Recent content on I've been thinking
Karpathy has been posting about AutoResearch — pointing an agent at a research question, not a task list, and letting the experimental loop run. His framing: the bottleneck in research isn’t compute or even ideas, it’s experiment throughput. If an agent can close the design-run-interpret cycle autonomously, you’re not just going faster — you’re doing structurally different…
The question I’ve been thinking about: can a LLM model, a stateless machine, teach itself? Does it have the introspection to understand its mistakes and know how to improve? I spent the last few days running an experiment based on a paper called Training-Free GRPO . The core idea: instead of fine-tuning a model with reward signals, you extract natural-language “experiences” from…
The previous experiment ended with an unresolved anomaly. Three models improved when a strong teacher (DeepSeek V3.2) injected procedural experiences into their prompts. One — Qwen 2.5 7B — regressed, and kept regressing regardless of what experiences it received or how much token budget it was given. The cross-injection experiments showed it wasn’t the content; it was something about how…
This post summarizes a paper I co-authored with Benjamin Pikus and Pratyush Ranjan Tiwari. The full paper is on arXiv . Fine-tuning a language model with GRPO is expensive. Collecting and annotating training data is expensive. So if you can only afford to train on 10% of your data, which 10% should you pick? The intuitive answer might be: a representative sample. Maybe some easy, some hard, some…
Reinforcement Learning from Human Feedback (RLHF) has become the dominant paradigm for aligning language models with human preferences. But optimizing against a learned reward model introduces a subtle failure mode: reward hacking . The Setup In RLHF, we train a reward model $r_\phi(x, y)$ on human preference data, then optimize a policy $\pi_\theta$ to maximize the expected reward:…
Mar 2026 Update: Computer Use has gotten a lot better. I myself have been very impressed with Manus’s ability to take over a browser tab and architect actions. While this has weakened the initial premise, I still think this is an interesting viewpoint on how much can change in a year. And more importantly, the original argument — that enterprises with workflows worth automating have a SOP…
About me