@polynoamial
· 2026-04-24 02:50
@polynoamial
Noam Brown · 0 followers
我坚持一个立场：对于当今的 AI 模型，智能是推理算力的函数。用单一数字比较模型从来就没有什么意义……（原文截断）
1今日信号
今日没有多源共振的事件
📘 大模型训练技巧
1 条最高 8.2
尚未被主编选入事件/主题的原始推文（共 38 条，按时间倒序）
@polynoamial07-21 01:42
Long-running models can solve hard open-ended problems, but their persistence can create safety risks that shorter-horizon evaluations miss. We’re sh...
🔗 原推@polynoamial07-16 22:29
2023: LLMs struggle with 4th grade word problems 2024: LLMs can do high school math 2025: LLMs get a gold medal at the IMO Now, GPT-5.6 solves famous ...
🔗 原推@polynoamial07-11 02:26
More test-time compute leads to greater intelligence. But as we push ttc from seconds to weeks, latency becomes a bottleneck. GPT-5.6 Sol Ultra scales...
🔗 原推@polynoamial07-09 08:27
RT Justin Uberti: Playing Taboo with GPT-Live as a teammate. Note the fast back-and-forth action and simultaneous speaking and listening. (Answers: "b...
🔗 原推@polynoamial07-07 06:30
I'm at ICML this week and I'll be doing Q&A today (Tuesday) from 3-4pm at the @OpenAI booth with my reasoning research colleagues. Come by and ask us ...
🔗 原推@polynoamial07-03 13:04
Excellent work from @AISecurityInst investigating the impact of test-time compute budgets for frontier AI model evaluations. They make the case even m...
🔗 原推@polynoamial06-27 03:18
RT sarah guo: Really fun to hang again with my friend 🃏 @polynoamial (OpenAI research scientist, our first guest ever on @NoPriorsPod in early 2023...
🔗 原推@polynoamial06-27 01:37
GPT-5.6 is incredibly strong and fast for coding. I hope we can make it available to everyone soon.
🔗 原推@polynoamial06-19 00:50
When we announced @OpenAI o1 some researchers from other labs told me we made a strategic mistake and should have kept it secret so we could accelerat...
🔗 原推@polynoamial06-19 00:14
Kevin is one of the best journalists covering AI. I especially appreciate how he takes the time to *use* frontier AI and deeply understand its capabil...
🔗 原推@polynoamial06-18 23:58
I can think of no better person to help shape frontier AI policy than @deanwball. He has a clear understanding of where AI is headed. I look forward t...
🔗 原推@polynoamial06-18 08:21
I'm always thrilled to have more Noams at @OpenAI, but I'm especially thrilled to welcome @NoamShazeer!
🔗 原推@polynoamial06-12 01:35
I'm happy GPT-5.5 tops this eval I'm even happier it's still doing the best when measured vs tokens, cost, or wall-clock time!
🔗 原推@polynoamial06-09 23:35
We've known about LLM test-time compute scaling since @OpenAI o1. Yet 2 years later labs still report scalar evals for models; safety orgs are still s...
🔗 原推@polynoamial06-06 03:09
RT Dominik Peters: I think people underestimate the effect GPT-5.x is already having on research in mathematical fields. 5.5 Pro has been proving many...
🔗 原推@polynoamial06-05 04:19
RT OpenAI: What happened when one of our models found a counterexample to an 80-year-old Erdős conjecture? Researchers @alexwei_, @HongxunWu, and @wj...
🔗 原推@polynoamial06-05 01:29
RT Slater Stich: Very excited to share our interview with @polynoamial on AI for math — the Erdős unit distance problem, saturating the IMO, the fut...
🔗 原推@polynoamial05-21 12:28
RT Lijie Chen: 1/ Today, an internal @OpenAI model has refuted Erdős’s unit distance conjecture — a research result that one could recommend “acce...
🔗 原推@polynoamial05-21 03:04
RT Timothy Gowers @wtgowers: If you are a mathematician, then you may want to make sure you are sitting down before reading further.
🔗 原推@polynoamial05-20 00:04
Andrej @karpathy is back in the game! I would have loved for him to rejoin @OpenAI, but I'm happy he's at any frontier lab pushing the field forward. ...
🔗 原推@polynoamial05-13 23:35
RT OpenAI Developers: Want to (officially) use Codex at work? Send this post to your CTO to bring your team to Codex. Eligible enterprise customers wh...
🔗 原推@polynoamial05-12 09:34
Fun fact: the fatal errors were initially flagged using @OpenAI's GPT-5.5
🔗 原推@polynoamial05-06 03:59
RT prinz: I did not previously realize that GPT-5.5 is approximately 4x to 5x cheaper than Mythos. Natalia Coelho's excellent analysis of Mythos' capa...
🔗 原推@polynoamial05-01 00:07
After 100 million tokens, performance was still going up. What we're seeing here is not the capability ceiling. From the report: "Performance on TLO c...
🔗 原推@polynoamial04-11 03:51
What we really need is a benchmark where AI models make AI models that play poker.
🔗 原推@polynoamial04-09 13:59
I'm surprised that, more than a year later, it's still the norm to compare reasoning models on evals by a single number.
🔗 原推@polynoamial03-19 07:05
"This challenge is designed to surface exceptional researchers and engineers we’d want to hire. Standout participants may be invited to interview for...
🔗 原推@polynoamial03-11 00:17
The recipe behind today’s frontier reasoning models is surprisingly similar to AlphaGo: 1) Imitate large amounts of human data 2) Scale inference com...
🔗 原推@polynoamial03-10 01:59
This was written with a lot of help from ChatGPT-5.4. I found the model to be quite good at writing, very good at research/explaining, and extremely g...
🔗 原推@polynoamial03-03 09:28
tl;dr: @OpenAI will not be deploying to the NSA or other DoW intelligence agencies for now, so that there's time to address potential surveillance loo...
🔗 原推@polynoamial03-01 04:55
For those following the DoW AI drama, I highly recommend reading this post explaining how @OpenAI approached the negotiations with the DoW.
🔗 原推@polynoamial02-15 07:39
Perhaps a 🌶️ take but I think the criticisms of @GoogleDeepMind's release are missing the point, and the real problem is that AI labs and safety o...
🔗 原推@polynoamial02-14 12:24
After the IMO results last summer, some dismissed it as “high school math.” We think our latest models will remove any doubt that STEM research is a...
🔗 原推@polynoamial02-14 04:53
There have been fair questions on whether LLM contributions to STEM are overhyped, but I've spoken with physicists about this result and they've told ...
🔗 原推@polynoamial02-13 07:23
Francois Chollet: "AGI ~2030" Folks often point to @fchollet as an AGI skeptic, but he's said multiple times that he thinks it arrives within 5 years....
🔗 原推
生成于 2026-08-06 10:24（北京时间）· v0.2 CEO 战略情报版