Transcript: 'How Sam Altman Uses Dots to Take Back His Time'

The transcript of The Every Podcast with Sam Altman is below. Watch on X or YouTube, or listen on Spotify or Apple Podcasts.

Timestamps

Introduction: 00:00:30
Why AI hasn’t replaced developers: 00:00:55
OpenAI’s platform bet on the rise of small-team companies: 00:04:45
Why the AI era is more “Renaissance” than “Industrial Revolution”: 00:05:45
How Sam’s Dot, OpenAI’s always-on AI agent, gave him back his mornings: 00:07:40
Sam’s process for building new features with his Dot: 00:12:00
Inside OpenAI’s “live documents” that agents and humans can build together: 00:16:00
What bring-your-own-subscription changes for app developers: 00:20:55
Why “OpenAI just killed X startup” never comes true: 00:24:00
Sam’s life on Ultrafast and the push to make speed as cheap as intelligence: 00:27:55

Transcript

(00:00:00)

Dan Shipper

You do all your prompting on Ultrafast?

Sam Altman

I do.

Dan Shipper

That’s sick.

Sam Altman

Yeah.

Dan Shipper

That’s crazy.

Sam Altman

Seven people’s dots working together can create something new that kind of existed in all of our heads to some degree, but was an emergent thing.

Dan Shipper

Hey, this is Dan Shipper. You might be wondering what happened to AI & I. This is the same show—we’re just rebranding it to the Every Podcast. I’m really excited for our first episode, with Sam Altman. Here it is.

Sam, welcome to the show.

Sam Altman

Thanks for having me.

Dan Shipper

It’s nice to have you here. We’re here at Dev Day. I think if you had told me four years ago that the coding models were at this level of facility, I would have been like, “Wow, I don’t know if there would be a lot of developers at Dev Day.” But there are more developers than ever, it seems like, and more work to do than ever. What has riding the increasing capabilities of coding models and seeing the impact on engineering taught you about the future of engineering and, more broadly, the future of work with AI?

Sam Altman

I think people just want so much more stuff. People’s expectations are so high, and I think that’s a wonderful thing. I think this is why we’ll all have this incredible future in front of us to explore. Developers have never been able to do more, but there’s never been a higher bar either. And I think that’s great. I think it’s a nice thing about society.

Dan Shipper

Is that something that surprised you? Did you expect that to happen?

Sam Altman

I was never as doomer-y about the job stuff as other people, but it has still surprised me somewhat on the upside—just how much latent demand existed in the world. And also how much people have remained at the center of everything that is happening with AI.

Dan Shipper

Do you think there’s a limit to that? Or do you think that translates pretty cleanly to other areas of work?

Sam Altman

I think there will be some areas of work where there’s actually not much creativity required and we don’t care about the person doing it, and many of those tasks or jobs will be done by AI. But there’s so much other stuff people want in the world, there’s going to be no shortage of new things to go do. I think that’s been one of the lessons of recent times.

Dan Shipper

Why do you think it’s so human-centered, even though we talk a lot about automation?

Sam Altman

One observation is we just don’t care that much about what the machines do. And in some sense, we never have. We’re obsessed with other people. We want to know about the person behind the art or the person behind the company or the idea or the product. And I think this is a wonderful positive surprise that we probably should have all known before.

Dan Shipper

I agree. I’ve been thinking, if an agent solves Navier-Stokes in a data center, but no one finds out, did it really solve it?

Sam Altman

I think that was a great example, where 10,000 agents worked super hard for four days in perfect coordination with each other and did this amazing thing. And almost none of the public reaction was about the problem actually being solved. There was a lot of talk about the interpretation, a lot of talk about the people involved in it in different ways. But that was a good thing.

Dan Shipper

It’s Dev Day. You launched, by my account, 22 new things today. I’ve been at all four Dev Days, and I think last year was nine and the year before that was four or five. So it’s a lot more than usual.

Sam Altman

Just wait for next year.

Dan Shipper

Okay. I’m tired.

Sam Altman

We cut a bunch of things that we were going to launch today just because you have to fit it in 50 minutes. But we try to practice what we preach. If we are saying these tools can make people much more productive and build more, and that people want more stuff, we’d better build more ourselves.

Dan Shipper

What are the new bottlenecks and best practices for doing that? Because I can see for myself—and I assume you guys are seeing this too—now that you can build everything, you’re like, what should we build, and how do we make it all coherent? How do you think about that?

Sam Altman

Moving to this world of persistent intelligence, active intelligence, whatever people want to call it—this idea that you’re going to have these agents that are running in the background 24/7, connected to your systems, doing stuff proactively for you—is clearly a direction that feels important to explore. I think we’re going to see amazing stuff happen there.

Another thing we really wanted to do: There are now so many ideas that are capable of being built, and what we are missing is the creative spirit to figure them out and build them. So we wanted to really open up our ecosystem. Let people build apps in ChatGPT the same way that we get to, let people bring their subscription wherever they’d like, have the marketplace, and really try to let people get connected to all the stuff that we won’t be able to build.

Dan Shipper

Looking at all the platform features that you launched and the way that you talk about platforms, it reminds me a lot of the way that Bill Gates used to talk about Windows as a platform where you create more value than you capture. You said this in your keynote, and I’ve observed it as well: At Dev Day, there’s always an app store. You’ve been taking swings at app stores for a while. What have you learned about building a platform, and what do you think makes this one different?

Sam Altman

We clearly have not yet figured out the app store equivalent, but we have figured out how to build a platform that a lot of people are building incredible value on top of. I think there are really two schools of thought among people building AI model companies. There are people that believe or hope that there’s only going to be one company in the world. And then there are people like us that believe we are going to see a renaissance in the number of companies, and an absolute lowering of the bar for one-person companies and small-team companies to create incredible new products and services. We want to go as hard as we can with the platform to enable people to do that. And although we haven’t figured out the app store yet—that’s totally fair—I think we have figured out how to help millions of people build and create significant new products and companies.

Dan Shipper

You talked at the keynote about an industrial revolution versus a renaissance, which you just mentioned now. If you had to narrow in on the key things that make the difference between those two worlds—maybe OpenAI’s part, but maybe even the broader ecosystem—what do you think those things are?

Sam Altman

One take on this would be that the Renaissance was fundamentally centered around people, and the Industrial Revolution was fundamentally centered around machines. There were a lot of good things about the Industrial Revolution that had a hugely positive impact on the world in many ways, but I think a lot of people did not have a purely positive personal impact from that. People had these rich lives and did all these things. And yeah, of course, they struggled and wanted more and better stuff, and that’s what the Industrial Revolution delivered in some ways. But they were turned, in many cases, into this cog in the machine, doing some extremely specialized thing and kind of disconnected from a lot of the parts that made life good.

I get sad when I hear people talk about how AI is going to automate everything, there’s going to be no role for humans, we’re going to be like house cats, and the AI will decide how to reward us and tell us what to do. I have heard people say that. That’s certainly not what we think. What I am hopeful for is that this is a huge new wave of scientific discovery, art, and company formation—everybody that wants to start a new thing. And that people do get the benefits of all the economic growth and the new stuff happening here, but also that people get back control of their lives and the ability to do the things they want and spend their time the way they want.

Dan Shipper

One of the things that strikes me about the Renaissance versus the Industrial Revolution is that in the Renaissance, you have a different understanding of what it means to be a person—an individual, as opposed to the cog in the machine. Do you feel like you’re starting to see any difference in how we think of ourselves as humans that’s comparable to the way we changed how we thought of ourselves during the Renaissance?

Sam Altman

So much happened during the Renaissance, it’s hard to say there’s this one thing. But definitely a big part of it was about people reconnecting to a lot of things that had existed in the distant past. And then also this combination of that with this crazy futurism and discovery of science and engineering, and this feeling that human potential had massively expanded. You do hear people talk about both of those things simultaneously with AI, which is kind of how I got interested in this as an analogy in the first place.

This is a very trivial example relative to what is going to be possible or already is possible, but Dots, the new thing we launched today, is the first thing that has given me back control of my mornings and my mental space in the mornings. I used to love the early mornings. It’s my favorite time of the day to work. It’s maximum creativity. That’s when I would do the really deep, complicated project of the day. And now I run this big company, and something is going wrong all the time, and overnight a lot of stuff happens. If I didn’t check my phone right away, sometimes it was fine, and sometimes there was a real disaster that, if I ignored it for an hour, just got much worse. The problem had been that as I checked my phone, I polluted my brain with all these other things, and I lost the deep work period.

But now I trust my dot to have good judgment about what I need to think about and what I don’t, to tell me if there’s something very urgent, and to otherwise handle things for me. And I get this space for creativity, or to just hang out with my kids, back. It sounds small, but it’s been a significant improvement to my life.

Dan Shipper

That’s what I was going to ask. It seems like it’s deflecting a lot of things that you might have had to deal with immediately, that you can just leave for later because you’re confident you’ll get the thing you need when you check it.

Sam Altman

Yeah. Two things happened today on that. During the keynote, we had a big sev in production. This is the risk of live demos and trying to deploy many new features at once.

Dan Shipper

I was watching.

Sam Altman

So that happened. Thibault got a message from his dot five minutes before the demo saying, “Just to let you know, I see that you’re in the DevDay keynote. I know you were working on it. I know what the demo is, because I was able to look at the presentation in Space, and the sev is probably going to make the demo not work. Would you like me to try to fix it?” There were a few minutes to go. It couldn’t fix it yet, but maybe the next version will be able to go fix that in prod. And think about all the stuff that had to come together to make that work.

Then, right before I came here—I’ve been in back-to-back stuff all day and I’m very, very behind on all of my normal stuff—my dot, just before I walked into the last thing, said, “You have one little break between now and this thing you’re going to with Dan. You have to pay attention to this one thing.” I never would have found that otherwise, and it would have been a big mess. Just the fact that I don’t have to worry about that feels super different.

(00:10:00)

Dan Shipper

What’s your dot’s name?

Sam Altman

Dotty.

Dan Shipper

Dotty.

Sam Altman

I had Alfred earlier, and then many people are calling it Dotty for now, but I’m trying to think of a great name.

Dan Shipper

Mine’s named Boo.

Sam Altman

That’s a good one. What do you think so far? I know it’s super early and still rough.

Dan Shipper

I like it. It has become the default way that I use ChatGPT, which I think is a very promising thing.

Sam Altman

Me too.

Dan Shipper

I’m in the beta, so I was in the early period, and I think it has limits. It’s been a little bit buggy for me. There are a lot of different surfaces it can be used on, like mobile, and it has its own cloud computer and all that kind of stuff. I think getting those kinds of things cleaned up is a big deal. But yeah.

Sam Altman

I think this is the kind of thing where it’s going to get better rapidly each week from here, but it’s definitely going to take some time to smooth it all out.

Dan Shipper

I had a similar example where I was trying to reschedule a flight, and it was like, “Hey, you have a meeting during this flight that you’re rescheduling.” But it wasn’t on my calendar. It had just read it in Slack.

Sam Altman

Yeah. That ability to bring context together from across your entire life and really do things in ways that would be hard even for an attentive person to get right has been super impressive. That’s been one of the things I’ve found most useful about it.

Dan Shipper

What had to happen from a research perspective to make this possible?

Sam Altman

This is a very complicated system with a lot of stuff, but maybe most important was Astra. You needed a smart enough model. There were a lot of other things you had to build too, but getting the model over the intelligence bar was a key part.

Dan Shipper

You’ve told me there have been a couple of times—not just for scheduling or urgent things like filtering—where it’s helped you see something or make a decision at OpenAI that you couldn’t have hired someone to do. It would just be too complicated. Do you have any specific examples you can share?

Sam Altman

There was one new feature that I was kind of excited about, and it would have been very hard to explain it to people in enough detail to get it built. And even then there were a lot of ways to do it, and I wouldn’t have been confident in which one. So I would just constantly jot notes to my dot as I thought of something, or when I was on a walk or whatever. And then the dot went off and built five or six versions. I was able to give feedback. I was able to iterate. Whenever I had a spare moment or had an idea, I could just dump it in raw form, and the ability to get and give pretty continuous feedback on it was very cool.

There are many categories like that, where there’s an idea you’d be so unsure about, or it would be too much effort for you or for someone else to build, and you can actually just continually iterate on it. It’s pretty cool to go to sleep after leaving a long voice note with a bunch of thoughts as you’re drifting off, and wake up in the morning, and not only is the thing built, but it went off and did the next seven steps.

Dan Shipper

It’s the best.

Sam Altman

It’s like, all right, this is a very cool way to express stuff. And then there’s also a set of tasks that I would have given up on before. There was a thing I was sure I had seen in Slack. Slack search is obviously terrible, so I mostly use Codex to search it. I had been searching for this thing because I wanted to find it again—I wanted to reference it. I spent a good 10 or 20 minutes trying to search with Codex, and I kept trying different things and didn’t find it. And I was like, “Man, maybe I dreamed it, and this didn’t really happen,” because Codex doesn’t usually fail. And I kind of gave up on it.

Then the next day my dot was like, “Hey, I noticed you were trying to do this thing and you didn’t find it, but I kept thinking about it overnight and I found it for you.” And I was like, “What?” It was a screenshot. I had forgotten. In my head, it was text in Slack, not a screenshot. And it just sat there very patiently, looking for something to do to be helpful to me, and looked at every image and every screenshot and found the thing. I would have given up on that.

Dan Shipper

You’re constantly feeling like there’s way too much to read in Slack and email and all that kind of stuff, and it’s perfect for that amount of context—to just get you the right thing at the right time.

Sam Altman

People have talked about the need to have an agent deal with the onslaught of the internet and all of the various things trying to hack our attention and distract us. And I have definitely learned, in the few weeks that I’ve been using Dots, that although technology is very wonderful in many ways, that crazy attention hacking of social media and phones and notifications had gotten way worse than I realized for me personally.

Dan Shipper

Tell me about Space. It’s a big bet to do three office suite-type tools in one launch.

Sam Altman

We didn’t set out wanting to do this. And then we realized—and I’m sure you’ve realized your own version of this too—that software was really not designed, on the whole, for collaborating with AI.

Dan Shipper

Yeah.

Sam Altman

You can use it yourself, or you can have AI use it on your behalf, but to really collaborate closely—especially when it’s multiple people and multiple people’s AIs all working together—a lot of the current abstractions we have break down. So this is definitely an ambitious bet, but I think the space of human-AI collaboration is really important to explore.

Dan Shipper

How are you working with it? Are you reviewing docs in Space, or are you making docs? Do you have expectations for how much of it you’re typing versus how much your dot is doing for you? And with the docs you get, are you reviewing them from that perspective too?

Sam Altman

If there’s a new idea, I like writing the old-fashioned way because it helps me think about it. So I still do that, and then maybe have my dot or other people’s dots do the follow-ups. But one of the things that I have found most amazing about it—and I don’t think we have a word for this yet—is what, for lack of a better word, I’d call live documents: creating a doc that is going to automatically update as people interact or as other things change in the world. And it’s something that many of our dots can work with together, not only to share information. Seven people’s dots working together can actually create something new that kind of existed in all of our heads to some degree, but was an emergent thing. That’s been very fun as we think about new features to explore.

Dan Shipper

Is there a specific live doc that has been most useful for you?

Sam Altman

The dashboard. I think everybody’s probably got a dashboard that they look at every day for how the company’s doing, in the ways they care about. I used to share one with everybody else, or many other people. And now I have one that is a changing summary of the business just for me, based on the evolving things that I’m caring about or not caring about. It looks at all my Slack messages, makes me a new thing, and constantly updates it. And it’s like, oh, this just feels like a living document.

Dan Shipper

That’s sick. What do you care about now?

Sam Altman

A lot on my mind recently is just how we’re thinking about the safety, alignment, and security challenge in front of us. But six months ago, I probably cared more about revenue and growth and metrics—the traditional part of the business. So watching that evolve... I’d be very curious to see what my dot thinks I care about most each day, over time. And we don’t currently have a good way to visualize that.

Dan Shipper

That is really—

Sam Altman

I have not been able to get my dot to figure out a cool way to do that. And I would love to see that over the course of years.

Dan Shipper

I think that’s really cool. I’ve been wanting one that’s a live dashboard of my mood. Because when I get a Slack message and I’m like, “I don’t know if I want to look at this Slack message,” I have my dot go look at it. And then I’m like, “Should I look at this now or later?” And it’s like, “You should look later.” So I kind of want a live thing of my mood, so I know—

Sam Altman

Totally. Yeah, I want that too. I’m going to use that.

Dan Shipper

Very cool. Tell me about the Decisions API. I feel like it’s one of the sneaky most important things today, because Jev is this new hot thing that has honestly changed how we think about building things internally. Did you start the Decisions API because of Jev? Or has it been in the works for a while?

Sam Altman

Diogo used to work at OpenAI, and Diogo is awesome. He was talking about this class of ideas for a long time, and we had been working on trying to figure out how to make it work. This is an example of how people always want more—cheaper, faster, better. The number one request with Luna, even though people love that model, was “I want it to be even cheaper and even faster.” So we have been thinking about how we could push it further in that direction. Seeing the response to Jev definitely made us more excited to prioritize things in that direction, and Diogo has inspired me for a long time in thinking about this. But I think you will see us try to push in different ways in the spirit of making the technology super fast and super low-cost for common use cases.

(00:20:00)

Dan Shipper

If something like Jev comes out—a new type of model—and then OpenAI can add something like a Decisions API, how do you think that changes the race between neolabs and labs, and what OpenAI can cover versus what a new lab can focus on?

Sam Altman

As I understand the approach of Jev, it’s pretty different. And I don’t think we can go—well, maybe we could go do something from scratch like that, but I think we have other things to go do. We want to make the best, cheapest, fastest, most capable model at every point on the Pareto curve of cost and speed and modality. But I think there will be new approaches, new architectures, and new technologies that we just can’t cover all of. For the mainline workhorse, though, we will try to deliver that.

Dan Shipper

I’ve been in your ear a lot about Codex native apps for a little while now.

Sam Altman

We finally did it.

Dan Shipper

You finally did it. I’m psyched. So: plugin extensions. And also being able to bring your tokens from your ChatGPT subscription to an app developer. It changes the equation for someone like me, where normally you’re thinking about how much AI you can let your users use so that it doesn’t bankrupt you. And then you’re like, should I use lower-priced models, even though the experience is worse? But if someone can bring their subscription, I’m like, great, you can use whatever model you want under your subscription. What are you most excited about for the kinds of applications and products that can be built at the application layer in this new world?

Sam Altman

I’m most excited for the ones that we haven’t thought of. There are many reasons that we wanted to do this, but one of the biggest—and why we do this big Dev Day thing—is this very deep belief that there are way more good ideas outside of our building than inside. I don’t know what I’m going to love the most as a ChatGPT-native app or a plugin extension, but I am very confident that there will be things that I really do love that I cannot imagine today, and that no one at OpenAI could really imagine with conviction today. That’s one of the reasons why this is so important to do. We want to find a way to get that creative spirit into people’s hands, and we want to find a way to help builders connect with people who are going to find the technology useful.

In terms of bringing your ChatGPT subscription with you everywhere, we just want that ourselves. There’s all this great stuff that people are building, there are all these new apps, and we want to try them. I hate having to sign up for 100 subscriptions. And also, when I try something, I’d like to be able to use it with full-juice AI. So we want to make it super easy for people to take it with them. It’s your subscription—you should get to use it wherever you want.

Dan Shipper

Do you think anyone at the app layer is doing this well today?

Sam Altman

Building interesting apps? Well, not with the new stuff, just because it’s brand new. But yes. For all of the fear people had a couple of years ago about, “Man, AI model companies are just going to eat everything,” there is an unbelievable number of significant new companies and great new products building this stuff all of the time. I’ve never seen startups get as successful as quickly.

Dan Shipper

You were part of Y Combinator before OpenAI. Does that change how you think about it when people ask you how to start a startup, or how to grow, and all that kind of stuff?

Sam Altman

I would love to be a startup investor again, because, oh man, what a fun time. Watching what people are doing right now, getting to spend time with some of these founders—which I don’t get to do nearly as much as I’d like—it is a golden age.

Dan Shipper

What are you looking for? Obviously besides the founder themselves, what are you looking for in the shape of things that would make you say, “Oh man, I want to invest in that”?

Sam Altman

I was always a very founder-driven investor. I have my pet projects like anybody else, but my whole MO as an investor was to find the absolute best founders I can that are working on ambitious projects and be willing to take a lot of risk with them.

Dan Shipper

I feel like there’s always that narrative that the labs are going to eat everything, and at Dev Day, the tweet that always works is “OpenAI just killed X startup” or whatever.

Sam Altman

The tweet works because people like doomerism, but it never actually happens.

Dan Shipper

Why is that?

Sam Altman

Why does it never happen? Well, A, I think we have a lot of conviction that empowering builders—and not just trying to be the one giant evil megacorp—is important. I think we took a lot of steps today that no other AI lab would take, in saying, “We’re going to give you all of the same tools we use.” We’ve always pushed the frontier on model access. So I think we’re uniquely committed in the industry to doing this. And we really do love developers. I think that’s partly because I just love startups, and that’s where I came from.

But I also think, speaking of a new renaissance, there is so much to explore. People are going to do such incredible stuff. We are going to have such a tiny fraction of that vision ourselves. The world has a fundamentally new enabler. I don’t even know what—this is like fire or something. This is a real change to what’s possible. Maybe we brought it down from the mountain, but we want to give it to everybody, and we want people to do stuff with it.

Dan Shipper

There’s a lot of basic startup lore—like you should always have a co-founder, or even the structure of how Y Combinator does investments—that is based on a previous cost structure, for example.

Sam Altman

Yeah.

Dan Shipper

If you were going to start those things from scratch, or think about them from scratch, what do you think changes?

Sam Altman

There are a few things that are probably just invariants. You’ve got to make something people want. That advice was good.

Dan Shipper

Or agents want.

Sam Altman

Well, yes. But at least by our understanding of the world today, at some level, people have got to want it. So there’s some stuff that just hasn’t changed. And then there are a lot of other things—like how much work you can get done with one person or two people or whatever—that feel totally different. I don’t think there has been enough reexamination of that in the startup conveyor belt of Silicon Valley yet.

Dan Shipper

As a person who’s running a gigantic company and generally probably wouldn’t have time for building stuff and announcing it like you are, how does being able to make stuff on the side change how you’re able to lead, or what you’re able to do?

Sam Altman

We are not a gigantic company, first of all. Maybe the big tech companies have 100 times more people than we do, something like that. But I really believe that if you lower friction a little bit, people build way more. And I have really felt that now with Dots, with these little chunks of time I have, or even a long chunk of time where I can have this interactive thing get a lot done. I feel like I will do something that is not just a little side project for myself, like I used to do with Codex. I hope I launch them into the world. I hope I actually make a cool new thing and get it over the threshold.

Dan Shipper

Does that go out on samaltman.com, or are you talking about a real OpenAI product?

Sam Altman

I meant I would try to build a plugin extension or something and put it out.

Dan Shipper

That’s—

Sam Altman

Maybe I’ll ship a real OpenAI product. Feels like all bets are off now.

Dan Shipper

And so you have to make that decision—do I—

Sam Altman

You know what, I will do that. I’m inspired.

Dan Shipper

Do you have—

Sam Altman

No, I don’t know. I didn’t even think about it, but now I want to do something.

Dan Shipper

I love it. But you do have to make that decision—you could just put a team on that. So what does it add that, instead of having to put a team on it, you can now be like, “Oh, I’m going to make a plugin”?

Sam Altman

Well, I think today was an example of this. You talked about how the amount of stuff at each Dev Day has gone up. We could not have done all of this stuff without having AI to help us. And if this works, next year we will put the teams on it and I will do something, and more people will do something, and we’ll have a crazy amount of amazing stuff. Or just one really great thing—that’d be fine too. But hopefully the total amount of utility we put out into the world goes up.

Dan Shipper

Do you do all your prompting on Ultrafast?

Sam Altman

I do.

Dan Shipper

That’s—yeah. That’s crazy.

Sam Altman

No, I’ve been thinking about this. I worry that in some ways we get out of touch with people building on our tools, and we try to put all of these constraints on ourselves. I’m very aware that not everybody is going to be able to use Ultrafast all the time for everything, but I have such a hard time dragging myself back over.

Dan Shipper

Given that you’re living in that future—and I think all of us can experience a little taste of it—what is different about the Ultrafast world in terms of how you can use these tools?

Sam Altman

There was an amazing talk, I don’t know, 15 or 20 years ago, from Bret Victor about the principles behind how he did his work. I’m going to get this a little bit wrong, but one of them was something like, creators should have an immediate connection to what they’re creating—how speed and the feedback loop were so important. And I did not realize what a difference it makes to my iterative thinking loop to have responses come in so quickly. In terms of trying to build something and trying to think through a problem, it’s a wild difference.

Dan Shipper

I know you’ve been thinking for a long time about AI as a democratizing force that allows anyone to build stuff. And in the Ultrafast world, or the world where Astra is a more expensive model, I know you’re also thinking about what it means that people pay more for better intelligence. So what are you thinking there?

Sam Altman

Historically, what we have mostly focused on—and I think done a great job of—is bringing down the price of IQ. We have made hugely more capable models that get dramatically, dramatically cheaper every few months or every year. Now there’s this new vector that is clearly going to be very important: speed. We are putting the investments in place. Our own chip is going to be an incredible contributor to getting very fast inference into the hands of everybody, to say nothing of what the models will do to improve research and our inference stack and everything else. And I think we now realize that we’ve got to invest in the democratization of this like we have for model intelligence.

(00:30:00)

Dan Shipper

So when can I expect to use Ultrafast for all my prompts?

Sam Altman

You can pay us for it and do it now.

Dan Shipper

I don’t think I can afford it.

Sam Altman

I think we can deliver 8x at a great price in the not-distant future. But then we’ll have 100x faster, and there’ll also be GPT-7 Galaxy or whatever. So I think the right way to think about this is, as far as I can see, there will always be a real frontier in terms of intelligence or speed or whatever else at a higher price. But you can also bet that we will keep doing what we’ve done: a ridiculous, put-Moore’s-law-to-shame, relentless focus on driving value per dollar up and up and up, in this big compounding way.

Dan Shipper

Amazing. Sam, thank you so much.

Sam Altman

Thank you.

Dan Shipper is the cofounder and CEO of Every, where he writes the Chain of Thought column and hosts the podcast AI & I. You can follow him on X at @danshipper and on LinkedIn. To read more essays like this, subscribe to Every, and follow us on X at @every and on LinkedIn.