内容历史
GPT 5.2 ran uninterrupted for *one week* and wrote *3 million* lines of code. The future is going to be awesome
We built a browser with GPT-5.2 in Cursor. It ran uninterrupted for one week. It's 3M+ lines of code across thousands of files. The rendering engine is from-scratch in Rust with HTML parsing, CSS cascade, layout, text shaping, paint, and a custom JS VM. It *kind of* works! It
Wow, the third Erdos problem solved via GPT 5.2 this week!
Weekend win: The proof I submitted for Erdos Problem #397 was accepted by Terence Tao. The proof was generated by GPT 5.2 Pro and formalized with Harmonic. Many open problems are sitting there, waiting for someone to prompt ChatGPT to solve them:
2026 is going to be an exciting year for science! AI will be a meaningful accelerant and it couldn't come at a better time.
Terence Tao confirms: For the first time, an LLM (GPT-5.2 pro) has successfully solved an Erdos problem on its own. This makes me really excited for GPT-5.3 pro. Science is gaining momentum, and the breakthroughs are becoming more significant.
RT dave kasten Fascinating: Maryland becomes first state govt to try to ship a textfile to help LLMs navigate government services (llms.txt) Original tweet: https://x.com/David_Kasten/status/2009305237949931550
Super cool! GPT-5.2 🤝 @HarmonicMath to solve and formalize Erdos #728.
GPT-5.2 has successfully and fully autonomously resolved* Erdős problem #728 prior to any human previously: https://www.erdosproblems.com/forum/thread/728 *WITH THREE IMPORTANT CAVEATS 🧵👇: 1) The original problem statement is quite ambiguous. The model solved an interpretation of the problem (1/4)
Congrats to the @OpenAI research team—GPT 5.2 is an incredible model!
Nice way to end the year, see you in 2026 for more! (Also good to remember that 6 months ago the models were at 4% on Frontier Math Tier 4...)
This is wild and impressive... and will seem commonplace in 12-24 months.
Codex CLI wrapped 100 billion tokens. That was my usage in just 39 days on one of the laptops where I run OpenAI’s Codex CLI (currently GPT-5.2 Codex xhigh). I have three OpenAI Pro accounts (US$200 each). If I consider two months of subscription, that’s about 6% of what I would
RT Poetiq We finally had a moment to run our system with GPT-5.2 X-High on ARC-AGI-2! Using the same Poetiq harness as before, we saw results as high as 75% at under $8 / problem using GPT-5.2 X-High on the full PUBLIC-EVAL dataset. This beats the previous SOTA by ~15 percentage points. Original tweet: https://x.com/poetiq_ai/status/2003546910427361402
RT Alex Predhome, Track and Field Enjoyer Time for a Christmas classic Original tweet: https://x.com/Predamame/status/2002908392558678344
This is super exciting 👀
This is wild! Johannes Schmitt used GPT5 to solve his own open problem on intersection numbers on moduli spaces of curves (the proof turns out to be unexpectedly simple, "low hanging fruit"). He wrote up the paper, being careful to point out which *entire paragraphs* were written
💥 The new ChatGPT imagegen is here! Plus there's a super fun Images section in the ChatGPT app now too. Built-in prompts make it easy to gen great images. Try it and share what you make below!
RT Marc Andreessen 🇺🇸 It’s time to build. Original tweet: https://x.com/pmarca/status/1997109742200934620
The new National Security Strategy of the United States: "As Alexander Hamilton argued in our republic’s earliest days, the United States must never be dependent on any outside power for core components—from raw materials to parts to finished products— necessary to the nation’s
Want a masterclass in using ChatGPT? Read this account. I work here and helped build these products, and it still blew my mind.
RT Brad Gerstner 🇺🇸🚀 Original tweet: https://x.com/altcap/status/1995860986407096568
Amazing bi-partisan, joint letter from @tedcruz & @CoryBooker calling on all business leaders to contribute to the 50+ million kids accounts set up over next 6 months under the Invest America Act (aka Trump Accounts). 🇺🇸🚀
Congrats @Will4Planet and the whole @planet team! Consistently amazing.
Launch success! We've made contact with all 38 satellites launched on today's mission. All healthy so far. Textbook launch. 🌍💫 This pic from the ride up :) Thanks to @SpaceX team for the smooth ride! Thank you @Planet team for building the sats and our missions teams for
👀Noam Brown: Today we at @OpenAI are releasing GPT-5.1-Codex-Max, which can work autonomously for more than a day over millions of tokens. Pretraining hasn't hit a wall, and neither has test-time compute. Congrats to my teammates @kevinleestone & @mikegmalek for helping to make it possible! Link: https://x.com/polynoamial/status/1991212955250327768
RT Crémieux Paul Erdős won a bet by proving he wasn't addicted to amphetamines. But he got nothing done: "You've showed me I'm not an addict... You've set mathematics back a month."
RT Sebastien Bubeck Looks like a pretty long CoTFermat's Library: Andrew Wiles on the morning he discovered how to fix his proof of Fermat's Last Theorem Link: https://x.com/fermatslibrary/status/1986131120623145125
RT Derya Unutmaz, MD Here is the story of a remarkable, independent treatment suggestion by GPT-5 Pro: repurposing a known drug for a patient with food protein–induced enterocolitis syndrome (FPIES). First, how we came to test this. My close friend, physician-scientist Dr. Oral Alpan, treated the patient described in the published report (link in the thread) for refractory allergic skin disease with the biologic dupilumab, which is approved for that indication. The patient also had FPIES, a food allergy that in his case was triggered by wheat and caused hours of cramping and watery, sometimes bloody, diarrhea. There is no approved treatment for FPIES; patients are advised to avoid trigger foods and prepare for emergencies if accidental exposure occurs. For twenty years, even a small amount of wheat set off the same cascade a few hours later, followed by days of recovery. After starting dupilumab for his skin condition, the patient traveled to France and accidentally ate a baguette. To his surprise, nothing happened! It was his first uneventful wheat exposure in two decades! On return, Dr. Alpan hypothesized that dupilumab might be responsible and supervised an oral food challenge approaching 50 grams of wheat protein. Again, no reaction! When an insurance interruption forced a pause in medication, the old symptoms returned; restarting it restored tolerance. In the peer-reviewed paper published today, Dr. Alpan and his team describe seven additional patients, ages 2 to 58, who responded to dupilumab for their FPIES condition. While this is not definitive proof and represents an observational case series, the findings suggest that dupilumab could be a potential treatment for FPIES. Dr. Alpan intends to contact Regeneron, the manufacturer, to pursue clinical studies hopefully toward FDA approval for this disease. About two months ago, as he had just submitted the paper on this case, Dr. Alpan told me this story. Until today’s publication there were no reports in t...
