Jerry Tworek
@MillionInt
Berry farmer @ OpenAI | o3, o1, GPT4, ChatGPT, Codex, Solved Rubik’s cube with robotic hand | cautious AI optimist
We trained a model and it is good in some things
We're releasing a preview of OpenAI o1—a new series of AI models designed to spend more time thinking before they respond. These models can reason through complex tasks and solve harder problems than previous models in science, coding, and math. openai.com/index/introduc…
very very true
There is something special – far more rewarding than money – about working with an epic team to make breakthroughs
This guy gets it
i’m much more inclined to say that the RL *system* inside OpenAI is AGI rather than than any fixed model checkpoint which comes out of it
It is
It’s truly a privilege to be able to wake up every morning, see where the latest intelligence frontier is, and help push it a little further.
You start with sparks and get a fire
Well past the sparks of AGI paper now ..
To summarize this week: - we released general purpose computer using agent - got beaten by a single human in atcoder heuristics competition - solved 5/6 new IMO problems with natural language proofs All of those are based on the same single reinforcement learning system
Sheryl (@sherylhsu02) was our first hire onto the multi-agent team. Within a few months of joining, she helped to make this possible. We're so lucky to have her on the team!
Watching the model solve these IMO problems and achieve gold-level performance was magical. A few thoughts 🧵
Why am I excited about IMO results we just published: - we did very little IMO-specific work, we just keep training general models - all natural language proofs - no evaluation harness We needed a new research breakthrough and @alexwei_ and team delivered
I need to print out some copies
I feel this may be helpful to some of you today:
Many problems have been solved, many problems still remain unsolved
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
I love how ChatGPT keeps automating all the easy, mindless, hassle-free tasks, so I’m free to spend my valuable time on the tedious, frustrating, soul crushing ones!
I feel connected to @FakePsyho through intertwining threads of fate
we're competing in the @atcoder World Finals programming contest. real nailbiter — OpenAI has been #1 for most of the contest. looked like it might be over when @FakePsyho pulled ahead, but we've just retaken the lead. 1 hour and 20 minutes to go!
That could be true
The OpenAI open source model is going to be really, really good🍓
Road is full of forks. Every day something earthshatteringly worldchanging happens and no one notices because we don’t know the counterfactuals
Automating strawberry production
European company @Dyson is now automating Strawberry production. Showcasing that Europe is starting to lead technologically 🇪🇺
That’s about right
Right now my AI usage is something like 66% o3-pro, 33% o3, 1% Veo 3, 0% everything else
Good model
On my YouTube stream tonight, @OpenAI's o3 and o3-Pro correctly solved not 1, not 2, not 3, not 4, not 5, not 6, but 7!!! graduate-level physics problems in quantum mechanics within a single reply *Technically, it was more like 6.5/7 but still 😮