Anthropic still leads everyone in pretraining, especially on data, researcher argues
Guodzh · x · 2026-11-23
Guodzh argues that in 2024-2025 OpenAI didn't believe much in pretraining and big models, while Anthropic remains ahead of everyone in pretraining, especially on data. His reasoning rests on the "prediction ≈ compression ≈ intelligence" thesis — an opinionated take on the frontier lab landscape, not confirmed by either company.
More from Companies & People
- Allie Miller shares the highest-leverage AI and non-AI spending on her small team — alliekmiller · 2026-11-23
- Sequoia's Doug Leone: 30% of VCs on boards are incompetent, 40% are no-ops — aldeamartinez · 2026-11-23
- Replit ships Windows-hardened security model for desktop AI, wins Christie's — amasad · 2026-11-23
- Insider claim: OpenAI never believed much in pretraining and big models through 2024-2025 — Guodzh · 2026-11-23
- Ortet Benchmarks 34 Open-Weight Health Models on 50 Tasks Across 15 Benchmarks — kchonyc · 2026-11-23
- CMU Robotics Institute Opens Annual Faculty Hiring, Deadline December 16, 2026 — CSProfKGD · 2026-11-23