Lawsuit filing reveals Anthropic's early internal codenames for Claude
voooooogel · x · 2026-10-15
A lawsuit filing has surfaced Anthropic's early internal codenames for its models, some of which observers find remarkably fitting.
The documents also include an early internal slide titled "Interpretability & Alignment, or Bankruptcy" arguing alignment teams must either save the world (if alignment is impossible) or save the company (if it's manageable), plus a note that the founding team was once called Sparrow Systems — led by GPT-3 alumni, physicists, and scaling-laws engineers who set out to train a large language model from day one.
More from Companies & People
- SGLang Summit 2026: BenchFlow founder to talk on building evals for superhuman intelligence — ying11231 · 2026-10-15
- Nvidia's Jensen Huang responds to Pope Leo's AI safety criticism — jeremyakahn · 2026-10-15
- 88% of AI Pilots Never Reach Production, But the Baseline Is Only 20-30% Anyway — rseroter · 2026-10-15
- Epic CEO Tim Sweeney says Fortnite's Linux support was vibecoded — Dimillian · 2026-10-15
- Anthropic acqui-hires Opaque Systems team, licenses its confidential computing tech for AI — rohanpaul_ai · 2026-10-15
- A Curated Who's-Who of Legal AI: Founders, Investors, Scholars and the Model Builders — EJWalters · 2026-10-15