying to put in it is um I mean all these researchers are very competitive.
1:06:58
They all want to be at the leading edge.
1:07:00
Um, they know the industry is moving quickly.
1:07:02
They're going to put a ton of pressure on themselves.
1:07:04
Me telling them that something should get done in 9 months or 6 months or whatever isn't going to help them do their job.
1:07:11
It's only going to put another artificial constraint on it that makes them suboptimize the problem.
1:07:13
And I want them to go for kind of the full thing.
1:07:15
I it's like we're going for trying to build AI that can um improve itself responsibly and that can and where we're basically building these models that bring all these modalities together to um deliver the kinds of experiences that we're talking about and um yeah I I mean I think me me putting a deadline on that is not going to be helpful.
1:07:38
So yeah, so I'm very focused on and that's the nature of research, right? It's not engineering.
1:07:41
It's not like engineering is when you know how to do something and you go and and you you need to put together a complex process to build it.
1:07:48
Research is when there's several >> unknown problems.
1:07:50
Um in AI, I don't even think we have a sense of how many unknown problems there are.
1:07:55
Um >> for something like glasses, we have a sense.
1:07:58
It's like, okay, there's like 10 areas of unknown problems that we need to go solve.
1:08:02
Like how do we get the right wave guides?
1:08:04
How do we get the right laser display?
1:08:05
No one's ever done this, but like we can try 10 different things in each one of those and kind of run it forward.
1:08:10
And AI, I don't think anyone can definitively tell you how deep the problem space is.
1:08:14
Um, so it's very much research and >> that's fascinating.
1:08:21
>> Um, so yeah, so what do you do to to to do that as well as well as possible?
1:08:24
You get the very best team, talent density, um, make sure that people have the res resources that they need and um, and clear all the other stuff that comes from running a big company out of the way.
1:08:35
and that's kind of my job for them. >> Yeah.
1:08:36
You were talking about the capex and the data centers.
1:08:38
You obviously see something on the other side of that that will warrant that being worth it.
1:08:42
Um but I'm wondering do you ascribe to these these bubble fears at all that people are talking about that we're in this massive uh overspending getting ahead of the skis bubble and maybe a company like metal will be okay because you guys do have a core business that makes a lot of money.
1:08:59
But >> yeah, >> how do you think about this bubble talk that has been going on for the last few months especially?
1:09:06
>> Um >> I mean I I think it's quite possible.
1:09:09
I mean I think basic if you look at most other major infrastructure buildups in history you know whether it's railroads or um fiber for the internet you know in the dot bubble um these things were all chasing something that ended up being fundamentally very valuable.
1:09:28
In most cases it ended up being even more valuable than the people who were kind of pushing the bubble thought it was going to be.
1:09:33
But in in each in at least all of these past cases um the infrastructure gets built out, people take on too much debt and then you hit some blip whether it's some macroeconomic thing or maybe you just have like a couple of years where the demand for the product doesn't quite materialize >> and then a lot of the companies end up going out of business and then the assets get distressed and then it's great opportunity to go buy more.
1:09:59
So I think that that it's not it's obviously impossible to predict what will happen here.
1:10:06
Um there are compelling arguments for why AI could be an outlier and and um and basically just um you know if the models keep on growing in capability year-over-year and demand um keeps growing then maybe there is no collapse or some or something.
1:10:25
Um, but I do think that there's definitely a possibility, at least empirically, based on past large infrastructure buildouts and how they led to bubbles that um that something like that would happen here.
1:10:42
Um, from Meta's perspective, you know, I I think the strategy is actually pretty simple.
1:10:50
It's um you know it's at least in terms of building out the infrastructure you know it's no one knows when super intelligence is going to be possible.
1:10:59
Is it going to be three years?
1:11:01
Is it going to be five years?
1:11:01
Go be eight years whatever.
1:11:02
Is it never going to happen?
1:11:03
But but I don't think it's never going to happen.
1:11:05
I I I'm more ambitious or optimistic.
1:11:06
I think it's going to be on the the sooner side.
1:11:08
But let's say let's say that you weren't sure if it was going to be three or five years like in a conservative business situation.
1:11:17
Maybe you'd like hedge building out your infrastructure because you're worried that if you build it out assuming it's going to be three years and it takes five, then you've lost, you know, maybe a couple hundred billion dollars or something.
1:11:29
I mean, my view is that >> that's a lot of money. >> Well, no.
1:11:34
Well, I was going to say in the grand scheme of it is it is objectively a huge amount of money. Yeah. Right.
1:11:40
>> I mean, didn't you just tell Trump you were gonna spend like 600 billion? I mean, that's >> I did. Yeah.
1:11:43
Through 2028, which is uh >> that's a lot of money. >> It is.
1:11:46
And and if and if um and if we end up misspending a couple of hundred billion dollars, I think that that is going to be very unfortunate obviously.
1:11:57
But what I'd say is I actually think the risk is higher on the other side.
1:12:00
If you if you um build too slowly and then super intelligence is possible in 3 years, but you built it out assuming it would be there in 5 years, then you're just out of position on what I think is going to be the most important technology that enables the most new products and innovation and value creation and history. So, um so I don't know.
1:12:25
I mean, it's um I don't want to be kind of cavalier about it.
1:12:28
I mean, obviously, these are very large amounts of money and we're trying to get it right, >> but I think the risk, at least for a company like Meta, >> is probably in not being aggressive enough rather than being somewhat too aggressive.
1:12:44
But, but part of that is like we're not at risk of going out of business or something like that, right?
1:12:50
if you're one of these um you know companies that like an open AI or an anthropic or something like that where you know they're raising money as as the way that they're funding their buildout and you know there's there's
1:13:05
obviously this open question of to what extent are they going to be able to keep on raising money and that's dependent both to some degree on their performance and how AI does but also all these macroeconomic factors that are out of their control. I mean the market could
1:13:14
I mean the market could get bearish for reasons that have nothing to do with AI.
1:13:18
Maybe you know something bad happens internationally and um >> and then it could just be impossible to fulfill the kind of the the compute funding like the compute obligations.
1:13:27
So I think that's that's it it might be a different situation if you're in one of their shoes.
1:13:33
But um but I think for us you you um I think the clear strategy is it's it's just you I think it's creates more value for the world.
1:13:44
Um if we kind of assume pretty aggressive assumptions about when this is going to be possible and takes some risk that maybe it takes a little bit longer.
1:13:56
>> Do you feel like the US is in a better place now to help with this and to help American companies succeed?
1:14:01
I think you've you've done a lot of work with this new administration and when we were here last year you were saying you wanted to kind of stay out of it.
1:14:08
You didn't but seems like you realized is it it was the realization that this is just so important that I I have to play ball with this.
1:14:16
>> Oh well I mean the thing that I want to stay out of is partisan politics. Okay.
1:14:20
So um but I mean our we will always want to work with and have a good partnership and collaboration with governments right so um >> and that's going to be especially true in our home country but it's you know also true in in other countries around the world where we >> serve large amounts of people >> um the so yeah but but I'd say yes I mean I think the this administration for a number of reasons is definitely more
1:14:52
forwardleaning on wanting to help it build out infrastructure and um and that has I think been positive and I think that that aspect of this is um I think the next few years are going to be very important for the AI buildout
1:15:09
and the AI infrastructure buildout and I think having a government um which it's important both at the federal level and in the states where you work that want that buildout to happen is fundamentally a helpful thing. >> Yeah. Um there was a line in your super >> Yeah.
1:15:23
Um there was a line in your super intelligence memo that you wrote where you said uh over the last few months we have begun to see glimpses of our AI systems improving themselves.
1:15:35
>> Yeah, >> I was really interested in that line.
1:15:37
What specifically did you see that made you write that?
1:15:39
Well, I mean, one of the early examples that we saw was a team that was working on Facebook that took a version of Llama 4 and made this autonomous agent that could start to improve parts of the Facebook algorithm.
1:16:00
and it basically checked in a number of changes that are the type of thing that like a mid-level engineer would get promoted for >> really.
1:16:12
>> So yeah, so I think that's like very neat.
1:16:16
It's like you basically have built an AI that is building AI that makes the product better um that improves the quality that people observe.
1:16:25
Um, to be clear, this is still a uh a low percentage of the overall improvements that we're making to Facebook and Instagram.
1:16:37
Um, but I think it'll grow over time.
1:16:40
So, that's what that's that's um one of the things that I was talking about when I said glimpses, right?
1:16:46
It's I mean, this isn't like okay, the AI is improving itself at like an exponentially fast rate or something like that.
1:16:52
It's I I think that what we're seeing are early examples of AI's autonomously improving AI in ways that are having a positive impact on the experience that people get to have.
1:17:02
>> And that's how to think about super intelligence broadly is that it's you're when you're there, it's AI that is rapidly improving itself. That's what it means. Or is that too simple?
1:17:13
>> Yeah, I think AI that can improve itself >> and >> and that's that's that's beyond human level.
1:17:18
I think that there's there's this dynamic today where all the AIs are trained on data and knowledge that people have produced.
1:17:28
So a lot of the systems seem to be kind of very broad and have all the knowledge of humanity.
1:17:33
Um so maybe that's in >> some dimensions it might feel like it is smarter than any one given person and in in kind of the breadth of what it knows.
1:17:42
But um but I still think today the systems are basically gated on human knowledge and um there is a world beyond that right where I think you're starting to get into that with some of the thinking models where they can go out and solve problems in the future that no person can solve and then can learn from having solved that problem.
1:18:05
and and um the pace of that improvement to me is somewhat less important than um than just the process of it.
1:18:15
I'm not like a super fast takeoff believer because um in the sense that I don't think it's like going to be okay one day it can improve itself and then the next day it's going to take over everything.
1:18:26
it's going to take over everything. Um I mean I I think that there's like way more physical constraints like it takes time to build data centers like if you a lot of frontier human knowledge comes from empirical experimentation right so you um >> you know if you're you develop some new drug you want to see a if it works and b
1:18:49
if it's safe for people how do you do that you run a test where you give it to a handful of people um maybe more than a handful but some some statistically significant group of people and you observe how that goes for a while to see both whether the positive effects are are kind of longived and whether there's any negative effects. Okay. Okay.
1:19:11
Well, if you're trying to run a six-month or a 12-month trial, you can't do that in less than six or 12 months.
1:19:16
I mean, maybe you can get a negative result sooner, but you're but you're not going to be able to validate that you you can get the positive result that you're looking for without having done that test.
1:19:27
So I think um I think that's also going to be true with AI, right?
1:19:32
There are going to be some things that maybe something that's like a super intelligent system can just intuit it or reason from first principles using the knowledge that we already have.
1:19:43
But I think a lot of learning is going to be experimental.
1:19:46
And I I do just think these things take time, right?
1:19:48
It's you have to run long-term experiments if you're trying to make long-term changes in the world.
1:19:53
And that I think is going to be true for the AI too.
1:19:54
Now, maybe it'll on average run smarter experiments.
1:19:56
Um, so per experiment, maybe it'll learn more.
1:19:59
Um, it'll I think it will probably be able to figure out some things from first principle.
1:20:06
It will definitely be able to figure out a bunch of things from first principles. But, but I don't know.
1:20:09
I I'm I'm not like I'm I'm not on the the the camp of people who think it's going to be like overnight this changes.
1:20:16
I think it's going to be this very steady progression.
1:20:20
We're just making our lives better. >> All right.
1:20:22
Well, uh, it's gonna be a wild few years.
1:20:24
Yeah, Mark, I appreciate you doing this. >> Yeah, happy to.
1:20:27
Yeah, congrats on the on the new show. >> Thank you. Appreciate it.
1:20:30
>> Alex, I enjoyed that interview.
1:20:30
Uh, but I have to ask, how does it feel to have Mark Zuckerberg in your neurob?
1:20:38
>> I hope he's not actually in the band, but I guess we we don't know for sure.
1:20:41
But seriously, thanks to Zach for taking the time to be the first guest on Access.
1:20:45
You can read more about what we talked about in my newsletter, sources. news.
1:20:51
And uh Ellis, what are you plugging?
1:20:54
>> Yeah, you could find me on Twitter, which I will continue to call Twitter at Hamburger and uh at meaning. com.
1:21:02
>> Access is produced in partnership with Vox Media. Please follow us. We're a new show. We need your support. We need your follows.
1:21:10
Hitting that notification button to get new episodes.
1:21:12
You can find us on Spotify, Apple Podcasts, all the other podcast apps that I don't know about.
1:21:19
We're also on YouTube and video.
1:21:22
Please check us out there at AccessPod.
1:21:25
You can also find us on all of the socials at AccessPod.
1:21:30
>> Smash that like button. Smash it. >> All right.
1:21:33
>> Can't wait to tell my friends.
1:21:36
>> Alice, that's our that's our first episode.
1:21:37
Uh we'll see everyone next week.
1:21:43
>> Support for access comes from Warp.
1:21:43
Warp is a topranked coding agent and one of the fastest growing AI companies.
1:21:50
Trusted by over 700,000 developers and 56% of Fortune 500, Warp is redefining how software gets built.
1:21:56
It understands your machine stack and codebase.
1:22:00
Combining the power of a terminal with the interactivity of an IDE, giving you a tight feedback loop with agents to prompt, review, and ship production ready code.
1:22:08
Find out why developers save an average of 5 hours a week.
1:22:11
Try Warp Free or Unlock Pro for just $5 at warp. dev/acess.
1:22:23
Thanks to Notion for their support of access.
1:22:25
What does Notion and their Notion agents actually do?
1:22:28
How about track beats, sources, open record requests, and fact check?
1:22:32
An agent can spot conflicting dates, flag missing citations, and even tee up follow-ups.
1:22:39
And it could also even run the show.
1:22:41
guest research, run of show, edit checklist, asset library, and delivery.
1:22:46
You could tell agent to prep for Thursday's taping, and it'll assemble a brief from notes and links, fill the rundown template, and ping everyone who needs to be in the studio. That's useful.
1:22:55
So, yeah, it can do all that and more, whatever your workflow.
1:22:57
See for yourself and try Notion and agents free at notion. com/ox.