X post alleges Moonshot distilled Anthropic’s Fable for K3 and used GB300 servers
Dr_Singularity · x · 2026-07-23
A post circulating on X alleges that Moonshot AI distilled Anthropic’s Fable to help develop K3, and that the company used an internal platform to switch access methods and avoid detection during large-scale distillation against U.S. models.
The same quote also claims Moonshot acquired GB300-equipped servers and accessed GB300s in Thailand, suggesting the company used overseas infrastructure to train its models more quickly.
More from Infra
- AWS on agent prompt caching: explicit + implicit patterns to slash token costs — AWS ML Blog · 2026-10-27
- Thieves spoofing GPS steal $100M+ in data center cargo since June — Ars Technica AI · 2026-10-27
- Lam Research: weekly token usage up ~250x in 18 months — Beth_Kindig · 2026-10-27
- AWS shows how to build enterprise agentic retrieval on Bedrock with CDK — AWS ML Blog · 2026-10-27
- Encrypted GPT-2 speedrun challenges FHE inference: best is 3.031 s/token, goal sub-1s — tensorqt · 2026-10-27
- Open-source Latch compiles PyTorch into training megakernels, 2.75x faster on H100 — sguada · 2026-10-27