Every post of theirs in the library
Loading their posts
Nathan Lambert
Writes about AI and Hiring. 88% of posts open with an observation. A typical post reaches 472 likes.
- Posts
- 34
- Likes
- 27K
- Typical
- 472
- Best
- 4.4K
What they write about
How they open
Every post of theirs in the library
| Post | Topic | Likes | Replies | Reposts | Score | Posted | Actions |
|---|---|---|---|---|---|---|---|
| @natolambertAirbnb CEO Brian Chesky: “We’re relying a lot on Alibaba’s Qwen model. It’s very good. It’… | AI | 4.4K | 119 | — | 58 | ||
| @natolambertKimi K3 license. It's inspired by MIT but distinctly non-commercial, where any company mak… | Revenue | 3.8K | 203 | — | 57 | ||
| @natolambertMajor restructuring at Gemini (Dean out, Hassabis no longer CEO). This story will be studi… | AI | 2.5K | 94 | — | 57 | ||
| @natolambertMany AI leaders in the US accused Chinese LLMs of subtle manipulation of the user (without… | AI | 1.3K | 40 | — | 47 | ||
| @natolambertThe gaping hole that Qwen imploding would leave in the open research ecosystem will be har… | AI | 1.1K | 42 | — | 45 | ||
| @natolambertThinky with a ~1T param, 41B active, apache-2 model Benchmarks are a clear step up from Ne… | AI | 983 | 29 | — | 45 | ||
| @natolambertA good time to remind people that in my time doing LLM research I feel like a minority of … | AI | 961 | 40 | — | 45 | ||
| @natolambert> be me > be zuck > need llama 4 to land > send a model/prompt to LMSYS to get a top1 scor… | AI | 794 | 18 | — | 39 | ||
| @natolambertI'd put good money on this being an high-impact finetune of one of the large, Chinese MoE … | AI | 784 | 35 | — | 43 | ||
| @natolambertIf anyone at @AlibabaGroup needs a report on Qwen's impact in order to make changes to kee… | AI | 760 | 29 | — | 43 | ||
| @natolambertQwen’s biggest models have never been open weight recently. Something is changing. You can… | AI | 712 | 39 | — | 42 | ||
| @natolambertVery happy to show that we can do RL finetuning on 405B models with open-source code, beat… | AI | 658 | 26 | — | 38 | ||
| @natolambertOpenAI skips o2, previews o3 scores, and they're truly crazy. Huge progress on the few ben… | AI | 621 | 14 | — | 37 | ||
| @natolambertI read the Anthropic open-weight model piece and despite there being nothing new in it, it… | AI | 596 | 47 | — | 41 | ||
| @natolambertThoughts on Kimi K2 Thinking Congrats to the Moonshot AI team on the awesome open release.… | AI | 512 | 14 | — | 39 | ||
| @natolambertStill hard to expect the unexpected with AI. It goes to show how skilled many of the scien… | AI | 491 | 12 | — | 39 | ||
| @natolambertThe best part of all these Claude 5 Fable safety measures is I bet the jailbreaking commun… | AI | 472 | 15 | — | 38 | ||
| @natolambertThis looks like a model that's competitive with GPT OSS 120B or similar Qwen3.5 models on … | AI | 471 | 9 | — | 38 | ||
| @natolamberto3’s weird hallucinations could indicate they used llm as a judge (or other softer verifie… | AI | 415 | 11 | — | 34 | ||
| @natolambertSeems like Llama 4’s reputation is maybe irreparably tarnished by having a separate unrele… | AI | 394 | 16 | — | 34 | ||
| @natolambertThis is confusing and stands to fracture the AI community unless they are being misleading… | AI | 377 | 44 | — | 38 | ||
| @natolambertThe only reasonable expectation if you're a fan of open weight models is that if there's a… | AI | 373 | 38 | — | 36 | ||
| @natolambertI'm simultaneously very happy for Dean, but also sad to have another extremely bright, ind… | AI | 347 | 5 | — | 35 | ||
| @natolambertThere's a lot of sloppy thinking around open models. You can ban them and make it impossib… | AI | 344 | 36 | — | 37 | ||
| @natolambert“Due to GPT‑5.3-Codex being so different from its predecessors, the data from alpha testin… | AI | 344 | 14 | — | 36 | ||
| @natolambertBoth OpenAI Pro sub and Gemini Ultra sub right now are pretty bad deals for most people at… | AI | 340 | 37 | — | 34 | ||
| @natolambertLlama 4 was a messy release: unreleased finetunes boosting scores, rumors of training on t… | AI | 301 | 20 | — | 32 | ||
| @natolambertThe current pace of token-efficient reasoning improvements across minor Claude Opus/GPT mo… | AI | 289 | 12 | — | 34 | ||
| @natolambertOpen weight base models + incredible ML teams in a specific niche can create immense value… | AI | 280 | 18 | — | 35 | ||
| @natolambertGoogle's models today are absolutely crushing it today at IO. I'm still not optimistic tha… | AI | 219 | 9 | — | 29 | ||
| @natolambertInkling was somewhat inevitable once Tinker took off -- integration of post-training servi… | AI | 218 | 8 | — | 32 | ||
| @natolambertNvidia is the tip of the spear for the growing US open model efforts. Is a long time comin… | AI | 212 | 13 | — | 32 | ||
| @natolambertI expect a lot more remarkable plots like this showing how fast the frontier models are pr… | AI | 172 | 7 | — | 30 | ||
| @natolambertThreading the needle in this post of anthropic has done some bad things for AI governance … | AI | 150 | 14 | — | 23 |