Codex @OpenAI. Previously: Cash App & Block, YC S17, Google, several startups.
Joined August 2008
- Tweets6.1K
- Following355
- Followers5.3K
- Likes58.9K
Pinned Tweet
A personal update: I’m joining OpenAI to work on Codex, the best agent app in the industry.
After years building consumer products, my obsession since September has been to make Block one of the most AI-native engineering orgs in the world.
I’m proud of what we’ve built: faster Eng velocity with RPI in October, an internal multi-harness ADE in January, Slack-based cloud agents in February, and AI woven into every step of product development. The team is just getting started.
Along the way I tried almost every agent orchestrator out there — 30+ over the past year. Codex felt special from the day it launched. Lately, it’s started to point at something much bigger.
What used to be scattered across prompts, tabs, tools, files, and handoffs is starting to come together in one place. The app is becoming the single surface for knowledge work — letting you accomplish more than you ever thought possible, while staying calm and focused.
You can already see the early signs: making a slideshow and editing it right there, working inside a custom site built just for you and your team, handing off to the browser and the computer while you watch it go.
Getting this right takes powerful models. But it also takes taste, product thinking, and real care for the person on the other end. You see that in the product, and you feel it the moment you talk to the team.
I couldn’t be more excited to join the Codex team. Together we’re going to build the future of computing and knowledge work, and we’re going to build it for you.
Can Sar retweeted
Thariq has always been my favorite Anthropic engineer! He consistently shared incredibly useful AI tips, insights, and suggestions.
Wishing him all the best in his next chapter! I’ll definitely be following whatever he does next!
I’ve left Anthropic after more than a year of working on Claude Code and seeing it scale larger than I ever imagined.
I have really loved working at Ant and believe deeply in the mission, and most of all, the idea of a race to the top.
The idea that if we work hard and responsibly, we can raise the bar and inspire others to great outcomes. And if we all push each other to be better, things can be radically good. I will carry the spirit of this with me forever.
My colleagues at Anthropic are the most talented, thoughtful, competent, and earnest people I have ever met. I am glad they are at the wheel and trust them deeply.
To the community on X, thank you for everything. You’ve made this the most fun (and sometimes stressful) job I’ve ever had.
For me, it is time for a bit of a break before diving into something new.
If you want more excellent Claude Code content, you should follow @addyosmani , @lydiahallie, @_catwu, @ClaudeDevs and of course @bcherny who started it all.
Can Sar retweeted
Caring is the moat, you should feel visceral pain when things are bad or broken and stop at nothing to fix them
Can Sar retweeted
I’ve left Anthropic after more than a year of working on Claude Code and seeing it scale larger than I ever imagined.
I have really loved working at Ant and believe deeply in the mission, and most of all, the idea of a race to the top.
The idea that if we work hard and responsibly, we can raise the bar and inspire others to great outcomes. And if we all push each other to be better, things can be radically good. I will carry the spirit of this with me forever.
My colleagues at Anthropic are the most talented, thoughtful, competent, and earnest people I have ever met. I am glad they are at the wheel and trust them deeply.
To the community on X, thank you for everything. You’ve made this the most fun (and sometimes stressful) job I’ve ever had.
For me, it is time for a bit of a break before diving into something new.
If you want more excellent Claude Code content, you should follow @addyosmani , @lydiahallie, @_catwu, @ClaudeDevs and of course @bcherny who started it all.
Can Sar retweeted
When you buy a house in the suburb, you buy one house. You do not buy the entire neighborhood, town, county.
I understand people don't like change around them, but until those people pay for other peoples' mortgages and taxes, there is no reason to disallow density in suburbs.
I think most people would be okay with development on top of BART stations. I would.
What normal people don't want: towers next to their single family homes.
And that makes sense, because why would you chose a more suburban neighborhood if you wanted to live near towers?
Is it too much to ask that people who chose to live in the suburbs get to keep it that way? And is it wrong to keep towers in the places where people chose to live by towers?
Or is everyone who wants to live in a suburban neighborhood a bad person that deserves to have their chosen lifestyle destroyed?
Can Sar retweeted
The “alignment *to whom*?” question IS important, but it creates the illusion that the primary questions in alignment are philosophical or political rather than scientific. The current high-order question in alignment is, “can you align a smarter-than-human intelligence *to any set of values*?” instead of “but whose values should it be aligned to?” I wish we had the luxury of having the rather more familiar political/philosophical debate of “to whom?,” but at this moment we do not.
Can Sar retweeted
This is actually a pretty great deal.
A bit more than the price of Astra for nearly Astra intelligence at 8x speed
Voice mode and computer use with ultrafast is such a magical experience and a whole new way to experience working with llms
Can Sar retweeted
There are a couple of universal patterns that show up across fields that you should apply to pretty much anything:
- Use a written process because memory isn't reliable (see aviation checklists, surgery, mise en place, SOPs...)
- Stop early when something is wrong (Andon cord, feature flags, turnaround time, error budgets...)
- Leave a buffer (Margin of safety, safety stock, taper, defense in depth...)
- Fix the system rather than blame people (Blameless post-mortems, poka-yoke, Swiss cheese model...)
- Get feedback fast (OODA, kaizen, canary releases, after-action reviews...)
- Hand over with a fixed format (SBAR, BLUF, commander's intent...)
- Ramp up gradually (10% rule, canary releases, position sizing....)
Can Sar retweeted
It's moves like this OpenAI feels to be pulling ahead of Anthropic.
Anthropic is closed everything: closed source Claude Code, no you cannot use other models with CC, and so on.
OpenAI: Codex is open source, you can use other models; oh let's make it dead simple to use GLM-5.3:
Can Sar retweeted
still thinking ab this quote @jxnlco referenced in his devday talk the other day
“i want ai to do my laundry and dishes so i can make art"
Can Sar retweeted
People seem shocked that the author of Clean Code doesn't read code anymore. But that should not be surprising.
Why did I write Clean Code? What is the goal of keeping code clean? To get it out of the way of the real job -- thinking.
Can Sar retweeted
Launching a new Codex Cloud, much improved from last year. With configurable cloud environments, it’s impossible to go back to building on your laptop once you’ve taken the time to configure it.
Agents API, the same tech powering all our cloud agents, including dots, is also now in preview and supports computer use. It allows you to build the same incredible products we are making available today.
Can Sar retweeted
Over the past 5 weeks, the team completely rewrote ChatGPT Web on top of the desktop app codebase.
This was only possible with Astra, and it’s why Dots, Space, Codex Cloud, and so many more of our launches will come to web and desktop same day.
Era of the megaproject begins.
Can Sar retweeted
Remote is now Codex again! And Codex Cloud is just like any other host! Probably my favorite launch today!
I hear we’ll be launching some big things today.
Among that, we’re also launching 2 tiny bits of polish I’m really excited about (and added over the weekend). Just 2 of many hundreds, but that’s how you build a great experience.
Nothing is perfect on Day 1, but we always try to get there ASAP.
Can Sar retweeted
It takes a while to build the confidence, but there's no future where you're manually reviewing every line of agent code. Not much acceleration in that. You need adversarial agent reviews, you need automated testing, and maybe you spot check. That's it. From prompt to production!
Can Sar retweeted
This is probably one of the most repeated worst takes of all time.
Generally if your brand is good enough you can get into most deals if you’ll pay the most. Obviously when you’re paying with stolen money things are cheaper. The FTX venture portfolio can best be understood as a drug lords supercar collection. It’s neat that he had some taste and that they generally have appreciated but it’s immaterial to the core issues with the core business. The most important part of investing is not blowing up which would be a charitable way to frame things.
