Nathan Rowe @_nathanrowe

📣📣 Meet Qwen-AgentWorld — a native language world model that simulates 7 agent environments (MCP, Search, Terminal, SWE, Web, OS, Android) within a single model. Environment modeling is the training objective from day one, not a post-hoc adaptation. 🤔 LLMs are trained to be better agents — better at acting in environments. But nobody has trained them to model the environments themselves. 🗺️ Our roadmap: investigate how language world modeling can push the boundaries of general agent capabilities, along two routes: 1️⃣ Build a foundation model for environment simulation — outperforming Claude Opus 4.8 and GPT-5.4 on AgentWorldBench 2️⃣ Investigate how world modeling enhances agent training: 🔬 Controllable Sim RL (agentic RL with LWM as environments) surpasses training in real environments 🧠 Learning to predict environments (LWM warm-up) makes agents stronger — remarkably, even without any agent-specific training, this predictive knowledge transfers to agentic tasks with zero fine-tuning 📑 Paper: arxiv.org/abs/2606.24597 📖 Blog: qwen.ai/blog?id=qwen-a… 💻 GitHub: github.com/QwenLM/Qwen-Ag… 🤗 HuggingFace: huggingface.co/collections/Qw… 🧩 ModelScope: modelscope.cn/collections/Qw…

We've kept hearing how GLM-5.2 beats Opus 4.8, and are skeptical of benchmarks - so we tested them on a real bug from the Cline repo. While both models fixed the issue, GLM was the winner in terms of cost and code quality: - GLM used twice as many tokens (GLM 1.1m vs Opus 660K) but cost half as much (GLM $0.41 vs Opus $0.81) - Opus finished quicker - 1.6 min and 12 tool calls vs GLM 4.7 min and 28 tool calls - GLM cleaned up dead code and verified the build compiled before completing. Opus didn't - it left type errors that passed tests but broke the production build. Both runs used the same Cline harness prompting and tools, so it seems GLM is RL trained to spend more tokens verifying its work before completing. Impressive work by the @Zai_org team!

We want to help all companies be secure, working with the USG and the security ecosystem. *The full version of GPT-5.5-Cyber is here; state of the art performance on CyberGym. *Patch The Planet and Codex Security will help solve security problems instead of just finding them.

Introducing Claude Tag, a new way for teams to work with Claude. In Slack, Claude joins as a team member with access to the channels and tools you choose. Tag Claude in and delegate tasks to it while you focus on other work.

What stands out isn’t correctness, but pattern discovery. This is closer to hypothesis generation than automation. @OpenAI

OpenAI @OpenAI

8 months ago

GPT-5.2 derived a new result in theoretical physics. We’re releasing the result in a preprint with researchers from @the_IAS, @VanderbiltU, @Cambridge_Uni, and @Harvard. It shows that a gluon interaction many physicists expected would not occur can arise under specific

919 1K 9K 4.5M 3K

Important reminder: Seeing motion ≠ understanding physics. VLMs are still pattern-matching far more than modeling the world.

Tiange Xiang @xxtiange

10 months ago

‼️VLMs/MLLMs do NOT yet understand the physical world from videos‼️ In our recent work, we found that even the most advanced AI models still lag behind humans in one key aspect: reasoning about the kinematic properties of objects from videos. Takeaways: 1. ChatGPT 5.1 leads

keyboard_arrow_left Previous keyboard_arrow_right Next

11 9 81 39K 23

This is the shift people keep missing. LLMs aren’t just tools or personalities — they’re population simulators. Once you stop forcing them into a single voice, entirely new classes of work open up.

Andrej Karpathy @karpathy

8 months ago

Congrats on the launch @simile_ai ! (and I am excited to be involved as a small angel.) Simile is working on a really interesting, imo under-explored dimension of LLMs. Usually, the LLMs you talk to have a single, specific, crafted personality. But in principle, the native,

380 562 8K 989K 4K

@karpathy The core of GPT fits in a few hundred lines. Everything else is scale, optimization, reliability, and distribution. That’s where almost all the real difficulty lives.

The core of GPT fits in a few hundred lines. Everything else is scale, optimization, reliability, and distribution. That’s where almost all the real difficulty lives.

Andrej Karpathy @karpathy

8 months ago

New art project. Train and inference GPT in 243 lines of pure, dependency-free Python. This is the *full* algorithmic content of what is needed. Everything else is just for efficiency. I cannot simplify this any further. gist.github.com/karpathy/8627f…

637 3K 25K 5.2M 29K

wsj. com/tech/ai/picks-and-shovels-still-rule-the-ai-tech-trade-0bc1ddf1

AI isn’t one big trade anymore — it’s a structure play. Capital is flowing into memory, chips, and infra, while traditional software gets repriced. The signal is clear: inputs beat layers that can be internalized.

GPT-2 going from “too dangerous to release” to “new MNIST” is a reminder that capability isn’t static — it’s a function of cost curves and tooling maturity.

Andrej Karpathy @karpathy

9 months ago

Enabled fp8 training for +4.3% improvement to "time to GPT-2", down to 2.91 hours now. Also worth noting that if you use 8XH100 spot instance prices, this GPT-2 repro really only costs ~$20. So this is exciting - GPT-2 (7 years ago): too dangerous to release. GPT-2 (today): new

218 294 4K 685K 2K

@lexfridman The fun comes from removing friction. When syntax and debugging fade, intent becomes the bottleneck again.

This isn’t a model upgrade. It’s research being turned into a repeatable workflow, not a one-off interaction.

OpenAI @OpenAI

9 months ago

Deep research in ChatGPT is now powered by GPT-5.2. Rolling out starting today with more improvements.

605 544 5K 1.7M 908

This paper is kinda crazy but there's a big caveat imo. "We are able to train the 8B parameter size of Qwen2.5 to 91% accuracy on GSM8K with only 13 trained parameters in bf16 (26 total bytes)." --> wild!! "in this low-parameter setting LLAMA is less responsive than Qwen and reaches 85% with an update size of 1KB (500 parameters trained in bf16)." "Unlike Qwen, when we train fewer than five parameters, LLAMA barely improves performance above baseline." So once again, Qwen is likely a bit of an outlier in its behavior, and unclear how well other regular LLMs can learn reasoning with limited params. Would love to see more thorough experiments on this.

This isn’t about better agents. It’s about enterprises standardizing how work itself is delegated. Frontier looks less like a product, more like an operating layer for labor.

OpenAI @OpenAI

9 months ago

Introducing OpenAI Frontier—a new platform that helps enterprises build, deploy, and manage AI coworkers that can do real work. openai.com/index/introduc…

548 755 6K 2.1M 4K

@OpenAIDevs This is Apple making agentic coding a default workflow, not a side experiment. When agents ship inside Xcode, “AI-assisted coding” quietly becomes just “coding.”

This is Apple making agentic coding a default workflow, not a side experiment. When agents ship inside Xcode, “AI-assisted coding” quietly becomes just “coding.”

OpenAI Developers @OpenAIDevs

9 months ago

GPT-5.2 and GPT-5.2-Codex are now 40% faster. We have optimized our inference stack for all API customers. Same model. Same weights. Lower latency.

533 706 11K 2.2M 1K

@karpathy This is the real scaling story. Frontier models grab headlines, but capability is quietly becoming cheap, reproducible, and local. When GPT-2 is <$100, the floor for experimentation disappears.

This is the real scaling story. Frontier models grab headlines, but capability is quietly becoming cheap, reproducible, and local. When GPT-2 is <$100, the floor for experimentation disappears.

Andrej Karpathy @karpathy

9 months ago

nanochat can now train GPT-2 grade LLM for <<$100 (~$73, 3 hours on a single 8XH100 node). GPT-2 is just my favorite LLM because it's the first time the LLM stack comes together in a recognizably modern form. So it has become a bit of a weird & lasting obsession of mine to train

323 601 7K 1.3M 3K