models @openai · prev @ figma, @ opensea · berkeley cs · endlessly curious
Joined December 2025
- Tweets190
- Following16
- Followers0
- Likes59
as ai keeps improving, i think we need to pay more attention to where all this progress is heading, not just what the latest models can do. the pace alone is changing how we should think about long term strategy, both at openai and across the industry
Personal update: Turns out that what I thought was leaving OpenAI became more of a sabbatical.
When I left a few months ago, I felt that the most important thing to focus on was making AI more useful to people. But we’ve since entered a different regime of AI progress. One in which getting the safety and security challenges of increasingly capable AI right has to become a central priority.
So I’ll be returning to OpenAI, joining @merettm to help chart a path through these challenges.
There’s much work to do, so let’s get going.
27
we've been working on textGrain at OpenAI, our approach to text watermarking for the EU AI Act.
starting with an EU rollout, we're also giving researchers access to the detector. the plan is to open source it too, so others can experiment and build on what we've done.
We're expanding our approach to content provenance to include text in response to EU regulatory requirements, while recognizing the significant limitations of current text watermarking technology.
Our tools already help verify whether an image or audio file was created with our models. This work builds on those efforts to help people better understand when content may have been generated or edited with an OpenAI model.
In the EU, we’ll start watermarking eligible text from ChatGPT and Codex over the coming weeks to comply with the EU AI Act.
Customers using our API can turn on text watermarking for select models worldwide today.
2
40
check out the tech report and blog post below
cdn.openai.com/pdf/e9508624-…
openai.com/index/eu-text-pro…
12
plenty of companies are already doing this. the key imo is having solid review agents, extra agentic checks for critical parts of the codebase, and teammates who know how to weigh risks. most importantly, ppl who take ownership and can actually fix things when stuff breaks.
We ditched code review at @DuckbillHQ (mostly)
About a month ago, we found ourselves with 60 open PRs for a team of five. They had been accumulating for a few weeks and we all had the sudden realization we were looking at two days of just code review.
2
20
never got to play this as a kid, so kinda excited to finally give it a shot. even better that my son gets to experience it too. love how nintendo still makes games that are just pure fun for everyone.
3
58
recently read the dog stars. wasn't a huge fan of the writing, but loved the world. rainy morning in nyc felt like the perfect excuse to watch the movie. big mistake lol. half the stuff wasn't even in the book. felt like they just borrowed the title.
1
2
17
really solid take, appreciate u sharing this
new post: the senior engineer death spiral
sunilpai.dev/posts/the-senio…
a friend just started a big job and asked for some advice. so I braindumped a monologue about a super common failure mode I see with engineers and posted it here, hope it helps whoever it can.
2
11
finally back home after weeks of travel and team onsites. i'm a homebody at heart, but getting to work w/ people irl is always worth it. imo, a week every 2-3 months is the sweet spot.
1
2
24
ipadOS27 is straight up broken for me. i almost always rock a hardware keyboard, and when one is plugged in the onscreen kb should never appear.
but now the second i tap any text field the virtual keyboard pops up and eats half the screen
2
127
the pacing, how he intros the product then wraps it up, feels like listening to music. first found him when he dropped hey. just built different.this is fire.
love how these models finally let us ship the dream software we always wanted
For years, maybe a decade, I've wanted to build a writing tool that just did a few things I couldn't find anywhere else.
A system that mimicked how I think when I write. A system that provided clear affordances, features, and places to do the things I wanted to do. A way to play with my writing as I'm writing it. A way to settle into just the right way to say something.
It never made sense to make it a 37signals product because it's not commercially viable, nor would it be worth pulling people off their other work to hack on this. Plus, it would have taken months the old way.
So this weekend I just made it myself, with Claude's help.
It's called Write_On and the short video walks you through it.
Essentially it's "alternative control" at the word, sentence, and paragraph level, plus a way to dim stuff back, and stash stuff near by. Not version control, but alternative control. You'll see what I mean when you watch the video below.
So here it is. Write_On.
2
13
muse wonder full holding its own on webdev arena. solid score, but still tons of room for us to level up
Exciting news: Muse Wonder Full (Max) by @AIatMeta has just debuted in the Code Arena: WebDev at #6 overall with 1726 pts, putting Meta in the top 3 labs!
This is +75 pts and 11 places above Muse Wonder (Max) at #17 (1651 pts), and +67 pts and 8 places above Muse Spark 1.3 (Max) at #14 (1659 pts).
Muse Wonder Full (#6) also landed +16 pts above Claude Sonnet 5.5 (#17), and +48 pts above Gemini 4 Argon (#10). With this score, Muse took the #3 lab spot (previously held by Google).
Across categories, Muse Wonder Full (Max) landed:
- #4 in Data & Analytics and in Consumer Product
- #5 in Brand & Marketing
- #6 in Content Creation Tools
- #7 in Gaming and in Simulations
- #8 in Reference-Based Design
Congrats to the @AIatMeta team and @alexandr_wang on this release!
1
25
In the industry, we call this framemogging
Deutsche Bank: Muse remains significantly ahead ChatGPT in terms of downloads and DAUs through 36 days since launch.
$META
1
8
The highest-ROI move in tech that almost no one tracks is investing in your people’s growth and ensuring they feel genuinely valued and recognized.
1
9