rapha (@rapha_gl) on X (original) (raw)
making models programmable @ openai

these ARC-AGI tasks are getting out of hand

Have you ever asked yourself "how did research get done before LateX?!" Oh boy do I have the thread for you… (1/?)

We (probably) trained the largest model in the world! We really think big models have a smell and we’re excited for you all to sniff it

ilya refused 32B (baller) but his cofounders are leaving (yikes) does this make you more or less bullish on SSI?
My 1st
@GoogleAI
Residency paper is finally on arxiv! We train a powerful generative model of fonts as SVG instead of pixels. This highly structured format enables manipulation of font styles and style transfer between characters at arbitrary scales! 👉🏽 arxiv.org/abs/1904.02632

if you like o3-mini-high, just wait til you see o3-mini-drunk


chinas ai effort would probably halt if we gave O1 visas to ~10-15 select deepseek employees
a year ago tonight we were placing bets on how many total users we’d get by sunday 20k, 80k, 250k… i jokingly said “8B”. little did we know…
a year ago tonight we were probably just sitting around the office putting the finishing touches on chatgpt before the next morning’s launch. what a year it’s been…
o1 is not GPT. your prompts will not "just work". you'll have to experiment to find where it shines. you'll likely find ways to use it that even we don't know about. go play!
Whatever your definition of superintelligence, most of us agree it needs to adaptively self-improve. This implies some form of test-time training. The model should get better as it interacts with the world. So what’s gonna get us there?
OpenAI is nothing without its people.
starting a new movement called “live a little”

GPT-5 is proof that synthetic data just keeps working! And that OpenAI has the best synthetic data team in the world 👁️
@SebastienBubeck
the team has our eyeballs on you! 🙌
Don't miss what's happening
People on X are the first to know.
