Stanford CS329A Self-Improving AI Agents | Part 2 | Test-Time Compute Scaling

Explains how repeated sampling plus a verifier can improve results without changing model weights, with the tradeoffs. Useful when deciding how to spend inference budget in agents.

• test-time compute — Spending more computation when the model answers — thinking longer, trying multiple attempts — to buy accuracy without training a bigger model.
• inference — Running a trained model to get answers — the phase where AI is actually used, as opposed to trained.
• fine-tuning — Taking a trained model and training it a bit more on your own examples so it gets better at one specific job.

• videoStanford CS153 Frontier Systems | The AI Native Company: How One Founder Becomes a 1000x Engineer
• videoStanford CS25: V5 I Large Language Model Reasoning, Denny Zhou of Google Deepmind
• videoStanford CS25: Transformers United V6 I From Language Models to Native Multimodal Intelligence
• videoStanford CME295 Transformers & LLMs | Autumn 2026 | Lecture 1 - Transformers

• articleScaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model ParametersCharlie Snell et al.
• postScaling Verification Pushes Agents to New State of the ArtStanford AI Lab
• articleMaking LLMs faster without sacrificing accuracywww.amazon.science

Checking sign-in…

Loading comments…