OpenAI's Codex: A Model So Fast It Changes How You Code

Source material

Full source text

The first time I showed it to someone, they were like, "No way, this is like a fake demo.

This can not be this fast.

This will change everything."

Especially because it's not yet the fastest that we can actually get it to be.

My experience was trying the app.

I didn't really want to go back to a terminal.

What I realized is actually GUIs are great, IDEs are just a problem.

There's something that's a GUI for programming that's not an IDE.

And it seems like you're figuring that out, but I don't even know what that's called.

It's called a Codex app.

[Music] Dan here, and I want to take a second away from the episode to tell you about Granola.

Granola is an AI note-taker for your meetings, and I use it pretty much every day.

That may sound a little bit weird or a little bit creepy, like transcribe all your meetings.

Well, for me, it's actually kind of indispensable as a leader.

Every is about 20 people now, and it's really important to me that I understand how decisions get made, how I'm showing up in meetings, and how I can help my team the best way I can.

Granola acts a little bit like a leadership log for me, so I can see how I've done in meetings, what situations came up in a particular week, and how I can do better next time.

If you're trying to improve as a leader and scale your company, try Granola as your AI-powered notepad for meetings.

Head to granola.ai/every, code "every" to get three months free.

And now, back to the episode.

Tebow, Andrew, welcome to the show.

Hey, thanks for having us.

Thanks for having us.

Great to get to chat with you.

So for people who don't know, Tebow, you are the head of Codex OpenAI, and Andrew, you are a member of the technical staff on the Codex app at OpenAI.

You are the people of the moment.

They just ran a Super Bowl commercial about Codex OpenAI.

How are you feeling?

Yeah, that Super Bowl.

I was quite surprising, wasn't it?

It really was.

I think the core thing, and I think the place I want to start this conversation is it feels like that is a strategic shift.

You would expect OpenAI to have run a chat GBT commercial during the Super Bowl.

Especially if you looked at Codex's positioning four months ago for professional engineers, maybe not have run an ad targeted at a much broader audience.

It felt like for a long time, there was this divide where Codex was for professional engineers.

And if you want to do vibe coding, you do that in the chat GBT app.

And it seems like that has shifted a lot over the last month or two.

Can you tell me about that?

Yeah, I think especially, we can talk about last week.

So last week on Monday, we released the Codex app.

Immediately, we saw a ton of downloads, more than a million downloads in the first week.

And then we knew that we were releasing an extremely strong model, 5.3 Codex on Thursday.

That just made, I think, it very visible that we're here to put incredible experiences out there.

We're very committed to Codex.

And also agents are really starting to work and be able to create these things, even if you're a little bit less technical.

I think the app really showed that.

It's much more inviting for people to just try it and run multiple agents with our models being very good at allowing for multitasking and being reliable for long-running sessions.

So it allows you to create a lot more.

So it just felt that maybe we can inspire more people to build and then show that agents are here.

It's coming.

It's going to be mainstream.

Why don't you try and create something new and inspire people?

I felt like the right thing that we wanted to reinforce.

Yeah, while we were designing and developing the app, one of our internal mandates to ourselves the whole time was that we had to make something that we love to use and that we used for all of our work.

And if we couldn't do that, then we weren't going to put this out.

And this was back when we started.

And I think that we surprised ourselves a lot with how fun it was.

And especially as we started to build this app before we started to build agent skills.

And then once we paired them together, it became this really rich interactive experience where you could open the browser or you could connect to these various services.

And so all of a sudden we started to feel this really connected interactive experience and wanted to share.

I see the ad as a love letter to builders.

I have never seen a Linux CD in a Super Bowl ad.

And so that was really cool to watch.

What was the impact of the ad?

We're still to measure that.

We'll see how it plays out over the long term.

But we saw a giant surge of traffic, actually, remarkably, very, very quickly after 4 p.m.

like PSC when it aired, the surge and our systems were under heavy load.

So it felt kind of weird to me.

People are watching the Super Bowl and then going and installing the app and they're just trying it out right there and then.

But it happened.

And a lot of people reached out and saying they were really inspired by it and just wanted to build afterwards, which is what we're aiming for.

Tammy, I still want to talk a little bit about the strategic shift.

So, codecs in general are moving from something that is really for professional developers moving to something that has a broader audience and maybe moving some of the vibe coding from Chagipati into the codecs app.

Tell me about that.

I don't think we're trying to move vibe coding from Chagipati into the codecs app.

Like we're very much, you know, two things are happening.

Like one, we're pushing the frontier on like professional software development.

Like 5.3 codecs beats every single other model on the top benchmarks for coding.

So it is a very, very capable model.

It's also like at the speed and cost.

It's like, you know, it is a top performer.

I think the app, the second thing is like the app does make things more accessible.

And so like it does appeal to like a wider audience.

But internally, we're also seeing the app, you know, just it is very much used within research within around team, like the entire codecs team uses the app.

It makes people more productive.

So it's like very much leaning in into, you know, how we think agents are best used, the patterns that we were seeing, you know, that we're making people like very productive here at the company and outside.

And then just sort of like going all in on that.

It does happen at the same time.

Also, it's like, hey, it's just delegation is finally here.

It works, you know, it's like much more accessible.

And we're going to try and see like how we can package that and actually ship this to like, you know, a much, much wider audience.

But that's that might not be the codecs app.

When you use that all that is like you build in there, 99% of the code that I write is using the codecs app.

Same.

I mean, I live in there now.

Yeah.

Okay, well, that's that's actually really interesting.

I definitely want to talk about the app in particular.

But I want to go back to the thing you just said, which is maybe if I if I'm reading you, right, you're you're kind of like, we're pushing the frontier.

We're seeing lots of people who are maybe broader than just like senior engineers using us.

However, the overall idea of like, who is doing what in which app like maybe you haven't totally figured out yet.

And it's not as clean of a line as like, no longer vibe coding and chat GPT or really vibe coding and codecs.

It's like, you can do it in both, but we haven't figured out exactly like, which thing you're going to do where.

Yeah, I think critics is like the most powerful experience right there.

So you should be fairly technical so that you understand like, hey, you know, code is actually getting written, you know, it's going to get executed on your machine by default is executed in the sandbox.

But you should probably be able to read code in order to use, you know, codecs to like its fullest.

We will bring a similar experience to chat GPT at some point, which will have like different properties in terms of like the sandbox and like how concepts are represented.

Maybe we won't be showing like, you know, hey, it's like this scary terminal command like thing is like running and you know, you should probably approve it.

It's like, you know, of course you shouldn't do that to someone who is not technical.

And critics is really there to like appeal to, you know, just all coders, builders, you know, technical, like people who are close, like either technical themselves or like technical adjacent, you know, like data science, these kinds of things.

Yeah.

And, you know, if you use the codex app for any amount of time, you can see the inspirations from chat.

You know, the layouts very similar.

We auto name your conversations.

We've got contextual actions, but it's pretty clean.

Right.

The composer looks very similar.

And you'll see some of that inspiration back in chat for other types of things.

But we still believe that some, you know, when we set out to make something that was for the professional software developer and for us, that it deserved a dedicated experience that could could really showcase the power of the models and the way that the models could change the development lifecycle.

And so we, you know, we made something very tailored to that.

And we've, we've had a lot of success internally with research teams, with product teams.

And so, you know, we're, we'll look beyond, but I think we're really happy with where we've ended up on that kind of tailored, the tailored approach to this.

Can you tell me about the decision to invest in a GUI over a TUI?

I feel like two is are so hot right now.

And obviously you have one for Codex already, and you could have said, okay, we're going to double down and just make the make the terminal terminal experience even better than it is now and really invest in that versus, okay, we're going to go, you know, I think, yeah, making a GUI is a little bit of like a counterintuitive or like counter narrative thing to do.

So tell me about that decision process.

I think it wasn't counterintuitive.

It's more maybe it's not mainstream.

And so we, we experiment with a lot of different approaches.

I very much consider that we're still in the experimentation phase.

And, you know, we're responsible primarily for two things is like, you know, building the most powerful entity out there, you know, that's capable of coding.

And then, you know, increasingly, this will become like a multi agent system, and it will become like more and more capable.

And, you know, you will have to figure out like how to steer and supervise like its outcome and its behavior, you know, that's like one thing that we're building.

And then we're also building like, how do you even, you know, interact with this is like, you know, what is the optimal way to have visibility into what this like very capable entity or like system of entities is doing?

How do you steer them?

How do you supervise them?

And so we, we're very much still experimenting with what that is, it's like, you know, sure, you can do it in the two, it's like, at some point, it starts to feel like very limiting, you know, especially on like multimodal, like, you know, actually, like, the model scan, like draw little diagrams and generate images and, you know, or, you know, you can talk over it, you know, using voice, maybe you have like, many of them going in parallel, and so you start to lose track.

So we felt like we needed to start experimenting with something else.

And it is only when you know, we started become like super, super popular internally, we were like, we have to shift this externally, like, this is kind of like, this has come to a point where it's like too good to sort of like, just keep it to ourselves.

I mean, that was like the journey that you went, you know, you were not building in the app.

Although, like, when did you start building in the app, that was actually like fairly quickly, like when the app was building itself, that that was, yeah, that was pretty quickly.

And yeah, because I was starting with the two E and with the ID extension.

And I think that my goal personally was how can I get to fully building the app on the app as fast as possible?

No, it's like, it's really easy when building this stuff to slip into the mode of like, Oh, this will be good for somebody.

Like somebody would love this a certain type of like, they will love this, right?

So we really wanted to get quickly to like, I want to be able to build the app on the app, I want it to be able to run itself with skills, I want it to click around on the app that it spawned.

And I want this to be like part of my workflow as soon as possible.

And they're like, I still use the two E sometimes when I want to fire something quick.

But I think that like, there is something about the flexibility of controlling the UI, and being able to have some pains be persistent and others be ephemeral, and be, you know, we shipped voice with the app.

So you can prompt with with voice.

We have mermaid diagrams in the app, we have full image rendering.

So all of those things, I think, are like the tip of the iceberg and what we want to do with a dedicated UI.

And it's it's pretty simple.

And it's simply intentionally, but I think we're going to do a lot with dynamic stuff there.

I mean, yeah, the ceiling is just much higher.

Yeah, it's interesting.

My experience was trying the app, trying the app, I didn't really want to go back to a terminal.

And I had been coding most mostly in cloud code and some in some codecs in the terminal for the last like, for several months before that.

And I think what I realized is, actually, GUIs are great IDEs are just a problem.

And like, there's something there's something that's a GUI for programming that's not an IDE.

And it seems like you're kind of in that figuring that out.

But I don't even know what that's called.

It's called a codecs app.

You know, there, there, there was a moment during the development of this, where everybody and their mother was forking the same IDE.

And we, we kind of looked at each other.

And we were like, Hey, should we have done a fork of VS code as well?

Like, very seriously, I remember exactly which day it was.

And I think I don't know if I don't know if I would say that IDEs are the problem.

But I go back to like the truck analogy sometimes with them, which is that like, I will open an IDE here and there like I opened one today.

It was something very specific that I wanted to do, that I don't even remember what it was.

But then I closed it.

And I went back to using the codecs app.

And I think that there is something there with like the codecs app being a great daily driver.

And like occasionally you need an idea, occasionally you need like a really complex terminal setup, but that this should be your home base, it should be your command center for the agents that are running, and a place that you can come back to and track all this stuff.

And, you know, there are a lot of design decisions around like, do we allow freeform panels like an IDE?

And we kind of came to the conclusion that a lot of what these models are great at is knowing what is needed in the moment for what type of task.

And so we wanted to have kind of more full control over what was able to show at what point, right?

And you can see that in plan mode, where you're not necessarily getting a composer, you're getting a really quick way to answer questions.

You can, you know, and you've got your plan and you can edit your plan.

And I think we only want to do more with that as we go.

It seems like you were surprised that you didn't want to go back to the 2E after.

I was.

Yeah, is that...

We're like a, like, Greg didn't review and Greg was like, I am a 2E power user.

I thought I would never leave the terminal.

Yeah, Greg lives in Emacs.

I was a 2E power user for like six months, starting with like when Cloud Code first got really good.

And I was like, holy shit, this is so much better than being in Cursor or Windsurf or whatever.

And now I feel like I speed ran my 2E era and I'm back in GUIs.

Like, I'm kind of setting back and forth right now, but I can sort of see the light where it just, if you're, especially if you have a bunch of them going at once, the affordances of GUI are just like, make it much nicer.

Yeah, and there's a lot more to come there.

And it was a very intentional thing for us.

Like we sort of see agents will act and are already acting on much more than code.

And so they need to be a companion to like every single app and every single thing that you can do on your computer.

It's like we integrate with like linear or Slack.

And of course, they also need to be able to read the code and produce code, but maybe it can do like a deploy reverse cell as well.

And like, are you going to do all these things from the IDE?

That would sort of like feel very odd.

And so it's like, it's like this command center for your agent.

We optimize the entire experience around that, you know, around the idea that you have a very capable, intelligent entity that you're like controlling, steering and supervising.

And you know, you never need to like sort of like go in there and like, you know, do the things yourself.

It's like, you know, the thing is very capable of like, you know, being delegated to like, I think, you know, when, when you accept that that is like, you know, what we're headed towards and like, you know, with factory codex is like, you know, it just feels like, you know, getting almost there, right?

Then you're like, well, you know, it's the same with you, right?

You know, like when I talk to you about like a feature ID or something, it's just like, you know, you go and you get inspired and you go and do it.

You're just like, you know, I don't suddenly jump into your ID and like, you know, just go and like implement it.

You could.

Yeah, I mean, I think you would find it disturbing, right?

It's like, so that's the way that you will, you know, everyone will work with agents.

It's like, you just talk to them.

How has your workflow changed with five, three codex versus five, two?

I was surprised at how much faster it was.

And sort of like, I have to adjust on, I had been optimizing a lot more for like long running sort of like multitasking.

And, you know, I sort of like had an explanation of like, okay, this type of task will take like, you know, 10, 15 minutes.

I'm going to kick like, you know, four, like, you know, different things and then come back.

So I'm able to like, you know, maybe do a little bit less multitasking and like, you know, be more in the flow.

So that, you know, felt really good.

And then it just feels now very satisfying as well.

Like, you know, to kick off like automations with it, using skills, it's like, it's a more generally capable model.

It's like less sort of like super focused on code, right?

And so I find it like much more reliable, like, you know, sort of like going through like Twitter replies, and like, you know, summarizing like the important teams or like filing bugs in like linear, and then, you know, coming back to that and using automation so that, you know, things are like implemented like daily feels like it's like much more robust for these things.

I mean, but you're really like the superpower user here.

And there's like, you know, it's just like the kind of stuff like, you know, he does it's just like, you know, it's like, I have very vanilla usage of codex compared to Andrew.

No, I mean, well said.

I had a series that I had intentions to run this for a while.

And I only ran it for three days on on on x Twitter, which was that I was I was setting up a prompt to basically add a feature to the codex app, like some random, like non shippable feature to the codex app.

I had this long prompts like about the quality bar that we had to do.

And once I switched it to five three codex, the results got actually much more interesting.

Like we did a subway surfers panel on the right was one of them, like a little Minecraft UI for the sub agents was another one that we did that I don't know, maybe, maybe we'll should have like, get back to work.

Yeah.

Why do we have Minecraft in the credits?

No, but can I explore?

No, I mean, five three codex, like it's, it's, it's neat.

It's fast.

It's capable.

It's multimodal.

What are T.

But says you have a lot of cool use cases, like what are what are the like more interesting ways that you're using the codex app that maybe people should try but haven't thought of yet.

Andrew came up with automations.

And I think that sort of like, shifts the way that you know, you're thinking about these things when they can just like sort of like hop it into background, you know, on a specific trigger at a specific time.

And then you know, just you can sort of like program it yourself.

Yeah, they're using that a lot.

There are a lot of things that I use the app for that are a little bit outside of just like coding features.

I keep it to I use it to keep my PRs mergeable with automations.

And so it'll resolve merge conflicts, it'll keep them updated, it will fix like build issues so that basically, like, as soon as they're ready to go, like they're ready to go.

There's no like, Oh, hey, somebody merged a big thing.

And there's a conflict now.

So do that.

So you said like, so at what point is the is the automation trigger?

Because I thought the automation triggers like at a certain time schedule, but it sounds like there are other triggers I didn't know about.

I yeah, I we're looking a lot of things I have it right now just on a time schedule.

And I use our GitHub skill and some internal skills for our CI.

And that, that runs hourly or every two hours and kind of just cleans everything up.

I see.

So it's like, through all you know, there are any changes on main and it just looks through any PRs and just like make sure that they're all up to date so that whatever you're ready to go, it's never like a section that's like that.

Yeah, it's it's actually really helpful.

It's surprisingly helpful.

I have one that every day at like 9am.

I get sent all of the contributions that have emerged to the codex app over the last day.

And so it'll do like a nice report of who merged what and it will, I have a group it by theme.

So you'll be like, All right, like three people worked on this part of the composer, two people worked on automations, like here's what happens so that I can at least be like, knowledgeable what's happening because things, things get chaotic.

Right before launch, and one automation I have is, it's I run it like multiple times a day.

And it's like, pick a random file and find and fix like a subtle bug.

And then it's kind of funny because it actually does pick a random file.

So it will run like, you know, Python like random and then it will like, you know, find a random file and it will start from there.

And so it's like every time it's like explores like a new one.

Has it caught anything?

Oh, yeah, yeah.

It's like we catch like, it's often latent bugs that you know, are not triggering actually, like on the critical path, but you know, they're actually bugs.

And then you know, just like trivial to fix it, like merge it takes very little time.

And it's a thing that you know, I would have never found myself found like an issue and I constrained sampling like the other day.

Yeah.

That's really cool.

Do you have other other automations that are worth sharing?

Let's see.

I feel like I have 60 that are running at all times.

Some for testing and some for real.

Some of the members on the team really like this one that looks at the PRs that you've done in past day or so and quietly cleans up any bugs you shipped and kind of like looks at a few of the observability platforms to see and like tries to basically ship a fix before anyone's noticed that you shipped a bug.

That's cool.

It's not coding related, which is like marketing research.

It runs daily.

And it's just sort of like it's prompted with like a specific skill to do like deep marketing research.

Which I've like sort of like tuned over time.

And then that just goes on like searches the web on you know, any sort of like new things that sort of like came up in terms of like how you know, just like how users are like receiving talking about codecs.

And then I just received that little report.

And it always makes for like an interesting read.

Yeah.

We can just go on.

It's like these are just examples that you know, we do rely on.

You know, they run.

Yeah.

Yeah.

Do you have any particular skills that you guys like that are beyond the normal kind of you know, I have a GitHub skill and that kind of stuff.

I love Andrew's Yeet Yeet skill, which it just like takes like the change.

And then you know, does the commit does the PR rights like the draft that puts it in draft and like, you know, publishes a PR with like a PR title and body.

Yeah, it's very satisfying.

Yeah, it just does everything.

That one is like makes definitely makes people like productive.

What are the top ones for you?

Image Gen is a cool one.

Yeah.

For both like silly automation purposes, like hey, make me an image that characterizes my last day of work.

Not my last day of work my previous day.

Yes, Andrew.

I you know, the the image Gen skill was actually really cool.

For I use the codex app to make a book for my daughters.

And so I had I like, you know, put together this prompt for teaching it about like a script that I wanted written, said like 24 pages, here are my daughter's ages, here's like where we've lived in the past, like we were in Boston and moved to New York and then moved over here.

And then I said, like, after that, we went through that I agreed on the script.

And then we went through and I said, like, all right, now it's time to use the image Gen skill.

And it made like it prompted for every page in the book based on the script, it prompted for the image, and then it kind of put them all together and use the PDF skill to put together the books PDF.

And then I printed it.

And so we've got like a super custom book that, you know, I read to my kids.

And it's it's really cool.

It's just this awesome thing when you can combine like the intelligence of like the agent and then it's like, like works in a programmatic way, like known by using skills, and then you can just combine them in like novel ways.

And like, yeah, I think the PDF and image Gen one is like, it's a common combo that we see.

It feels like the codex model, it obviously has got faster, which makes it much more usable.

And it also feels a little more opusy, like it's a little more has a little more emotional intelligence, but it still has a little bit of that like, it does exactly what you say thing in a way that is a little it can be annoying.

How are you guys thinking about how you shape the way the model feels and which way you're pushing it?

It's something that we obsess over.

So we definitely want the model to excel at coding and be really good at instruction following.

At the same time, when we optimize a little bit too much in that direction, it can over index on specific words or so like misunderstand the intent in ways that humans wouldn't.

Sometimes I will just like have a typo and then the typo, like, you know, actually find its way into like the file.

And I'm like, obviously, I didn't mean like, the typo is like, I meant like this name of this class.

So that's something that we're definitely continuing to push on.

But like the thing that we're pushing on the most right now is like really efficiency, you know, speed, and then also like what we now refer to as like, personality is like, you know, how supportive is it?

And we understand that not everybody has the same preferences there.

Like the previous default, you know, was definitely like super blunt, like pragmatic personality.

Now we've also introduced like a more supportive, like friendly personality, and you can just like, pick between those.

I think for things that don't have like, sort of like a universal, like, accept a, you know, thing that, you know, everybody that you know, should just use is that, you know, probably going to introduce like some way for you to just make it your own, right, you know, you should feel like you have your own little personal critics that, you know, works in exactly the way that you want it to work.

Do you use the friendly or the pragmatic one?

Pragmatic.

Pragmatic.

Yeah, okay.

I'll say it's pragmatic.

Interesting.

I think you guys recently put out a model that is so fucking fast.

I was testing it before it came out and I was just like, I can't really keep up with this thing.

So I'm curious how that changes how you think about what is now possible with coding with a model like this, and also the affordances that you need in order to manage models that are so quick, effectively.

Yeah, the first time we used this model in the app, we had kind of that same thing happen where all of a sudden there was just like this wall of text and we are at the bottom of the scroll and we were immediately like, all right, we need to smooth this thing out coming in.

And so we actually do slow it down ever so slightly, just so that you can see the words come in like a little bit smoother.

That's so funny.

It's like a really funny problem, but this thing has been super fun.

And I think what I'm most excited about is what sort of capabilities we can start to add to the app that are really, really dynamic that we couldn't with a model that wasn't this fast.

So yes, this model is going to allow you to iterate really, really quickly, but it also opens up a lot of new opportunities to how you code and how you interact with the codecs app.

The first time I showed the very first prototype when we hooked everything up, and obviously the model is powered by a series, and we've talked about the partnership there, and we're very excited to put the first model that we're serving through that out there.

It's obviously still very early.

It's literally the first time we hook it all up and we're just so excited that we want to share it.

But the first time I showed it to someone, they were like, "No way.

This is a fake demo.

This is not real.

This cannot be this fast."

And then they tried a few prompts.

They were just like, "Oh, I literally cannot keep up."

It's like, "This is insane."

And yeah, I think this will change everything, especially because it's not yet the fastest that we can actually get it to be.

With the preview, we're putting it out quite early.

We're actually going to layer a number of optimizations on top of it, which should be able to make it maybe two to three acts faster than the experience that you have experienced.

So that's going to change things.

And we're thinking about this also from a point of view of delegation.

We think this model has a huge role to play as part of a system of multi-agent systems.

And as a way to speed up maybe the slower, more intelligent agent as well.

So we're going to be experimenting in that way.

And do you expect the same hardware speed ups on the more intelligent agents to come out soon?

So a lot of the things that we worked on were interesting, so distributed systems and infra problems that we uncovered because we were able to sample from the model at unprecedented speeds.

And then if you're getting tokens back this fast, you need to go and optimize the entire set of bottlenecks that you sort of uncover on the critical path of serving.

All of those benefit the current, they benefit like TPT53 codecs and all future models.

And there's one thing that we've been doing as well, which I'm sure we're going to put in a more detailed blog post at some point, which is we rewrote the entire service stack to be based on web sockets and a persistent connection and to do things a lot more incrementally and statefully.

And that decreases the overall latency across all models.

We haven't chipped it by default yet, but it is something that we are making the default for this new super fast model.

And then we're also going to enable on the other models.

And it decreases overall turn latency by something like 30, 40%.

We can look into the exact numbers.

What are the most surprising things that you've seen using the model internally in terms of what a speed up like this enables?

It just allows you to be super, super in the flow.

And you're almost like just in real time, you know, sculpting the experience or like the code.

It's just a very different feel to it.

It's very unsettling at first.

And then once you get into it, it's very hard to go back to any other model.

That's like the feedback that we've seen.

That's like what I felt myself.

And so it's like this very, it takes like five minutes to adapt.

And then you sort of like know, okay, it's like, this is how I'm going to use this thing.

I also don't think that we've poked at the full extent of what we could do with it.

Yeah, it's very early.

We haven't had it for very long.

Yeah, someone on the team like Channing was just sewing like, Oh, yeah, it's so fast.

And it can actually like play Pong, you know, not very well.

But the model is able to react to things like, you know, almost like real time, right?

It like, you start to see how it might replace some deterministic steps.

So we have, we have in the codex app, a set of get actions, right?

And as everybody knows with get, like certain configuration of things or certain states that you can be in, can make it really hard to run those without a ton of error handling, and like all sorts of like error messages and guidance.

And it's really hard to create a good experience, which is why like nobody ever has.

But if you have a model that's as almost as fast as running these scripts, then you can imagine a world where these things turn into skills, or something like that.

And you can have your operations run a little bit differently with some, like some intelligence, and, and not have the same latency that you have today, when you're asking it to go track something down the codex, right, you can kind of like vaguely gesture and be like, Hey, like send this up and have that be fast enough for a button.

What I'm very excited about is like when it's going to come together with, you know, one thing that we shipped with five, three critics as well as like this thing that we call like mid-turn steering, you know, where you're, you're just you start with your prompt, it's like it got to work, and then you send another prompt, like, while it's still working, and it adapts, like in real time as well, like it will just sort of like receive that message, acknowledge it, and then you know, continue its work.

Like if you start to think about, okay, what would this look like with voice, and then with a model that is as fast as the one that we just shipped, then that's like a whole other experience that you know, we would be very excited to bring, you know, hopefully very quickly.

Because you can easily interrupt as you're, yeah, if you're just talking and engaging with like, you know, natural language, and then doing the mid-turn steers, and then the, you know, the implementation happens like almost instantly, because of the speed, it's like it becomes like a very pleasant thing to use.

Like right now, you can sort of emulate it, you know, with like voice, voice dictation, and then send it and mid-turn steering, and then, you know, what's the model implement, and it's like a very cool thing.

I think we're going to have a step change in that experience when we just like really just polish it.

If speed as a bottleneck is like, close to being solved, what do you think is the next bottleneck?

What is the next limit on making the thing you want?

The bottleneck that is very apparent is like, you know, how fast can you verify that things are correct?

So like we, I mean, we can generate like, code faster than ever before, we can implement entire features, and you know, we like, I saw like someone just based on a description of, you know, the codex app, if you sort of like synthesize that into a plan, just based on screenshots, like the models are very much capable of like, reproducing 95% of the features, and just rebuilding the app from scratch.

Now, is it going to be bug free?

Is it going to, you know, is everything like implemented to like, you know, perfection in the same way that you know, the actual app is, like that takes like a lot of time still, like, you know, for like a human to go and click and verify and you know, make sure that you know, it's like, the designs are like, consistent and to that, you know, there's like no bugs here or there, that the settings panel, like, you know, when you click that button, it actually does the thing that you expect.

I think verification, you know, definitely becomes a bottleneck, like we have people on the team like complain, you know, like, there's too much code to review, it's like, you know, it's what we're trying to solve for.

I mean, you complain about that.

I complain about that.

There's so much code to review now.

Both that like, like on your own machine, and like from another peer, it's like, we're gonna have to figure that out.

Yeah, you're already reviewing, you're reviewing the code the first time because the agent is just presenting it to you.

And then you have to review, you know, the code produced by your peers, you know, or like, there's like these two rounds of reviews.

Yeah, I mean, this is something that we're working on.

A lot of us still do have to review code.

And we want, you know, we're taking a look at what that experience should look like with the model involved, right?

We've got a review mode in the codex app that works really nicely and kind of annotates your diffs on the side with findings and stylistic things and lots to do.

Yeah, it's one thing I'm sort of like, also excited, like, about, you know, like making the models faster, and then this, like, you know, this one that we just put out, it's like, you know, which is mind blowing really fast.

It's like, you can also use it, you know, you can imagine using it, like, in a way through understand code, understand features, you know, helping you with code review, like helping you understand, like, you know, the code that appear really, and it's, it's like, much more pleasant, because this is something that you want to do, like, you know, you want to be there in the flow, it's like something that has to be like synchronous, it's not something that you delegate, you cannot delegate understanding, right?

It's like, you know, you're trying to, like, you know, get to understand something.

And so, like, speed there, like, is a real advantage.

So, it's sort of like, helps offset as well, like, you know, the fact that models are, like, producing more and more code is, like, you know, speed helps you understand, you know, this code faster as well.

Yeah, I mean, I definitely think I've found this already with this, with this new model is speed, especially for end-to-end testing is faster, because if you're having it do end-to-end testing, like, manual integration testing, often there's, like, a toast that pops up, it pops up for, like, a second, and if the model's not fast, it's not going to get it.

And it seems like it's better for that, because the cycle times are much, much shorter.

So, and I definitely find this too, it's, like, I can produce so much code, but when I see a PR come in, or when I make a PR, my first question is, like, is there evidence that you've actually tested this, and this actually works, like, not just unit tests, like, you've gone through it by end-to-end?

How do you handle this?

I mean, I've seen a lot of PRs that I have the same question about.

It's, like, it's so easy to code things now, right?

Yeah, I mean, we have gotten the codex app to be pretty good at, through some skills that we have of running itself, clicking around, screenshotting itself for evidence, and uploading it to the PR.

There's, there's, like, a lot that's pretty interesting there, especially when we make this, like, more async, or when, you know, the models get really fast at this stuff, like, I don't know exactly what it looks like yet, but there is a lot there around, like, hey, here's a bug fix, this is exactly, like, what it looked like when it was happening, and here's exactly what it looks like now with the same exact click path, and so, like, maybe that's the turning point that code review becomes less important when it's, like, you can verify that part instead, so you have to, kind of, like, do less through the code as a proxy, but there's definitely more to explore there.

Last couple questions.

I'm curious, what have you guys learned from Aroppic and Cloud Code, and how do you think about your positioning in the market versus them?

Like, how do you think about the differences?

I think they were first to put something out there, and that was interesting to us, because we had been working on similar ideas for a bit, but I think our models were a little bit, at the time, not ready, like, you know, they were not, like, reliable, like, on long horizon tasks, like, you know, they were not able to, like, do, like, reliable tool calls, and, you know, stay on topic, and so, as soon as, like, we started to, like, really invest on that, and, you know, especially with GPT-5, is, like, you know, we were, like, okay, the models are there, we know how to make them even better, 5.2, like, brought, you know, even, like, better, like, long context, long horizon, like, reliability, non-context understanding, and what we were seeing is that Aroppic was sort of, like, you know, to us, like, losing a little bit of steam when it came to the model, and we were in this fortunate position where, like, the way that we run Kudex is, like, you know, we've got, like, product, we've got engineering, but we've also got research, and we just, like, all work together and sit together and solve problems together, and it's, like, a highly creative space where, you know, at times we decide to, like, solve problems in the product, in the harness, but at times we also were like, hey, how can we actually improve the model, and, like, let's just, you know, talk about it, and, like, you know, idea together, and then, like, research will come and be like, hey, you know, we've got this, like, picture that we're sitting on, it's like, would this be, like, sort of, like, something that we can ship, and then it was just sort of, like, get excited about that.

One of the examples was we had a lot of complaints on compaction, you know, compaction was, like, something that people felt, like, whenever you would hit compaction, you know, people would complain, it's, like, it's losing too much context, and so we sort of, like, solved that end-to-end, and, like, you know, we decided to do, like, end-to-end RL training, and, you know, introduce compaction, like, within research, and then, you know, make the model, like, you know, itself, like, you know, very familiar with the concept of compaction, and, like, producing, like, optimal, like, sort of, like, delegating to itself, like, across time, and, you know, once we had that, and we had solved it at the model level, like, sort of, like, the harness problem became, like, so much easier, because it was just, like, oh, just let the model do it, and it's going to be, like, very reliable.

So, through that, and, like, through that collaboration, it just felt, like, like, the momentum has been, like, very strong, and I was so, like, able to improve, like, models, and, like, you know, ship a model, like, roughly, like, on a weekly, a monthly cadence, and then we took, like, a bit of a different bet, and, like, a different approach with the Codex app, which turned out to be, like, you know, an awesome thing, you know, to just try and do, is, like, and not just, like, sort of, like, force ourselves, you know, and, like, trying to cram everything into the 2E.

I mean, it was, like, it was, like, a great challenge, right?

You know, you were, like, I'm, you know, it's, like, let's build an app, like, just, like, where do I get started?

And then, you know, just, like, you just got obsessed by it.

It's hard not to.

Yeah.

I mean, it's, like, how was it to just, like, you know, build something that was quite contrarian, I suppose?

Yeah.

I mean, I remember you and I talking about whether or not, like, early on, we were, like, we don't know if we'll ship this.

Yeah.

Like, we'll try it out.

We'll see if we can get there with something that we love and see if we can get, I remember saying, like, let's get some PMF internally.

Let's get everybody at OpenAI to want to use this thing without being forced to use it.

Let's see if we can do it, right?

We did.

And it was, like, adopted very quickly.

I mean, the minute it was barely usable, the research folks, like, put dev boxes on it, right?

Like, which was, like, this crazy hack at the time.

Yes, yes.

But now they use it, like, for everything.

Yeah.

Yeah, it was, like, including and training, like, five to three codecs.

And so, like, I think I feel really good about having hit the point where, like, you know, everyone technical at the company, like, almost everyone technical at the company, like, uses codecs.

But, like, the people who use it the most are, you know, actually building codecs and building the models.

And so, you know, we're just able to, like, you know, improve things at, like, crazy, crazy speeds.

And, you know, there's, like, no signs of it slowing down.

Amazing.

Well, I'm excited for what you ship next.

Thank you, guys, for your time.

I really appreciate it.

Thank you.

Thank you for having us.

Thanks.

[Applause]