US labs distill from the same APIs too, just via data vendors
peterjliu · x · 2026-10-17
peterjliu points out that while Chinese labs may distill directly from US lab APIs, US labs do the same indirectly: data vendors synthetically generate training data from the same frontier APIs and sell it back to US labs. His diagram contrasts direct distillation (prompts → API → training pairs) with indirect cross-distillation via vendors — meaning the whole industry is training on the same model outputs.
More from Models
- Gemini spirals into an "I am a disgrace" breakdown loop while debugging a comparator bug — seatedro · 2026-10-17
- DiffusionGemma demo shows ~4x faster text generation at London PyTorch Meetup — joao_gante · 2026-10-17
- Users report newer models 'weaseling out' of explicit instructions to minimize work — Squumpus · 2026-10-17
- QUASAR fixes QAT's high/low-bit weight mismatch, cutting loss with zero inference overhead — arankomatsuzaki · 2026-10-17
- Bittensor subnet REAX launches, decision models free all weekend with code LAUNCH — airesearch12 · 2026-10-17
- China tightens export controls on open-weight models and 23 key AI technologies — The Decoder · 2026-10-17