0:00
You're never going to be able to fully replicate the feeling of two or three of us against the world.
Search this video
0:03
We have X amount of dollars and if we don't manage to make something work, we're going to have to go wind it down, return the money to investors.
0:11
>> you and Kevin back in the day. Exactly.
0:12
What's Anthropic's next big product after Claude Code?
0:14
This week on the show we've got Mike Krieger, the co-founder of Instagram, who now leads the team at Anthropic working on its moonshot bets.
0:24
One of the the dark days of my Anthropic tenure was naming it 3.
0:26
5 V2 and I can tell you why we ended up that way.
0:28
Mike and I recorded this conversation in person on the sidelines of Anthropic's recent Claude Code conference in San Francisco, where the company announced a big new compute deal with Elon Musk.
0:39
>> Hence you guys are apparently going to space with Elon. Exactly, right?
0:41
Yeah, finding new sources of unexpected sources of compute.
0:46
>> We get into what Mike is working on now, Anthropic's fierce competition with OpenAI, and the human work that Mike thinks will continue to matter even as AI gets better and better. This is Access.
1:01
Mike, it's good to see you here at Claude Code in SF.
1:03
Uh I was thinking back to our last conversation and you were more recent in your Labs tenure, but it's been a what, few months now?
1:12
>> Yeah, it's been almost 4 months now.
1:13
>> For almost 4 months now.
1:13
For people who don't understand Labs, I'd actually like to start there cuz it's a pretty unique setup.
1:18
We were talking about it uh when I was at the office uh a couple months ago. What is it?
1:22
What is the mission of of Labs at Anthropic?
1:27
>> Yeah, in a nutshell, the way I think about Labs.
1:29
And Labs this I would call this Labs V2 and we can go into what Labs V1 sort of did and and and what Labs V2 is meant to do, but um I think of it as doing two things.
1:37
One is closing the gap between what Claude is theoretically capable of and how that actually shows up for people in their kind of everyday lives, right?
1:45
So, what products or prototypes or projects do we build that sort of illustrate how you could unlock more of that sort of potential, like close more of that of that gap.
1:55
We can unpack what that means.
1:57
And the second piece is we're the sort of advanced scouts figuring out where the model is going to need to go in the future to meet different needs for people.
2:05
And so, you know, a successful Labs project might look like a prototype that we build and then say, "You know what?
2:12
The models actually aren't good enough to do this yet.
2:13
Let's put this on the shelf and we'll revisit it, you know, next model release."
2:17
Or we'll use that as an evaluation for future model development and basically iterate from there.
2:23
So, what, you know, versus you know, a product lab at a pure product company where maybe the success criteria is did you ship a product or not?
2:31
There's other ways in which this Labs initiative can sort of inform the direction of of where Anthropic is going.
2:37
And Labs has produced some hits, right? It's Claude Code is one. MCP. What are the others?
2:44
>> Agent skills is another, you know, big piece from from Labs as well.
2:46
And then maybe I'll talk about one other one which was one we didn't ship at the time but actually was one of those research informing ones which was computer use. So, Right.
2:54
Um, so I joined Anthropic May 2024.
2:57
Next week is my two-year anniversary, we call it.
3:00
Um, Is it anniversary or is it >> anniversary? Yeah, everything's ants.
3:04
Uh, which I resisted at first and now I just embrace it.
3:05
So, we don't dog food, we ant food.
3:07
Um, this is my two-year anniversary.
3:09
Um, and uh, when we started, you know, one of the I we were spinning up Labs and one of the first projects that was pitched was, "Hey, why don't we try having Claude use a computer?" Um, and do computer use.
3:21
And >> in the Sonnet what era, I mean? This was Sonnet 3. 5.
3:25
Um, it's the first model I was sort of part of launching with Sonnet 3. 5. That was my third week.
3:32
Like, we joke that Anthropic doesn't do onboarding projects, they just stick like a very hard project with you and mine was like, "All right, great.
3:36
Week three, like, launch a ship a model." Um, and and actually 3.
3:41
5 was a really interesting model cuz it was one of the first ones that really unlocked some coding use cases, not quite full agentic coding, but, you know, the beginnings of that. So, um, we put Sonnet 3.
3:51
5 and we built a computer use product around it and um it had a lot of problems.
3:56
Like it was too slow at doing the computer use piece.
4:00
It was not accurate enough.
4:00
Like it couldn't see very well and so it would see the the screen and say, "Oh, I need to click on that button."
4:06
and then it would click on something else.
4:10
That was actually very helpful to have built that sort of not quite working harness because then when Sonnet 3 at the time we called it 3. 5 V2. We can get into that.
4:19
That was That was one of the dark days of my Anthropic tenure was naming it 3.
4:23
5 V2 and I can tell you why we ended up that way. Okay.
4:25
Um and then now we've reckoned it to 3. 6. Try try 3. 6.
4:30
Still not quite good enough. Some signs of life. And then 3.
4:33
7, I remember this day really vividly cuz I was in New York on on on a trip on trip to visit the New York team um and I got a ping like, "Hey, we think this actually that thing that we had built in labs that that point was 9 months old is showing real signs of life with Sonnet 3.
4:48
7 and we think it's time to actually open up computer use as a capability that we talk about externally."
4:53
That was 9 months of, you know, trying out new models every, you know, couple of months and trying it in the harness.
4:58
So, even though at that point labs had kind of let that project go by the wayside, still was very useful in being an evaluation set for how, you know, computer use is evolving um in the models as well.
5:09
When you joined, I mean you were chief product officer when you first joined and I remember going, "Huh, Mike who, you know, co-founded Instagram who I think of as a very consumer forward founder is joining the enterprise AI company." Yeah.
5:22
Uh and we may even talked about it at the time but I was just like, "That's an interesting bet."
5:28
In hindsight, a correct bet.
5:30
Uh and you obviously the timing was incredible.
5:33
Um I'm curious, you know, from joining as CPO where you're overseeing all of product and product and AI is kind of a nebulous term and it changes quite a bit, right? >> Yeah.
5:42
Uh how did you get to fourish fiveish months ago going to labs where my understanding is you're you're an IC.
5:48
Do you actually manage anyone?
5:52
>> Yeah, I've never actually entered our performance review cycle and That's what you want to You want to get out of performance reviews. >> Exactly.
5:56
I was like I went into the portal, you know, where you get your list of all the reviews you're going to have to write and it was like you're supposed to write your own review and review your manager review. And that's it.
6:08
Now Claude does the reviews.
6:09
>> Claude does Claude actually does help with reviews which is which is helpful.
6:12
You know, it won't write all of it for you, but it's useful when you're even being like, "What did What did I do the last six months?"
6:15
Um I think the company's been through phases and and it's at different parts of it.
6:19
I think it's been more or less suited to like what truly drives me.
6:23
So when I joined, all of product engineering was maybe 30 people maybe split evenly across the eight >> Product and eng?
6:31
Uh like product engineers.
6:32
So we had an engineering team that was working on research infrastructure and and some scalability pieces, but in terms of like core people working on building at the time Claude AI and what we called the API.
6:41
We didn't call it the Claude platform at the time.
6:45
Maybe 30 people, maybe 35, like very very small.
6:47
Still felt very much like in the beginning and there was still so much I still think there is, but there was so much at the time shaping of what the product even meant.
6:54
You know, at the time Claude AI had no projects, no artifacts.
6:59
It was basically a list of conversations that you had with Claude with effectively no frills on top.
7:02
So it felt like joining a startup that was trying to figure out, you know, it had tailwinds for sure, you know, at the time I joined, uh the kind of Opus Sonnet and Haiku three series had come out.
7:15
And that was the first series where Anthropic had put out something that was sort of at least frontier adjacent or you know, close to the frontier.
7:24
So much left to do in terms of like, "What is this product going to be?"
7:26
And the reason that, you know, even though my background was more on consumer, the reason I got very excited was um the time between Instagram and Anthropic, one of the things that Kevin, my Instagram co-founder, and I did is spend a lot of time doing investing.
7:37
And we had a whole sort of set of investments that we were looking at in terms of what is the future of work and how is work going to get done.
7:44
And Anthropic felt very much like it could help unlock a lot of that kind of that theme, right?
7:49
Which is what happens when you have a really intelligent assistant helping you.
7:53
And I don't think I even foresaw how much how transformative it would be at the time.
7:58
>> "Oh, this interesting little AI company.
7:59
This may help with my investing." >> Right.
8:00
This might help with This might help with some of the you know, some of the themes that we've been thinking about.
8:04
But of course it was like it transformed it much more.
8:05
So, that was phase one where tiny team, you know, you could count the number of initiatives you had going on one hand.
8:11
Then we can talk to some of the different chapters in depth, but you fast forward it to the end of last year and the product team is hundreds of people.
8:18
We have a portfolio of projects and a lot of the work is deploying, understanding, customer facing management layers, all of these things that just learned over time.
8:32
There are people that love it and thrive it and I am super respectable of those and it's one of mine where my I have a great coach and she calls it your zone of competence.
8:39
It's like the things you are good at and like you're fine.
8:43
Like you're probably doing well, but it's not the thing that ultimately like fuels you and drives you and that's a good that's a almost the most dangerous place to be because you could keep doing it for a very, very long time, but it is not your place of maximum sort of like fire, right?
8:56
And so I entered Q4 of last year and started having conversations with Danielle and I like this you know, we've grown up already.
9:03
We've we've speed run a lot of the growth that usually would take five years, you know, so even though it'd only been about two years or Yeah, I'd say growth has been decent.
9:09
>> Yeah, growth has been you know, and the team growth and the product portfolio I was like, "I think I want to go and do another a new company."
9:15
And she was like, "Well, you know, is it about leaving Anthropic or is it about shifting what you want to do?"
9:20
And I was like, "I love the company.
9:21
Like the people here are amazing.
9:23
I love the the technology, the mission, etc."
9:27
And so that was also when we were spinning labs back up cuz labs we weren't so successful that all the projects graduated and left nobody behind.
9:33
And so we kind of wound it down and had it kind of on ice.
9:36
So, let's spin Labs back up and I'm going to go back to being a builder and it's, you know, everybody who sees me in and out of work is like, "Mike, you look so happy."
9:45
And it's >> Some of your colleagues were telling me that earlier today.
9:47
Yeah, it's I mean, Mike's Mike's having the best time Yeah, and you know, it's I still I have I am my own harshest critic, so I'm always every day I'm like, "How do I like, you know, what can we do? What can we build?
9:59
What what are what are we proving out here?"
10:01
So, it's definitely not easy, but it is definitely much more aligned with like the things that truly drive me.
10:06
>> We don't have to spend too much time on this, but I am fascinated by tech companies doing these kind of moonshot zero to one Labs things.
10:13
I mean, Alphabet's probably the greatest example, but I mean, history is littered with these and some of them work, some of them don't.
10:19
V1 of Labs at Anthropic certainly worked if you look at the products. >> Yeah.
10:23
I imagine that puts a lot of pressure on you right now to go, "Okay, I Claude Code is like the thing that I have to live up to." Right.
10:29
Well, it's funny because one of the like sort of internal mission statements I have is like, "Oh, we want to find the next Claude Code."
10:36
Which um was already a high bar when we started Labs V2, but it since then it's like grown so it's like even higher.
10:42
So, um I think there's a few things that you need to do like and I think it's helpful that I've had the startup experience a couple of times because you're never going to be able to fully replicate the feeling of there's two or three of us against the world.
10:54
We have X amount of dollars and if we don't manage to make something work, we're going to have to go wind it down, return the money to investors, you know.
11:02
>> That's like you and Kevin back in the day. Exactly.
11:03
And like that I remember that existential sort of question like lived with me every single day. Like, this doesn't work.
11:10
Like, you know, I'm not going to go have this independence.
11:13
I'm not going to be able to work on this thing that I really care about.
11:16
That's hard to manufacture inside a company unless you want to create some very complicated structure.
11:19
And I've seen that as well.
11:21
Like, before we kickstarted Labs V2, I went through and studied all as many of these sort of company Labs initiatives and you had ones where they get equity in the thing that they create or, you know, and all of Which all of that still does.
11:34
Yeah, and and, you know, they're all trying to patch around the sense of it.
11:36
patch around the sense of it. The thing that we have found works well, one of because I think Anthropic has attracted just a lot of really, um, high agency, ambitious, but also mission-aligned people, is that the thing that to watch out for is less,
11:50
make sure that people are incentivized to like do their best work or, you know, put, you know, everything they can into the project, and it's more, how do you make it not just a comfortable place for okay ideas to get like pulled out for months at a time. And so, the way we've
12:02
And so, the way we've gone about doing this is actually shortening the timeline.
12:06
So, Labs V1, we were on effectively four to six-week review cadences where a project would be sort of funded, and then they would have about six weeks to prove some sort of things.
12:17
Here, we actually run two-week sprints, and every two weeks, um, every project goes in front of the whole Labs leadership.
12:22
And it's not like a murder board-type meeting.
12:24
It's usually, you know, a reasonable meeting unless the project is really going off the rails.
12:29
But, we really hold every project, and we truly do, to this question of, what have you learned the last two weeks?
12:35
And have we learned enough?
12:35
Like, sometimes you've learned enough in the project, you say, that project was great.
12:39
Um, it did what it needed to do, and we don't actually need to spend another two weeks on it.
12:43
And right now, things are moving so fast, and you can build so quickly with these models, that the opportunity cost of letting a project languish for an extra four weeks is actually quite high.
12:51
And that's That's I think the best thing we've done with Labs V2 is is better sort of, uh, turn the crank so that we're at least measuring our learning, if not the external impact, every two weeks.
13:03
And most of the people in Labs are former founders?
13:06
A lot of them are former founders, yeah.
13:08
So, that's a criteria you have?
13:09
For We have sort of two roles in Labs.
13:11
There's the We call them all bets, cuz they're all, you know, pretty high high variance, right?
13:17
So, we have bets, and then we have bet leads.
13:18
So, you know, you if you're like kind of spiritual, directly responsible individual, those tend to be founders almost exclusively.
13:24
And then we have just people on Labs, and those might just be builders that were maybe the first employee at a startup, have seen that zero to one, or just love to build.
13:33
One person who joined, she was early at a her last startup, and then, you know, early to Labs.
13:37
So, she's got that sort of I can span the, you know, stack, I can do whatever needs to get done.
13:43
And that's as important, I think, as the You don't want just pure founders, you also need that complimentary like founding team energy, too.
13:50
My co-host Elas couldn't be here, but he sent me some questions to ask you, and one of them was is this founder supergroup, essentially, with Labs that you formed, is this a new model for company structure in general, especially with AI, or is this a thing that you think is unique to a company like Anthropic?
14:06
I It's a really good question.
14:10
I think there is going to be much more, many more teams like this where, um, a small group with the models, the models are not perfect, so you still need, you know, founders with taste and direction.
14:24
And there's, you know, I spent, just yesterday an example, I spent about two hours debating this Labs project I'm working on, like what the shape of, you know, in this case, like what multiplayer looks like within this product.
14:35
That was time that was really valuable and very human.
14:37
It was all like in a room, us hashing stuff out.
14:39
And then I had Claude work on all the ideas asynchronously for the next like 12 hours, right?
14:44
Um, but to have that taste of what are the things we need to debate up front, what are the things that we should just go and implement, right?
14:51
Because the models will make a lot of decisions for you if you don't specify them.
14:54
So, I think that's still a a useful, very important role that a like a founder type can play.
15:00
Um, but once you get that right structure, you can do a lot.
15:02
Like I was looking, I did a, um, little retrospective of what we built in Labs so far.
15:07
You know, it's been four months, and there were projects that we sort of conceived of in January and like shipped to testing in February, and wound down in March, or conceived of in February, and are now Claude design, right?
15:18
Like there's all these different, uh, things that have already happened.
15:21
And the velocity is really really high um and I think more companies are going to shift to this model of smaller teams, give them autonomy, have a person that feels accountable and sort of can drive that and then don't over staff it to begin with.
15:35
We've really learned that with Labs V2.
15:36
Like you don't want to add five people to a project up front.
15:38
You want to start with two or three much more like a startup.
15:40
This is the fun part where I try to get to tease out for you as much as possible about these projects.
15:45
So in March when we met last you told me that you were applying Claude to running for a really long time.
15:52
This kind of long horizon problem and you were leading that project.
15:54
Are you still leading that project?
15:57
>> that's a great one where we actually put that project on ice but it the learnings graduated into outcomes which we announced today in managed agents.
16:03
So um that whole idea of root brick based uh sort of autonomy for Claude to be able to perform a task towards a end and our goal rather than a um sort of just a individual prompt that you gave it very spiritually based on the things that we had been exploring in the labs.
16:20
And so that I was mapping out exactly that project and where all the tendrils became you know, and it's a couple of two uh themes that you saw on stage this morning that were kind of direct derivatives of that Labs project.
16:32
One was outcomes and the other one in Boris's demo you had um the uh Claude code actually verify its work by taking screenshots and testing out what it was doing.
16:42
That was a big part of what we were pushing on with that project because it's one thing for Claude to look at its code and say, "Looks good to me."
16:48
It's another one for it to fully explore and we in the project I was leading we were even looking at things like having Claude take videos of all the work that it had done and then have it scrub through its own videos and say, "Oh, that animation is broken."
16:58
And it wouldn't have been able to tell just from pure screenshot verification.
17:00
Um so yeah, that was one that we had been chatting in March was uh in development.
17:06
We've since uh basically put it on ice.
17:08
It still runs so internally we still have we still use it for for some demos but it's mostly been useful as a sort of source of upstream energy.
17:14
The other one was Claude choosing its own form was I think the way you put it and introspecting and being like for this next conversation instead of a command line interface, I'm going to be a website. Yeah.
17:25
Was that Claude design or is that something else?
17:28
>> So, there's a lot of spiritual alignment. It is something else.
17:32
Although we're the kind of question we're playing on now is like Claude design is an instance of effectively like agent plus canvas, right?
17:39
right? And you can imagine a lot more of those and even within Claude design, you know, I've done everything from writing technical specs is my new favorite way of writing technical documentation is actually in Claude design because you
17:49
can have it visualized like how does the information flow work and how could it work differently and actually just watch it evolve and that's actually my probably like number two use case behind slides which I use Claude design all the time for. Um but you can imagine other
17:59
Um but you can imagine other formats that could live in that canvas as well and so that's one labsy thing that we're pulling on is like like Claude design but for even more sort of types or applications or things like that and that's I think an exciting Is it just more generalized productivity software? Is that the idea?
18:15
I think that's a Yeah, that's a definite a theme to pull on which is like in productivity software that hopefully can be like pretty custom to you.
18:21
I think that's an interesting interesting trend to think about.
18:25
What are the other just kind of white spaces that you're thinking about right now in AI?
18:31
I mean, I think as the models improve, their ability to be useful partners in things like life sciences, I think is really interesting.
18:39
I've started to see more and more.
18:39
There was a great thread on X of somebody that did their full genome sequencing at home and actually he's now professional professionalized it so he'll come over to your house and do it for you.
18:47
I'm very interested in this because like a you know, self-knowledge geek so I'm like want to figure out how to do this.
18:56
And what I've told from people that like know the space is that even you know, the gap between models three four months ago and something like Opus 47 and being actually useful sort of parsers of you know, genetic data and like finding inferences or ingesting your labs like it's now actually useful versus like, oh yeah, cute.
19:13
Like, yes, that is the thing that a doctor would have said if it looked at it.
19:17
Um or, you know, we already knew from heuristics.
19:19
So, that's one white space interesting area.
19:20
Like, I'm very interested in that whole area of like personalized medicine um and where those things are going to evolve to.
19:28
I feel like we're at the cusp there, you know?
19:29
And I feel like there's still so much um overhang, you know, what we call like the gap between like what's now possible with the new >> all these startups popping up like, you know, Superhuman and Superpower and Ro and like all these companies that are Function Health, I mean, that are doing stuff like this. >> Yeah, exactly.
19:46
And we like in January when we did our our our like initial cloud for healthcare, one of the things we did is we partnered with Function so you can bring your lab results into cloud and do that.
19:54
I started using a uh service this week called Subco for supplements.
19:59
So, you put all your supplements in.
20:00
It um uh it knows uh how they interact and then it kind of gives you a recommendation of like, you know, you don't need to take all eight of these.
20:07
Oh, that's actually really helpful, yeah. >> four.
20:09
And now the next step, which they don't have, but like I could imagine somebody doing is, and, you know, based on your last run of your, you know, labs or based on your um genetics data, we think you might be a super responder to this or that.
20:21
So, that whole area, like, there's there's so much interesting and that's even just on like call those the like optimization side.
20:28
Think about all of the sort of like areas that are like currently underserved by medicine either cuz they don't have like local access to good healthcare and like we can help supplement that.
20:36
So, a lot of interesting white space there that I think is is super interesting.
20:39
The other area that um it's funny, I was on stage maybe a year ago at a consumer AI uh conference um that um Forerunner put together.
20:50
And I was talking to, you know, founders like, what what is the going to be the consumer AI breakout case beyond chat and list assistants?
20:57
And I still haven't seen it, you know, maybe health feels like it's not quite in the theme of what I'm describing, you know?
21:04
And uh like AI-powered dating feels like there's maybe something there but there's kind of a natural that's interesting.
21:13
Yeah, I don't think we'll see a cloud dating service but maybe you know, downstream cloud you know, one of our API customers might create something like that.
21:20
I don't think we would build that one in-house.
21:23
But yeah, like I think it's still interesting white space around can AI actually be useful in helping us either understand ourselves or our world or our communities in a way that is connecting rather than distancing, you know.
21:36
An interesting example I've been getting interested in um civic engagement and how you get representative groups together for civic debate.
21:43
So I'm not expert it's not an area I'm an expert in but I've gotten interested in it and that's where you know, AI could be helpful in sourcing representative group rather than trying to make decisions for you for example.
21:55
Like keep the humans in there but make sure that we're listening to the right voices for example.
21:58
>> wondering if you still have the social media bug in you or not.
22:02
Less the social media bug, you know, and it's probably the most interesting-ish idea in that has been something that they actually explored in in Sora that that was interesting was describe the algorithm you want and we will like degenerate it for you.
22:15
>> That's coming everywhere now.
22:15
I mean that's in threads, X has it to a degree.
22:18
I think that's going to be standardized. Spotify.
22:20
But I think it's yeah, use that in Spotify.
22:22
Spotify I create these really cursed playlists where my daughter and I can't agree what to listen to.
22:27
We just make a Spotify playlist that is the intersection of the two.
22:31
So we end up with like pavement plus frozen music. It's pretty good. >> Okay, I like that.
22:35
Spotify the Spotify AI DJ has has some thoughts. That's great.
22:39
But yeah, I think that that I think is one way in which it's it's personalizing us, you know, an idea I was kicking around more recently.
22:46
I don't think this is quite lab-shaped but maybe somebody should build it is can you know, can AI be a useful sort of like filter of all the noise out there.
22:57
I've started using dispatch and co-work for this reason like I have otherwise I am a like incurable news junkie and I will like check every news site and like read a lot and there's value to that but at some point I'm like am I I'm reading the same stories over and over.
23:09
So I have my like now kind of channeling artifact my previous startup like my daily brief of the places I tend to check and often I'll click through and read those stories but it's a useful sort of I don't feel like I need to go check these 12 websites, you know, first thing in the morning.
23:23
It can kind of give me the trends and I can go deeper on what's what's, you know, what's happening.
23:27
>> as you're still reading the Sources newsletter my key can Exactly.
23:28
That's that goes straight to the inbox. >> yeah.
23:32
I mean, I did actually read I don't know if it was a leak on X or what that you guys are thinking about something more proactive.
23:37
I don't know maybe it's Orbit or something where you may do that what you're talking about more proactively bring things into I don't know if it's co-work or another part of the interface but for the more consumer audience.
23:50
Do you do you think that's an opportunity?
23:52
Yeah, I think that proactive sort of we I anytime you see your power users create those kind of use cases it's an interesting question to ask like what would it mean to build more of that built in and that for sure is, you know, the scheduled daily briefing or you know, proactive, you know, monitoring of either, you know, we're we're a very Slack-heavy shop at Anthropic so sort of proactive Slack monitoring or email.
24:17
I've started doing this as well like where my my morning Claude routine is the news one I just talked about I also have it scan Superhuman which is what I use for my own launched a MCP so now it'll go through and it's great cuz it knows about the stuff that's like kind of like okay, you really should look at this one and like some of these are like things you could read later and it's it's actually what I check first before I open my inbox.
24:38
Oh, and then like there's like some shopping thing that I'm looking at where I'm like waiting for the thing to pop up and like rather than like obsessively check every morning I'm like Claude can go. Is this a sneaker drop? What's that?
24:50
I should I should use it for it's probably not even fast enough for sneaker drops.
24:52
You need something faster than that, you know.
24:56
Or it's all the rabbit holes, you know, and and you know, there's always that balance.
24:58
You don't want to remove the joy of Sometimes it's fun to get lost in internet rabbit hole every once in a while, but maybe do it a little bit more consciously rather than the thing that you're just like noodling on in the background. >> Yeah.
25:08
Um I am a co-work power user. I use it every day.
25:11
Uh I used it to work on this interview actually.
25:13
And I used it to write I used it it for everything.
25:17
And it's been pretty transformational.
25:20
And it it's something that I've like personally gotten my wife on and my friends on and um I do feel like it's still early.
25:26
Like I think most people are probably not doing what I'm doing.
25:30
Uh and co-work is still fairly I don't know.
25:34
Uh from a product perspective it's a little bit divorced still from chat, classic Claude chat, and and Claude code.
25:40
These are three things that I don't think really long-term will be three things. >> with that.
25:44
Um and I think some of your colleagues were telling me when I was in the office a couple months ago was like maybe co-work is the front-end interface for everything eventually.
25:51
Um and co-work was a labs thing before you got to labs, right?
25:56
So I'm I'm but you were CPO on that ship.
25:58
So I'm curious um yeah, what was the insight with co-work when you guys shipped that?
26:02
Like what was there a moment where you were like we have to abstract away the coding for normies and that's an opportunity or was it something else?
26:09
It was So I give Dario a lot of credit here.
26:12
He um you know, Anthropic has such a diverse set of things that we're thinking at a given time like the research side, the product side, the commercial side, compute, policy, societal impact that it's not a typical company where like the product and the go-to-market are the two things that matter and that's like what the CEO is spending a lot of time on.
26:29
So Dario's attention kind of shifts between different areas depending on what's most urgent.
26:34
But um shortly before co-work ship, he he's like, "Look, I've been seeing this interesting trend of people using Claude code for personal use cases.
26:44
It's very hard to do that for most people.
26:46
They don't They're going to get a terminal that feels like that."
26:47
And he's like, "What would it mean to build Claude code for everything else?"
26:50
And that was a very useful and I think that's like he operates.
26:53
He wasn't like, "Here's the mock-up."
26:55
or "Here's the like product." That was your problem.
26:56
It was like, "Yeah, right."
26:58
Like that's the product you felt like, "Go and go and solve that."
27:00
Um and um and it actually took the combination of like two of course that the many people were involved here, but I think of like two personalities that really like combined really well.
27:10
Um one was Felix who was the um creator he was like one of the main Electron maintainers like knows desktop software at least like desktop web software really really well and have been like noodling on this problem of like how do you enable people to do work on their computers.
27:24
And then Boris who has obviously like deep Cloud Code knowledge.
27:28
knowledge. And uh once we put the two of them together plus their teams, you know, working together, it was only a couple of weeks before we were able to ship Code Work because not because all the work happened in those two
27:38
weeks, but because a lot of thinking had been happening both on the Cloud Code side and the desktop side around how do we enable this use case for more than just coders because the models are powerful agentic engines and how do we give that engine to more people? But I
27:50
But I completely agree with you that right now it feels like it's less that we're shipping our org chart because Mhm.
27:57
It kind of feels that way a little bit.
27:59
>> But but we are shipping our um harness strategy or something like in a way that I don't think makes sense to most people as well and I think It's also a bad product experience because like my Code Work thread can't sync to my Cloud mobile app. Yes.
28:12
>> So like I couldn't take the interview prep doc I'd been spending time in Code Work with with me on my phone. >> Right.
28:17
And and some of these projects or products naturally suggest their next evolution.
28:21
Like I'm a huge dispatch power user.
28:23
So dispatch lets you access your Code Work remotely, but you got to run your computer, right?
28:26
So your next logical thing is like, "Well, this would be great if I didn't have to leave my computer on all the time."
28:31
So like how can we evolve it in that direction?
28:35
Whereas Cloud Code has been on that journey, but you know, a year ahead maybe six months ahead of it where Cloud Code remote, you know, I use all the time where I'm you know, out on the go and I can kick off a coding task and by the time I'm back it's often put up a pull request.
28:47
So, you can imagine similar things for co-work. But, I agree.
28:50
We're we're definitely in the uh we let a lot of, you know, things proliferate, which is great for innovation.
28:55
But, now I think there's even for uh like context and continuity value, the fact that you had a conversation here versus here shouldn't be something you should have to think of my wife is all the time like, "I don't remember.
29:05
Did I Was that a coding thing I did or did I ask it in chat or was it co-work?"
29:08
Like, that feels like a broken abstraction that we need to fix.
29:10
And it makes a lot of sense when I'd spend time with the OpenAI folks and their super app strategy and what they're going out with Codex because it is the end state of kind of what you guys got to first with co-work.
29:21
It's unifying all of that.
29:24
And I'm wondering if A, you think that super app approach is the right one for AI?
29:28
And if that puts more pressure on you guys cuz I think OpenAI has realized that this is a huge opportunity to bring coding, the power of coding to people who aren't coders. Right.
29:41
>> going at it very fast. >> Yeah. And Codex is good.
29:42
The new model's good and and Codex is doing well for them.
29:46
Um so, I'm curious, you know, Claude had that huge uh surge with coding and is still considered, you know, widely I think uh leading in code.
29:54
But, Codex is really nipping at your heels, isn't it? >> Yeah.
29:56
I mean, it's gotten I I always had the most interesting times at Instagram when we had like a snap, for example, like as a, you know, uh interesting competitor in the same space but with a different approach in some ways and, you know, in a fast-developing market, you end up like having some ideas that you borrow from them, some ideas they borrow from you.
30:15
But, I think >> So, is Anthropic Snap in this analogy? What's that?
30:18
I don't know which is which is which.
30:19
I mean, we're the the market dynamics are different.
30:22
But, at least like we're Yeah, we're a little different. Yeah.
30:25
And actually, the the approach the the personalities of the companies are very very different.
30:28
Uh but, at least in terms of, you know, the way I always thought about it was um I don't know that every competitive situation resolves to this.
30:35
But, at least that one felt like it did, which is we each had our own strengths and we each saw the future in sort of not exactly the same way, but directionally similar ways.
30:44
And it was a really interesting question of like how each company evolved to get to those different places.
30:50
For Instagram, for example, we were known as like the really precious once a week photo, and we wanted to get to a place where people were more free to share, and that's where, you know, being able to adopt stories really made a difference for us.
31:03
And then, I think for Snap, it was they wanted more of the influencer, celebrity piece as well, because that, you know, there was growing interest in there.
31:11
And it's like how many managing both as well.
31:12
Maybe an overextended metaphor, but I like but I think about for us, like it is really interesting, I think, to have open-ended space as well.
31:20
And I think for us, it's the two things I think are most important.
31:24
One is what we were just talking about, which is like reconciling and making sure that like our product makes sense, you know, whether it's on the web or whether it's on the desktop.
31:32
I don't think it needs to be a single app or super app, but at least that your sort of building blocks all talk to each other and all make sense together.
31:37
And then, two is continue to close that gap, because even co-work or any of the products right now, the gap between, even inside Anthropic, I was just talking to one of our recruiters.
31:47
He is like the most extreme co-work power user.
31:50
He's actually our top co-work user who's not a coder at the company.
31:54
And you see his workflows, and they're amazing.
31:57
But that is not how most of the recruiting team operates, even at Anthropic, right?
31:59
So, that is, I think, whether the shape is super app or not, like that is the goal.
32:04
Is can you, again, back to labs, like that's part of what we need to do is can we create these shapes, the form factors that actually make it so either that one recruiter's knowledge spreads really quickly and it's very easy for everybody else to adopt that workflow, or for everybody to get to that level of sort of confidence in using I was hoping to hear more about co-work today.
32:24
I mean, I know it's a developer conference, but I'm curious like how is co-work growing relative to Claude?
32:30
Claude's it growing faster? Anything on that?
32:32
>> I think index to I don't know relative to like overall coding.
32:34
Coding is growing really, really fast, but definitely on the like trajectory that Claude Code was on or faster, which is really exciting.
32:41
Like I feel like every couple of days, one of our slacks that is more like a productivity oriented will show some like new new milestone of our co-workers and it's like it's very exciting to see that growth.
32:52
I think personally very exciting because I always wanted us to reach beyond kind of pure code in terms of like impact.
32:57
Um so the growth has been really really good.
33:01
And we talked to enterprises and they'll say, "Great, I adopted co-work."
33:04
And I'm seeing that phenomenon of some people really get it, some people still need a lot of hand-holding, different departments are using it differently.
33:10
Um it's a good problem to have because you know the product is capable of the right things in the right sort of shape and with the right person, but there's definitely still a delta there and I think the remaining like it'll accelerate growth a lot further when we can actually feel like for most people it doesn't take a lot of ramp up to get to that level of like, "Oh, yeah, AI is actually really helping me." Yeah.
33:31
Um speaking of competition, this was a conversation you and I had, God, I want to say a year and a half ago at a conference was competing with customers.
33:42
And this was before you guys had launched Claude Code the product.
33:44
I think this was we were and we were talking about, well, you know, you have some of these like Cursor was a giant customer, like how is that going to work?
33:52
And you had to make that call and you don't have to do that as CPO anymore, I assume.
33:56
But as not as not the CPO anymore.
33:58
But um how is that these days?
34:00
Um and we've danced around it, but there was Claude Design and the Figma controversy and you were on the Figma board and you had to step down and Dylan said nice things about you and that didn't seem like it was necessarily you, but uh that was an example where I think the market and everyone broadly looks at what's happening and goes, "Oh man, Anthropic's just like coming for every key vertical." Yeah.
34:20
And that was a the Figma one felt a little dicey.
34:21
It's definitely more complicated than it was when we had, you know, the Claude the Claude AI AI, you know, I came in and I was like, "Well, this is not really competing with anything.
34:32
It's kind of like a brand new form factor.
34:36
Um and we're still, I think, careful and thoughtful about the products that we build.
34:40
And I think it's only when there is something the value that we are bringing by shipping something cuz we have amazing customers on the cloud platform, right?
34:49
Like in a lot of people inside the company who are really invested in these customers sort of, you know, success as well.
34:54
Is are we showing some direction that they think we think the industry could go?
35:00
And hopefully it's like a rising tide where a lot of companies kind of adopt that way of thinking.
35:03
Cloud Code being a great example where the attention wasn't in the terminal.
35:06
It was more in the editor, right?
35:07
And it's really shifted how like almost the entire industry has gone even though, yes, there were people who were thinking about the terminal beforehand, right?
35:13
Like that I think shipping Cloud Code in the way that we did and giving the attention that we did did that as well.
35:20
And I think that's my hope for any of these products.
35:21
It's not that they become the only product in their space. That would be a tragedy.
35:25
I like a diversity of products to exist.
35:26
But instead that they are, you know, one, they make sense within our product portfolio.
35:32
So, you know, the fact that you've got all your MCPs connected in Cloud AI should hopefully make these other products better, for example.
35:37
And two, that they showed the way forward where, you know, with Cloud Design, for example, it was being really agent first and giving the agent like a lot of sort of control over what it outputs.
35:50
It leads to a really specific product experience.
35:52
Like for me, it's really useful for things like, you know, slides where I can kind of think along with the agent but also be producing alongside.
35:58
It's not >> it as cannibalizing Figma? >> different use cases.
36:00
Like I like there is production and there is like refinement and that there is collaboration and that like is a real Figma strong suit.
36:09
I love those folks, right?
36:10
Um I think that like that product is really well tuned for that use case.
36:13
And then there is what I tend to use Cloud Design for cuz I'm not a full-time designer, you know, so I'm doing more like either production of, you know, a, you know, visual communication uh kind of artifact, lowercase a artifact, or some exploration of interactivity where the purpose is not the polish or production, it's the sort of early mock-up feel of it.
36:37
I was mocking something up in our iOS app in Cloud Design, and the pixels looked quite different than the things that we would eventually ship, but it pointed the way at what I thought we should be able to do in that space.
36:48
And so, I think they'll both I mean, use Cloud Design and Figma as an example, but in any of these, like, they're all continuing to evolve.
36:55
And again, I think good ideas from us will hopefully make their way in other places, and vice versa.
37:02
And I think it's more around we're all together figuring out what shape of product do we need to build to make the most of the agents.
37:09
I was having a good conversation with one of our researchers this week, and they were saying, "Look, the more you constrain the agent's behavior, try to be over specific on it, the less, sort of, naturally emergent magic is going to happen in there, as well."
37:23
And like, that was, I mean, Claude Code, I think, channeled that.
37:26
Even Claude Code channeled that a little bit versus Claude AI, where it was more like, "You are chatting.
37:29
You are doing these really specific things."
37:31
Um, and, um, I think we'll build products of that shape.
37:34
And that thing that that primarily does is have more people grok that and build really great experiences into their own products.
37:43
Like, great, cuz hopefully those are Claude-powered, and hopefully Claude remains the best or one of the best at being a great, sort of, agentic underpinning for these, sort of, next-generation AI products.
37:52
Um, but the headline is, it's definitely more complicated, but I I I I think the same principles we tried to apply then, which is, uh, you know, being thoughtful about where we enter, and then making sure that the building blocks are still available on the platform.
38:05
Like, managed agents, like, the things that you can now build on top of managed agents are, you know, as powerful as anything that we can build inside the house.
38:13
And we're not, like, other than for safety reasons, like, not releasing a model because, you know, it gives us an edge.
38:20
That's been a very principled approach that we've taken.
38:21
And then we're only going to ship a fully featured product if we think it says something new.
38:27
So, you and Dylan are good? I love Dylan.
38:30
And I, you know, we He's a former guest, so I got to I got to just make sure I clear that up.
38:36
>> I love Dylan and I like I have so much respect for what they're doing.
38:38
This is another LS question on this topic.
38:41
Everyone, especially in the early-stage community, is debating the viability of, I think, consumer startups more right now, especially with just like what is Anthropic going to do next?
38:50
What's the next Claude design?
38:51
Or what is OpenAI going to do next?
38:53
Um and enterprise feels a little safer.
38:58
Um Dario was talking today about, you know, he thinks there'll be a billion-dollar one-person company this year.
39:03
I bet that would be an enterprise company, but what do you think about yeah, uh the state of consumer startups and if you were yourself when you were starting Instagram with Kevin and you were looking at what's happening, Yeah.
39:17
what would you have done differently now because like the world's changed quite a bit.
39:20
And it's it seems like a harder time Yeah. To be doing this.
39:22
I think a radically oversimplified version of consumer like breakouts is there needs to be some like new capability or form factor.
39:34
Like for us it was the camera phone, right?
39:36
And nobody calls it the camera phone anymore because that's just what phones do. They have cameras.
39:39
Um uh you know, or you know, maybe it was more uh more uh rich media like the TikToks and Reels of the world because there was now the ability to stream, right?
39:51
And then some distribution mechanism that lets you break out.
39:55
And for us it was the fact that it was a little bit wild west still.
39:58
Like you could share Instagram photos to Facebook and Twitter and Tumblr.
40:01
Uh and Posterous just to date >> How old is Posterous?
40:05
>> Posterous when Instagram launched.
40:05
And all of those allowed you to freely link back to your product and so you could create that that cycle.
40:09
Um such a different time these days, right?
40:12
And so when I talked to founders um like Matt who founded Lockit, um Yeah.
40:20
they're you know, they have found ways of like, all right, you're in the conversation on Tik Tok and therefore like that leads you to some interesting kind of growth moments.
40:26
I think it's a much harder distri- like it's a it's a far less predictable sort of distribution um uh ecosystem than we were working with and so I think more than any AI player, I think that's actually probably the biggest trend or biggest challenge right now is how do you break out when you have you know, before maybe you're competing with uh Facebook and maybe their time spent, I'm kind of making this up, was like 15 minutes a day.
40:49
Now you're competing with Tik Tok, which I bet their time spent might be over an hour a day, which is frankly kind of crazy, but it is the way it is.
40:58
You're just like trying to displace um some of that.
40:59
So, if you have those two components, on the former side, you know, maybe it's around some novel interaction with some you know, AI-powered being or the opposite where it's AI helping you get outside and you know, touching grass and having more of those real-world interactions, whatever that is.
41:16
But you still have to solve that second piece.
41:18
And if you think about where a lot of attention is shifting these days, it is to these chat agents and I don't think anybody has like really figured out what it means to What does it mean to go viral on chatbot?
41:28
It doesn't It's kind of a nonsensical question today.
41:30
Maybe there's something there where you know, we launched MCP apps, you know, as an open standard.
41:36
Maybe That's a new app store kind of distribution potentially.
41:39
>> And you can bring that in and like maybe that will be the first like is the first MCP app to go viral is something consu- mery.
41:45
That would be interesting, right?
41:46
That's a novel distribution mechanism.
41:49
Um I was talking to um somebody Oh, it was somebody who works at like a nonprofit explorer and um we uh featured them in like a we did an article about how like MCP was being used for a new thing and he's like, "Oh, it's really interesting seeing like what constitutes viral growth for a nonprofit explorer thing."
42:06
Which is like huge numbers, but he was like, "I like I checked the connector list and we were in the top five next to like Gmail and Slack for a couple days.
42:11
And it like got a ton of attention. So, >> Huh. yeah.
42:14
So, maybe, you know, I'm thinking out loud here and this is very much like exploring thoughts with you, which is fun, but maybe, you know, we might not land on like the perfect answer.
42:21
But, maybe that is a thing that we'll we'll start seeing is can you provide enough utility within the sort of frame of you know, a Claude or a ChatGPT or a Gemini and then figure out how that spreads in some interesting way. I'd love to see that.
42:37
I think that'd be interesting.
42:38
>> Yeah, you did a pod with Dan Shipper from Every recently and I was listening to it and you were talking about using Claude to rebuild Bourbon, which for people who don't know was the app that predated Instagram that you guys turned into Instagram.
42:49
And it did it in like a few minutes.
42:52
And I was wondering A, like how did you feel emotionally that that experience seeing an AI just like build that thing that you and Kevin toiled on for I don't know how long like in a few minutes.
43:03
And also, if 2010 Mike had access to Claude today, would Instagram still happen?
43:11
Yeah, it's a really good question.
43:12
>> Because I would imagine the AI would have maybe not gotten there by itself.
43:17
Yeah, I think that was my my reflection having built it.
43:19
It was a one I knew exactly what I told it to build.
43:22
And actually, and I told Dan this was really funny cuz it over built it actually built filters into Bourbon, which Bourbon did not have.
43:27
We launched that with Instagram because Bourbon was totally web-based and the time there wasn't anything like WebGL to do filtering, right?
43:34
So, pretty interesting to see what it came up with even in that.
43:38
I think so much of product going back to our conversation on like the founder teams inside Anthropic, so much of the work is still asking the hard questions, putting in front of people.
43:49
I love both love and hate the feeling of like getting smacked in the face the first time you put out a product in somebody somebody and they're like, "What?
43:56
I have no idea what this is."
43:58
Or like, "I am so confused." Right?
44:01
My first job out of school was a UX researcher.
44:02
Week in and week out bring people into to lab. I still love that.
44:07
Bring people in and the fact that it was built using Claude or handwritten doesn't matter to the end user.
44:11
They just care if the product is full and usable and you know, just create some kind of moment of delight. That's still hard.
44:17
Like Claude is not going to solve for that.
44:18
So, I mean, my If we had Claude back in 2010, for sure there were like ways in which we were slower at at delivering on things that I would have wanted us to be, especially when the thing to build became really clear, right?
44:32
We launched, we knew we wanted to build app mentions and that was probably a week of wiring it up from the user interface, fighting with the text layout engine at the time, like persisting on the server.
44:42
Could have done that faster and delivered value to users quicker, but I don't think that the journey from Instagram from Bourbon to Instagram is something that Claude would have changed much of other than in some of the like coding sports that we that we did.
44:53
But zero to one coding, at least at the time I didn't feel like often like that was the limiting factor.
44:59
It was the thinking and the exploration process as well.
45:01
And so, I think the hard things are still hard, I guess is maybe how I would put it and um a fear or a concern would be if the LLM provides too much out of the box in terms of decisions that are being made, does it actually Yeah.
45:15
like prevent you from finding the orthogonal shape?
45:17
Now, you can use it to your advantage and say, "Great, I want it to generate three alternatives and then see like which of these feels good."
45:25
And I've done that before, too.
45:27
>> Um but yeah, I think it does definitely does not absolve you of making hard product decisions and in fact, it's important that you do so because >> more important. Yeah.
45:34
I feel that in my work, which is not even just building software, but just building media, is I could give myself over to Claude more and I could just let the skill file hopefully like remind it of what I like and all that stuff.
45:47
But then I do that and then I really read it and I'm like, like and I'm talking to it like I would talk to a reporter when I was an editor in the newsroom. Yeah.
45:55
And I feel like that's more important now as it gets more capable because I don't want to lose like my intuition in the process. Some of Yeah.
46:03
Yeah, I feel like a lot of people maybe are with how coding has taken off or starting to maybe give in a little too much to it.
46:10
And I hope that we kind of shift back towards that.
46:12
>> I think the the human instinct is I like the reporter editor piece.
46:16
Unless the reporter is like really gone off the rails, I imagine most of the >> and I did in the day, yeah.
46:23
Most of the time the editor's like, "Okay, this is like 90% close to what I want and maybe with some edits I'll get to 95." >> Yeah.
46:30
Or maybe it's like 80 to 85, right?
46:31
Versus if you're seeking it from the beginning, you might try to get to closer to 100%, right?
46:36
Not that these things are perfectly measurable.
46:40
I have that experience building even within labs where I've learned now that um it's much better if I sort of hash things out with Claude and much more of a conversation before it writes a line of code and then say, "All right, this feels good.
46:52
Like we've collaborated effectively on a spec and now go and implement the spec."
46:55
Versus if I just give it the high-level feature, it will build the feature.
46:59
It will like it will work.
47:01
All the verification stuff we're building will mean that it functionally works.
47:03
And then I'll look and I'll be like, "I would have done that a little bit differently, but now it's done and now you know."
47:08
And I don't like that feeling of it being like >> It's a little hollow. >> not not quite. Yeah.
47:11
So, like moving to a place where somebody put it really well internally, which is like the whole our whole job when coding with Claude is like expressing the North Star with clarity and then helping, you know, make sure Claude converges on that North Star efficiently and effectively.
47:26
And like that is the role of a sort of guide in some ways manager or the managers at least in software tend to be more on the like human development side.
47:35
So, it's really architect.
47:38
There's not a great word I've found for it, but it's that sort of Sherpa of, you know, project and Claude towards that end versus expecting that one it's just going to do it out of the box or two that you don't have a hand early in actually shaping that North Star.
47:52
>> Yeah, I mean there's so many interesting parallels with Instagram and a thing I remember you guys had suffered with a lot early in Facebook helped with was was computing like serving the app, right?
48:01
And your servers you didn't have the servers you needed and there was a moment where like you had to do something. Yeah.
48:06
And and that deal came together very fast and now it's funny like over a decade later like compute is still the problem.
48:12
And I imagine you know I know you're not working on compute directly but I imagine compute is a thing you have to think about a lot internally in a way that you probably didn't envision when you decided to join the little AI lab two years ago.
48:22
There's a maybe two sort of vignettes that really illustrated it for me.
48:27
One was in showing like you know I've learned a lot in two years but yeah when 3.
48:31
lot in two years but yeah when 3.5 launched you know day of even we're so much smaller than we are now but day of like people were adopting it we were starting to get like near like the you know like fully maxing out our like whatever chips we had allocated to at the time and I remember asking the infra team like great so what do we do when we're out do we just like add capacity cuz in my world like from the Instagram days like unless you were
48:57
doing some very specialized hardware on AWS there was more hardware out there to do and they're like no no no when we run out of this that's like that's it like that those are all the GPUs fully allocated and like yes we're working on getting more and I was like oh this is a
49:09
very different environment because you know they you know you can't just you know it's naive right you can't just like hit new instance and and and do that and so that's Hence you guys apparently going to space with Elon. Exactly right yeah finding new sources
49:22
Exactly right yeah finding new sources of unexpected sources of compute but that was a very quick lesson learned I learned that internalized that immediately but yeah we think a lot about you know even within a given product you
49:34
don't want to sacrifice intelligence so how do you deliver as much of that as possible without being sort of wasteful on computer or being thoughtful about how what you do asynchronously or being thoughtful about when you need the
49:44
biggest model versus small so it is a it is a consideration and I think it's a healthy one for a bunch of reasons but I think it also brings us closer to the I mean we were talking about customers before, you know, they're operating in a world where
49:57
they're buying tokens effectively and then reselling it in some, you know, sort of product um and I think we're very aligned with them in a lot of ways in in wanting to make sure that that's a healthy ecosystem for people. So you've
50:07
So you've seen a bigger shift towards token based sort of re-usage based pricing.
50:13
That has been a shift over the last year.
50:16
I remember looking at um you know, the ecosystem a year ago and saying, "Look, our customers want to charge more or they want they want to deliver more value and they're kind of constrained by their current pricing models and now so many more of them are able to actually deliver on intelligence or at least make it customizable, right?
50:33
Where you can say, "Great, I'm trying to optimize for cost.
50:36
I'm going to use uh like less frontier model, but know that I'm going to going to get this value."
50:39
Or "I really want to get this done.
50:41
I'm going to use fast mode.
50:42
I'm going to burn tokens, but I it's because I have like that that knob that I can that I can tune."
50:47
But yeah, compute definitely way more of a consideration uh here than than IG, but in a similar way like Instagram had but felt like and probably was if you do the math like exponential growth at least >> said you guys had like a 80x uh Q1 in terms of cloud code.
51:01
I don't know if Instagram had 80x in >> No, I think we definitely I mean other than maybe week one, right? Yeah.
51:08
From 100 to 100,000 we uh but yeah, it's you know, we had steady very good, you know, percentage growth, but nothing nothing like this. We can land it here.
51:16
I mean you you were you've been very early at Anthropic.
51:20
>> I mean I say early, it's been 2 years, but that is early in AI world and I think the culture of Anthropic is very fascinating and it's something that I've been learning more as I spend more time with the people in the company and I think people look at the headlines, they look at mythos, they look at Dar Hal Daria talks about, you know, maybe 50% of jobs will go away and then a lot of people are like, "Oh, they're just fear mongering."
51:39
Or there's like there's tactics here that are about regulatory capture, all these things.
51:42
The culture is very unique I find and I'm curious to hear from you, is there if you could identify a disconnect between how a lot of people externally perceive what you guys do and how people internally believe about what you're doing, is there something there that you could put your finger on?
52:00
It's a hard thing to convince people of because I can say all day like we are like a profoundly transparent company.
52:09
Like when we say things we are saying what we think, right?
52:11
And Dario uh probably more so than any person I've ever met like speaks his mind not in a calculated, oh this will like help me fund raise and like like that I've just like having gotten to know him and seen him operate both internally and externally like it is not the sort of motivating force behind like the communications or the rollout strategy.
52:32
Of course it's one thing to say that and it's another thing to live live that, but I mean I think if it's a helpful like illustration even internally you know, this year has been great from commercial growth, like great from usage growth, etc.
52:44
But I think the reason the company has stayed grounded is that the point of Anthropic is not to build a large commercial enterprise.
52:50
The point of Anthropic is to, you know, do our best to nudge the world towards a good future with AI.
52:58
And once you see everything through that lens, like it explains all the downstream decisions.
53:05
Now, I can't ask people not to be cynical outside.
53:08
So ultimately we'll have to continue to to hopefully demonstrate through our actions that like we're not trying to fear monger or we're not trying to like hold things back because we think it'll be, you know, perceived as more frontier.
53:17
Um I guess maybe I'll illustrate one thing which is like the internal belief is like we should be able to release a Mythos class model safely and the fact that we haven't yet is like bad.
53:31
Like it's not a we're not proud of it, right?
53:33
It's that if we do it right then all of the positive use cases can be had as well, you know.
53:38
We talked about life sciences earlier.
53:40
There's definitely some interesting things that could happen in in Methuselah like um I built a lot of software using Methuselah internally. It's really good at it.
53:46
Like there's there's all these >> There's no cybersecurity risk for your internal software. >> Right, exactly.
53:51
We have a way Well, actually we've like figured out how to like, you know, do do it as safely as possible.
53:55
But like that is how we see the world, right?
53:58
Like we want to usher We want to give people these capabilities.
54:00
We don't want to like hold them back.
54:02
And we're working really hard so that that's not the case.
54:06
So, you know, it'll be a prove it.
54:08
I think we'll That that that that's fair enough.
54:11
But like it does, you know, when I see people like, "Oh, they're they're they're just trying to like inflate their next valuation."
54:16
Or they're like trying to make themselves sound cool.
54:19
Or they don't really believe this.
54:22
Like that is to the extent that you believe me, that is not like the motivating force internally.
54:27
But again, I think it's a Our journey is to show that over time.
54:31
Okay, lastly, when people come to you, especially because you work at Anthropic, and they say, "Mike, I'm worried about my job.
54:36
I'm worried about my family member, my kid, my grandparent, their economic security." What do you tell them?
54:45
I don't tell them not to be worried cuz I think that there are they're going to be big shifts, you know, we don't know exactly how fast.
54:51
You can disagree on the timeline, but it is already shifting and happening.
54:55
And I get actually a whole class of email that I've gotten recently is parents of uh college graduate like people I know either socially or professionally who have kids who are graduating. And what do we do?
55:07
And um I think the thing that I tell people, which is what I believe, is that it is it is not one company or one sort of branch of government is going to solve this.
55:17
It's going to take the kind of sort of societal level of conversation.
55:21
If there's anything that we've been trying to do, it has been to try to spark that conversation.
55:25
They're going to get read as fear-mongering sometimes.
55:29
Um whether that's, you know, everything ranging from like figuring out like different tax structures, rescaling.
55:37
Like it's actually going to take all of these things.
55:38
And And one of the things that you've seen from us is being more specific on what some of those policy proposals might look like.
55:45
Um this is a long answer that I would tell somebody.
55:47
But it's a it's a complicated question.
55:48
But yeah, I think what I tell folks is you're not alone.
55:51
This is a shared kind of complicated thing.
55:53
I think the things that will still remain human and ineffable and important are still relationships and curiosity and creativity and the ability to organize people towards an end.
56:03
Like that's still going to be really, really important.
56:06
I don't see AI like replacing that anytime soon.
56:08
Um to the extent that that's a skill that you can pick up or at least develop or at least continue to nurture, I think that's really, really important.
56:14
And then also not get as uh like things will continue to shift.
56:19
So even if in the current moment with all the uncertainty, maybe you know, our friends kid doesn't land in exactly the job that they wanted.
56:26
Things will continue to shift.
56:28
Like nothing is set in stone.
56:28
And I think if people remain sort of uh curious and actively exploring what the frontier looks like, they might then either be part of creating a whole new category of jobs or progressing to a different place in their own companies.
56:42
Like I think the the landscape is still going to continue to shift.
56:45
So not treating the current moment as uncertain as and in some cases challenging as it is as like a static thing that will lock people in. All right.
56:54
Uh Mike, I'll let you get back to the lab. Thanks for doing this. to see you. >> All right, thanks. >> Thanks. Don't go anywhere.
56:58
After the break, Ellis is joining me to recap my conversation with Mike and talk about how old we are. And we're back. Surprise.
57:09
It's me, the Hamburger Helper, here with Alex Heath to discuss an event I was not at for a variety of extemporaneous, extenuating, spontaneous circumstances that may or may not involve founder clients and two-year-olds.
57:27
Uh but I'm very jealous that you got to go.
57:29
How was uh how was your trip? It was good.
57:31
I uh, yeah, I feel bad for you that you can't do the the Burbank to SFO day trip every week like I do.
57:39
You know, it's something that I wish on everyone. Uh, just kidding. But no, this was great. Uh, it was great to go.
57:43
This was Anthropic's second developer conference that they've ever done.
57:50
And the scale of it kind of maps to a how Anthropic has grown.
57:52
So, I didn't go last year, but it was only a couple hundred people.
57:58
And Anthropic has now obviously become like the fastest growing company of all time.
58:05
You know, valued around a trillion dollars.
58:07
And this was like 2,000 people with three satellite events in like different cities.
58:12
And I think they'll probably go even bigger next year.
58:14
So, it's actually kind of cool to just see this culture of this company.
58:19
The events are this really cool physical manifestation of how these companies are changing and evolving, especially in tech.
58:26
I mean, gosh, I, uh, I just finished with therapy 30 minutes ago.
58:30
So, now I can borrow some language and say, "Can we just pause and sit with that for a second?
58:35
How far Anthropic has come in just a year?" Oh, this is good.
58:39
So, you should always do therapy before we record. Yeah.
58:41
Cuz it's an absolute 180, isn't it?
58:43
I mean, gosh, I feel like OpenAI was just absolutely unassailable in terms of momentum, product.
58:52
And now all these companies are kind of seemingly trying to focus, bring back maybe some of the apps they'd built over the last year that were distracting them from gains as the market enters turmoil.
59:08
Uh, so much has changed, I feel like with the vibe.
59:10
There's been a vibe shift, one might say. There has.
59:11
And Anthropic is causing the turmoil, so I think they feel pretty good.
59:15
But it was also kind of interesting.
59:17
I saw a lot of chatter on social media about people like the day before seeing the, uh, Claude with code, uh, conference signage going up in SF going like, "Wait, this is happening.
59:27
Like people also didn't know that this was happening and it was also reflected in the reporters who were there.
59:33
There were the you know, typical you know, the tech reporters that I see all the time at these events, but I was amazed at all of the publications that did not have people there and it was clearly just like not on their radar that you know, maybe the hottest company in the world.
59:47
Certainly the most influential tech company right now was doing this big event with their co-founders and you know, Dario and Daniela were there and did a chat and all of that and the the headlines poked through because of the Elon thing where they announced that they were working with Elon on compute, but other than that I I don't know if this would have really poked through into the mainstream.
1:00:07
Um like I'm sure Anthropic hoped it would have.
1:00:10
Most importantly, where do all the reporters sit now?
1:00:12
I've always been a goody two-shoes sitting at the front of the class.
1:00:16
So back when I was reporting, I used to sit at the front uh right next to our old friend Josh Constine from TechCrunch.
1:00:24
Where do you sit these days?
1:00:24
They put you in a press room, you know, to kind of toil away away from the normal attendees, right?
1:00:31
Cuz part of that is they also don't want you interacting with the uh non-press attendees as much as possible.
1:00:36
So you you are you have to work harder for the scoops.
1:00:40
The creators don't have any hard questions.
1:00:42
Don't poison their minds.
1:00:46
But no, it was very nice.
1:00:46
They the Anthropic folks did a great job.
1:00:48
The food was actually quite good, which it usually is not at these events.
1:00:51
Uh so I give them props there.
1:00:53
Uh and they did have reserved seats for us in the keynote, which I always appreciate.
1:00:57
It's nice to still get something as a member of the media these days.
1:01:01
One of the other reasons I missed it, I was thinking back to my interview, which we can bring up on screen for those watching on video with Mike Krieger back in 2013, I think. Oh, wow. You beat me.
1:01:16
>> of a crazy amount of time ago. 13 years ago. >> 2013? Yes.
1:01:21
They'd just been acquired. >> gotten bought, yeah.
1:01:26
Yeah, just acquired by Fartbook.
1:01:26
And he has a quote about the last year being crazy because Instagram grew from 30 million users to 100 million users. Just incredible.
1:01:38
And he's talking about how as Instagram grows, he goes, "We got to think about the world here.
1:01:46
We got to optimize for a lot of phones.
1:01:48
Not everybody has an iPhone 5." And it's just like, wow. So much has changed.
1:01:55
>> Man, of course you beat me because he and I started before we were recording reminiscing.
1:01:59
And the last time I did something on camera with him for like a podcast type thing was in 2018.
1:02:04
And it was for the New York office opening of Instagram.
1:02:09
And it was the year before Kevin Systrom and Mike, the two co-founders, very dramatically, I'm sure you remember that, dramatically left Facebook, right?
1:02:17
Over like the power clash with Zuck.
1:02:18
If you were to told both of us where we'd be in 2018, I mean, Mike, you're a little bit more accomplished, but you know, it was uh definitely definitely wild to think back to that.
1:02:28
>> Well, so this is funny.
1:02:28
I'm actually looking at the URL for the article right now.
1:02:32
It was almost exactly uh 13 years ago. It was It was May 13th.
1:02:36
And when this episode airs, it's going to be May 14th.
1:02:42
Yeah, so that would have been That would have been a funny coincy-dink.
1:02:44
That makes you feel old, doesn't it? Yeah.
1:02:49
And that's the segway into my next question.
1:02:52
Uh well, wasn't really a question, but now I have a question.
1:02:55
So, I know that Mike, since he's been back at Anthropic and just recently has kind of, you know, let go of the reins for the chief product officer, now is kind of back to an IC role.
1:03:04
I feel like that's always kind of the dream for a lot of engineers, but I was also reading back in this in this interview I did with him that he was talking about waking up around 2:00 p. m. and coding until 6:00 a. m.
1:03:16
Is he still in that mode or is this a different type of IC mode?
1:03:18
Now being the uh world's most famous co-founder emeritus. Right.
1:03:24
I mean, Mike is looking very healthy and well-rested.
1:03:26
So, I would be shocked if he was doing that still, but who knows, man?
1:03:31
A lot can happen these days.
1:03:33
Like with modern advances in health technology, you can do a lot.
1:03:37
It's a very different job though, isn't it?
1:03:39
I mean, I just appreciate like he's always been uh very kind, humble guy.
1:03:44
And I also interacted with him a bit when he was on the board at the browser company when I was there for a couple years and he's just kind of kind of a real one, I think.
1:03:52
And it's fun to see people in that position who, you know, frankly have all the money in the world just kind of want to get back to work and help work on some skunk work stuff and actually be able to bring the urgency uh that in some ways only founders can if they can retain that energy, you know?
1:04:10
And I was talking with the Anthropic comms team too about this.
1:04:15
He's also like the dream spokesperson for a company, especially a company like Anthropic because he is, you know, legit. He's personable. He's charismatic.
1:04:25
He'll go and talk about things, right, that may be considered passé for, you know, maybe your more uh cerebral, you know, AI co-founders to talk about, but at the same time he he can he can say a lot without saying uh too much.
1:04:41
And that's a that's a rare skill.
1:04:44
Uh and >> he say too much when he made the Instagram for Snapchat comparison?
1:04:49
No, I think that's great.
1:04:49
I think that's like that's podcast gold, right?
1:04:51
I think uh Yeah, of course it is. Yeah, yeah, yeah.
1:04:55
Did the PR people bite their nails? Mm, I don't know. I don't know. I don't know.
1:05:01
I mean, he you saw how the way he reacted to Figma, which was, you know, uh very diplomatic.
1:05:06
I had to ask him about that situation for and, you know, uh I don't know. Mike's good.
1:05:12
He's good at what he does Uh and he's a great interviewee and I really appreciated the time with him.
1:05:19
I think it was nice to get into his head and hear kind of the areas that Anthropic is exploring next.
1:05:25
That was the goal for me because, you know, we talked about this like if the team he's on, the last big thing they did was literally Claude code, literally the thing that has made Anthropic the most valuable company in the world.
1:05:38
That's a tremendous amount of pressure and even for the co-founder of Instagram, that's a lot of pressure and so I wanted to hear the areas he was exploring and I do feel like he gave us a lot of good detail on that, which I appreciated.
1:05:49
Yeah, now that it's been a little bit of time between the interview, what do you feel like you've been thinking about the most and connecting the dots in your sources personal knowledge management system? What are the big themes? >> Right.
1:06:02
I need Mike to build me that.
1:06:02
A, um uh I I think I think that what I've been thinking about is it's clear that health is a thing they're going to go into next.
1:06:11
So, if you were an investor type you would short all the health stocks, right, given how things have gone with Claude design and Figma or what have you.
1:06:23
It's very clear they're focused on health.
1:06:25
There were two things he said that really stuck out to me about Anthropic's culture, which I think is very unique and I think people discount how unique it is and how much it drives the decisions Anthropic makes.
1:06:37
Um but he said at one point that the the point of the company is not to build a large, you know, commercial enterprise, which is ironic given that it's the fastest growing commercial entity of all time, but to quote nudge the world towards a good future with AI.
1:06:50
I think there's reasons to shine a light on that and a magnifying glass and go, "Well, like, okay, then how does that make sense partnering with Elon or the Department of War stuff or whatever?"
1:07:03
But I do I a lot of people, the vast majority of people inside Anthropic do feel that it I was talking about this um actually with uh another senior leader there um just off the record that day that the culture of Anthropic really reminds me of Apple in a lot of ways, the good and the bad.
1:07:18
a lot of ways, the good and the bad. It has the good of really focusing on the craft, um the commitment to detail, the world-building, the kind of reality distortion field of of what's going on, um the focus on product, but also you get a little bit of that chip on your
1:07:36
shoulder, a little bit of that reality distortion field, which can also, you know, potentially be negative, and it has a bit of a insular culture like Apple, where it's very much like you're in I don't want to say cult, but it is kind of like a cult, like you're in it or you're not, and you get it or you don't. It's not really
1:07:53
It's not really uh like a normal big tech company can just uh employee can just show up to Anthropic and necessarily succeed.
1:07:58
So, I thought that was interesting.
1:08:01
What do you think that's due to?
1:08:02
I mean, I'm a believer in kind of two schools of thought.
1:08:06
One is heavily documenting very specific beliefs and philosophy, and I think they've done that better than anybody else in AI probably.
1:08:14
I mean, I feel like I have like three to four browser tabs open that I just can't seem to make time for that are all the different essays that they've written.
1:08:22
Think the most recent one was kind of like teaching in Claude why.
1:08:23
It's been a lot of great stuff, you know, they had the constitution, uh obviously Dario's essays, but then on the other hand, it all just still comes back to people hire people like them, and it just comes back to the top uh and the founders, you know, hiring people all the way down.
1:08:41
I think that's what it is.
1:08:43
I think it's a very much a a safety-aligned culture, and the culture interview part of the hiring process is very serious.
1:08:51
It's way more intense than most companies. >> questions to Yeah.
1:08:54
They ask things about like what are the hardest like like moral tradeoffs you've had to make, like when have you, you know, put your morals ahead of like financial gain?
1:09:03
Would you be willing to do that at Anthropic? That kind of thing.
1:09:07
Did you ever go to detention in middle school? And if so, why? Right. Right.
1:09:12
Um but yeah, man, the culture's the culture's super unique.
1:09:16
Um and it's probably a rare thing for someone like Mike to come in and succeed in a culture like that because he was, you know, in the Facebook machine for so long.
1:09:27
And uh he joined, you know, two almost two years ago, two ish years ago, which is an eternity in AI.
1:09:33
Especially for Anthropic.
1:09:36
Anthropic was like not on the map really at all two years ago.
1:09:38
So, he obviously saw something and made a bet on something that he didn't need to make a bet on cuz he's, you know, he's the freaking co-founder of Instagram.
1:09:46
Like he doesn't need to do anything.
1:09:48
But he clearly saw that there was something happening here and uh yeah, the not that he needed another break, but he certainly got one.
1:09:57
I mean, it's a testament to him, too, just that he stayed in touch with everything happening uh in the world and in culture, which isn't always or maybe even usually the case uh for people who've kind of been around and had such big success.
1:10:12
I mean, I appreciated, you know, his recognition of kind of like the distribution challenges of building product today, like back in the old days, Facebook had 15 minutes of time per day that you had to compete with, you know?
1:10:25
And now it's an hour from TikTok and it's the most addictive content format created.
1:10:30
And he's clearly still plugged in, which I really genuinely appreciate.
1:10:35
You know, you meet you meet founders or people in this industry all the time who uh may even celebrate not being in touch anymore.
1:10:45
Uh just kind of moving beyond it, you know, whether it's with politics or otherwise.
1:10:49
or otherwise. And it seems like he's he's retained that uh even going to the point of rebuilding uh Bourbon from scratch, which was really fun, and then coming away from it with an actually cool insight, which is that even though
1:11:01
it helped us build it so quickly, I think he said something along the lines of like, if you just use LLMs for everything, it might actually prevent you from finding the surprising or non-obvious or culturally edgy things that might make a product work. Um
1:11:14
Um certainly speaking of Snapchat, I mean, what AI in the world would have ever suggested that ephemerality was going to be the solution to the pressures of social media?
1:11:25
Like, that was just a crazy one, and every single person said Evan was wrong about that. Yeah.
1:11:31
Until it started working.
1:11:33
No, I mean, his point about how human intuition and relationships and creativity become more valuable, I think, was really good to hear, and I do agree with that, and I think the people who are really plugged in on AI and where it's going are all saying that, and it was good to hear that from him.
1:11:45
I don't know if you also caught he's he's very focused on distribution still, uh having built something that obviously has tremendous distribution, billions of users with Instagram.
1:11:54
I thought his thing about, I don't know if you caught this, like MCP is a distribution mechanism was interesting, like an MCP leaderboard and hearing from partners uh on MCP that, you know, they get tremendous, you know, inflow of traffic via MCP if Anthropic features them.
1:12:09
So, it's almost like this new app store model being created where instead of downloading apps, you're in an MCP store or something.
1:12:18
I could see them productizing that as well.
1:12:20
It sounded like he was really focused on that, and that's his consumer mindset, right, being brought to bear on AI.
1:12:26
Yeah, I'm trying to picture what that might look like, and I mean, if MCP turns everything into a pipe, and there's a lot of competition, then the fundamentals underlying those pipes, whether it has to do with the fees that you get charged by platform or even to send a transaction, are just going to get lower and lower, I think, which is which is definitely a good thing for consumers.
1:12:48
Uh obviously, the moats are very thin, and we'll see what happens to uh venture capital and consequently to myself as a result of that.
1:12:58
But uh Yeah, it seems like >> going to be all right, Ellis.
1:13:02
That's That's my takeaway of the last year or so.
1:13:06
>> I just feel like everyone feels that way.
1:13:07
I think I think marketers and CMOs and brand people are going to become more valuable because that's how you're going to stand out, you know, when when anyone can make anything. That's what matters.
1:13:17
I guess that's why uh our friends at TPPN are over at Open-I. I know.
1:13:22
Uh are Have you Have you chatted with them?
1:13:24
How are they feeling being back in the slack of a big company?
1:13:27
I need to check in with John and Jordi about their X views cratering since the the deal and see how they're feeling about that. That's a good reminder.
1:13:36
Um but yeah, I know, look, man, this was a great episode with Mike.
1:13:38
Uh bummed you couldn't do it with me, Ellis, when we were actually charting out early guests that we wanted to have on the show, Mike was someone you put forward that you wanted.
1:13:47
Um so, I tried to do you I tried to do you right and ask a few of your questions you sent me.
1:13:53
I don't know if you caught one of them.
1:13:55
He He took He took a moment and said, "That's a really good question."
1:13:57
when I said it was Ellis's question. So. Which one?
1:14:02
It was the one about his lab being a broader structure for companies in general.
1:14:08
You know, these like these founder-lean, you know, based companies that don't have a ton of, you know, rank and file.
1:14:15
Um and the answer was basically yes, of course, but that's uh that's the future, man.
1:14:20
No, all I see is I'm picturing in 50 years we have 500,000 clones of Brian Chesky that are all spitting up $1 companies simultaneously.
1:14:33
With a fleet of AI agents. I mean, Brian's cool.
1:14:36
Like, come on the pod, Brian, but of Brian?
1:14:39
He's just He's got that, you know, New York hustler mindset, tone, body.
1:14:44
Uh who's the who's the perfect apex founder these days?
1:14:50
I feel like a lot of people say it's him.
1:14:53
You heard it straight from Ellis, folks.
1:14:55
How could any of us compete?
1:14:58
Well, this episode will be dropping the day that we're having our live show and launch party at Notion's headquarters with the CEO of Substack, Chris Best.
1:15:05
Uh Ellis, do you have your your toast for the room prepared?
1:15:10
Have you written a speech? Oh, Jesus. No.
1:15:15
But I did buy an Away suitcase.
1:15:15
You told me to get an Away suitcase instead of my Radial Travel Pro, so I did upgrade. >> Good.
1:15:22
And I did get some Nothing headphone A ear cans, so I could look cool on the plane as well. Oh, yeah. Shout out Carl.
1:15:29
So yeah, I'm trying to to to pump myself up.
1:15:33
I'm excited to have several former guests there and it's going to be a great group and hope we can do more live stuff down the road.
1:15:39
But yeah, uh Hopefully they show up instead of just talking to their agents. Yeah.
1:15:45
I mean, that's the one thing I think that may last in AIs is is events because what is the joy of sending your agent to an event? Mhm.
1:15:51
You can't taste the drinks that we've creatively named.
1:15:56
You can't have the apps, the past apps that will be delicious.
1:15:58
You can't hear the DJ we'll have.
1:16:02
You can't potentially meet your future partner, you know?
1:16:04
What kind of partner are you talking about?
1:16:05
Business or love partner, either one. That would be sweet.
1:16:09
That'd be quite an accomplishment if we had people like making out on the dance floor at our party. That'd be awesome. We'll see. All right.
1:16:17
Well, should we wrap it there?
1:16:18
And that's it for this week's show.
1:16:20
Thanks to Mike for coming on.
1:16:20
If you like the show, don't forget to like and subscribe everywhere you get podcasts. We are access. show online.
1:16:29
That's where you can find all of our links and you can find us in video, lovely high definition at access pod on YouTube.
1:16:35
If you really like this episode, we'd love a five-star review or even share with a friend or agent.
1:16:39
You can find my newsletter at sources. news.
1:16:43
And you can find me at hamburger on Twitter and at meaning.
1:16:45
company for your startup storytelling needs as long as they remain relevant. I think they will.
1:16:51
Access is part of the Vox Media podcast network and the show is produced by Hooked Creative. Bye.