Are LLMs just statistical search engines? A rebuttal on why search calls aren't fundamental
gandamu_ml · x · 2026-09-01
A debate on the nature of LLMs: one side argues LLMs are statistical search engines that merely use a different indexing algorithm, hence they can only copy, not create.
The rebuttal (gandamuml) counters that this is like saying "cooking is just particle physics" — a generalization so loose it loses meaning. In his view, there is no clean code/data distinction inside an LLM where data exists purely to be retrieved; consumer-facing products can call external search engines and do caching/RAG, but that's an engineering add-on (humans can look things up too), not fundamental to what an LLM is.
Related event: Debate: Are LLMs Just Statistical Search Engines?(3 posts)→
More from Models
- NVIDIA reportedly buys Hugging Face for $12.9B; GLM-5.3-Flash and Qwen4 preview land — altryne · 2026-09-01
- METR + Redwood GPT agent swarm probe could only analyze GPT variants, not Claude — geoffreyirving · 2026-09-01
- A nonprofit wants Qwen for everyday paperwork: how to pick local hardware — DerAndi_DE · 2026-09-01
- Thomson Reuters builds $40M legal LLM rivaling Claude Opus on Qwen — josh_wills · 2026-09-01
- Teknium: Hermes has long supported NVIDIA's openshell — corporate IT teams that missed it are uninformed — Teknium · 2026-09-01
- Users report Claude Projects favors older uploaded files as docs pile up — Electronic_Spell_154 · 2026-09-01