Image created with gemini-3.1-flash-image-preview with claude-sonnet-4-5. Image prompt: Animation cel style: A muscular blue-skinned genie with purple vest emerges from a brass oil lamp, one arm extended toward a floating open textbook with glowing pages. Magical teal wisps carry animated letters, mathematical symbols, and small diagrams spiraling upward from the book. Clean gradient background in warm amber tones, Disney-quality hand-drawn aesthetic with bold outlines, cinematic lighting, jewel-tone colors, horizontal composition with clear space at top for title text.
(24) ⚡️ Prism: OpenAI’s LaTeX “”Cursor for Scientists”” — Kevin Weil & Victor Powell, OpenAI for Science – YouTube https://www.youtube.com/watch?v=W2cBTVr8nxU
💥 Today we’re introducing Prism–a free, AI-native workspace for scientists to write and collaborate on research, powered by GPT-5.2. Accelerating science requires progress on two fronts: 1. Frontier AI models that use scientific tools and can tackle the hardest problems 2.”” https://x.com/kevinweil/status/2016210486778642808
Introducing Prism | OpenAI https://openai.com/index/introducing-prism/
A systematic attempt to understand how good GPT-5.2 Pro is at Erdos problems, the hard, generally unsolved math problems much discussed here. The answer is pretty good! But also it is wrong about a third of the time and you wouldn’t know that if you couldn’t check the results.”” https://x.com/emollick/status/2015949184734568842
Weak-to-strong generalization | OpenAI https://openai.com/index/weak-to-strong-generalization/
OpenAI Prism is *NOT* only for researchers. Uploaded Google’s “”Nested Learning”” paper for Continual Learning and had it generate a diagram that visualizes the paradigm in a simple way. Great if you’re a student or curious person who wants to learn too. Plus it’s free.”” https://x.com/daniel_mac8/status/2016554325691015604
Important new course: Agent Skills with Anthropic, built with @AnthropicAI and taught by @eschoppik! Skills are constructed as folders of instructions that equip agents with on-demand knowledge and workflows. This short course teaches you how to create them following best”” https://x.com/AndrewYNg/status/2016564878098780245
Can AI solve math research problems that have eluded human mathematicians? Our new benchmark, FrontierMath: Open Problems, is designed to help find out. AI hasn’t solved any of these yet, but the game is young!”” https://x.com/EpochAIResearch/status/2016188014540816879
People should read stuff beyond Tolkien. There are even better old-school fantasy/science fiction vocabularies out there. Wolfe’s The Book of the New Sun has enough names for a century of startups. Or here is a 600 page list of Jack Vance names:”” https://x.com/emollick/status/2015643546108248339
GPT-5.2 for language learning:”” https://x.com/gdb/status/2014373710392918082
GPT-5.2 Pro pointed out a flaw in one of the Tier 4 math problems:”” https://x.com/gdb/status/2014859263701839963
gpt-5.2 pro for mathematics:”” https://x.com/gdb/status/2014803896540267001
It is a bit weird that this is not a bigger deal in academic publishing circles. I suspect it is, in part, because GPT-5.2 Pro is not easily available due to its high price & fact that it is not part of ChatGPT Edu. Very few have used it. (There are other reasons too, of course)”” https://x.com/emollick/status/2014499691992424560
Interesting qualitative observations on GPT-5.2 Pro’s high frontier math score from one of the folks running the test.”” https://x.com/emollick/status/2015069180177809817
OpenAI Prism for scientific research: – a free, unified LaTeX workspace with unlimited collaborators – AI-assisted proofreading, citations, and literature search – with project-data aware workflows basically Overleaf with AI”” https://x.com/scaling01/status/2016211218633990391
Introducing Prism, a free workspace for scientists to write and collaborate on research, powered by GPT-5.2. Available today to anyone with a ChatGPT personal account: https://x.com/OpenAI/status/2016209462621831448
GPT-5.2 for reproducibility of academic papers:”” https://x.com/gdb/status/2014857297957421173
RL coding agents increasingly game rewards by exploiting their semantic and syntactic weaknesses. Can LLMs detect such behaviors from live training rollouts? We find contrastive cluster analysis is key! 🚀 GPT-5.2 jumps from 45% to 63%. Humans reach 90% Paper + data 🧵”” https://x.com/getdarshan/status/2017054360887611510
New record on FrontierMath Tier 4! GPT-5.2 Pro scored 31%, a substantial jump over the previous high score of 19%. Read on for details, including comments from mathematicians.”” https://x.com/EpochAIResearch/status/2014769359747744200?s=20
Leave a Reply