Siraj Raval: better plumbing beats a bigger model in his code-free VR museum build
sirajraval · x · 2026-11-06
Siraj Raval built a VR museum of all 528 of his videos without writing code, and his biggest lesson: the harness around a model matters more than the model's settings.
A lower-effort run with better plumbing scored higher and cost less. As his slides explain, the harness is the code around the model — it decides what the model sees, which tools it can use, and how many turns it gets. Same model, different harness, very different results.
His advice: fix the plumbing before you pay for more compute. Full details in his video.
More from coding & agent
- Amazon research: self-distillation pushes weight memory to 59.9%, beating retrieval — omarsar0 · 2026-11-06
- Liquid AI's decision recipe matches Jev; new release adds on-device, modalities, primitives — helloiamleonie · 2026-11-06
- Decision models: general-purpose classifier or LLM? LoRA fine-tune wins — helloiamleonie · 2026-11-06
- SFT classification datasets into a decision model; LoRA beats full fine-tune — helloiamleonie · 2026-11-06
- LLMs are trained to predict next tokens, not make decisions — helloiamleonie · 2026-11-06
- Turn any LLM into a decision model with ~15 lines of inference-only code — helloiamleonie · 2026-11-06