ilge

Researcher @OpenAI | there is as yet insufficient data for a meaningful answer.

[OpenAI @OpenAI]

Today, we share a breakthrough on the planar unit distance problem, a famous open question first posed by Paul Erdős in 1946.

For nearly 80 years, mathematicians believed the best possible solutions looked roughly like square grids.

An OpenAI model has now disproved that belief, discovering an entirely new family of constructions that performs better.

This marks the first time AI has autonomously solved a prominent open problem central to a field of mathematics.

[ilge @ilge]

The greatest game of all time.

[exQUIZitely 🕹️ @exQUIZitely]

Today, Descent II celebrates its 30th birthday.

Published on March 13, 1996, it is considered by fans to be the best in the franchise. One of its biggest strengths was the complete freedom of movement: no limitations, total six degrees of freedom (which set it apart from other FPS games at the time). But it also came at the cost of sometimes feeling lost or even nauseous.

The guide bot helped reduce disorientation, but you still needed a strong stomach for the rollercoaster sensation of flying through tight corners and labyrinth-like mine layouts.

Happy 30th, and thanks for the memories!

[OpenAI @OpenAI]

You can just build things.

[Sheryl Hsu @SherylHsu02]

2/n We officially competed in the online AI track of the IOI, where we scored higher than all but 5 (of 330) human participants and placed first among AI participants. We had the same 5 hour time limit and 50 submission limit as human participants. Like the human contestants, our system competed *without* internet or RAG, and just access to a basic terminal tool.

[Aleksander Holynski @holynski_]

Another one. Already a powerful painting, but moving around it yourself gives a totally different feeling.

Jacques Louis David's "The Death of Socrates" => #Genie3

[ilge @ilge]

It’s a good day!

[OpenAI @OpenAI]

We released two open-weight reasoning models—gpt-oss-120b and gpt-oss-20b—under an Apache 2.0 license.

Developed with open-source community feedback, these models deliver meaningful advancements in both reasoning capabilities & safety.

https://t.co/PdKHqDqCPf

[Alexander Wei @alexwei_]

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).

[Sheryl Hsu @SherylHsu02]

Watching the model solve these IMO problems and achieve gold-level performance was magical. A few thoughts 🧵

[ilge @ilge]

Research roadmap essentials

[ilge @ilge]

Congrats, @isafulf , @EdwardSun0909 and the whole team behind this launch! 🚀
Truly a new era in the space of agentic products.

[OpenAI @OpenAI]

Today we are launching our next agent capable of doing work for you independently—deep research.

Give it a prompt and ChatGPT will find, analyze & synthesize hundreds of online sources to create a comprehensive report in tens of minutes vs what would take a human many hours.

[OpenAI @OpenAI]

OpenAI o3-mini is now available in ChatGPT and the API.

Pro users will have unlimited access to o3-mini and Plus & Team users will have triple the rate limits (vs o1-mini).

Free users can try o3-mini in ChatGPT by selecting the Reason button under the message composer.

[Nat McAleese @__nmca__]

Epoch AI are going to publish more details, but on the OpenAI side for those interested: we did not use FrontierMath data to guide the development of o1 or o3, at all. (1/n)

[Chelsea Sierra Voss @csvoss]

don’t miss this part of today’s 12th Day of OpenAI: “Deliberative Alignment,” exciting work by the illustrious @MelodyGuan et al!

the technique achieves a Pareto improvement over previous approaches such as RLHF, and reduces overrefusals!

https://t.co/la6zthJaQP

[François Chollet @fchollet]

Today OpenAI announced o3, its next-gen reasoning model. We've worked with OpenAI to test it on ARC-AGI, and we believe it represents a significant breakthrough in getting AI to adapt to novel tasks.

It scores 75.7% on the semi-private eval in low-compute mode (for $20 per task in compute ) and 87.5% in high-compute mode (thousands of $ per task). It's very expensive, but it's not just brute -- these capabilities are new territory and they demand serious scientific attention.

[ilge @ilge]

Found him

[ilge @ilge]

I’m excited to attend NeurIPS’24!

[ilge @ilge]

o1 is finally here :)

[Kevin Weil 🇺🇸 @kevinweil]

Day 1 of the 12 days of OpenAI! 🎁

We're launching o1, which for the first time combines multimodality with the new reasoning paradigm, in addition to being smarter and faster.

We're also launching the pro tier, a $200/mo plan for unlimited access to our models (including voice) as well as o1 pro mode which uses even more compute to answer the hardest problems in math, science, coding and more.

Two launches on day 1! See you again tomorrow 🔥

[ilge @ilge]

I followed the trend of asking ChatGPT: “out of all the data you have on me, generate an image that captures me the way you see me”. 💓

[OpenAI Developers @OpenAIDevs]

Introducing canvas—your coding surface in ChatGPT.

✏️ Edit code inline
🐛 Review code and fix bugs
💬 Add logs and comments
🚢 Port to different languages

We’ll be adding more to canvas over time. ChatGPT Plus and Team users can try the beta starting today.

[hunter @hunterlightman]

come for my jokes, stay because @ilge and @polynoamial are explaining the next frontier of ai research!