An archived edition of vanemmerik.ai/Signal, preserved exactly as published — including that day's generative art seed.
That day's lede
On Sept. 1, Anthropic introduced Claude Fable 5.1 and Claude Mythos 5.1, calling them the company's most advanced models for coding and knowledge work, with early scientific-research capabilities. The two are the same underlying model with different safeguard levels — Fable 5.1 is generally available, while Mythos 5.1 is restricted to trusted-access programs for cybersecurity and life-sciences work. Anthropic cut prompt-cache-read pricing on both by 75% (to $0.25 per million tokens, down from the standard $1 rate), which the company says brings costs down roughly 25% for typical workloads and up to about 45% for highly agentic, cache-heavy work, while base pricing stays at $10/$50 per million tokens. Anthropic also detailed new cybersecurity safeguards that cut false positives by 60%, and separately announced Enterprise Frontier Safeguards, a customer-controlled-infrastructure approach to combining zero-data-retention privacy with misuse monitoring.
Claude Code · Sep 1
Version 2.1.257, shipped Sept. 1, added support for Claude Fable 5.1 as the new default Fable model, introduced a "Containment Escape" auto-mode rule that stops cloud metadata-credential fetches, egress evasion, and cross-tenant reach from being auto-approved, added configurable time-format/timezone settings, and made dozens of other fixes and VS Code extension changes. A same-day follow-up, 2.1.258, fixed Claude Code failing to launch on macOS 12 Monterey (a regression from 2.1.255) and a bug that broke remote/scheduled sessions after a re-sent permission approval couldn't be applied.
Claude Developer Platform · Sep 1
Alongside the model launch, Anthropic's Sept. 1 Claude Platform release notes specify that thinking blocks from the new models can only be replayed to the same model or a newer one, and that for accounts created on or after Aug. 31, 2026, replaying a thinking block after the system prompt, tools, or an earlier message changed now returns a 400 error. The release also added beta support for turn-scoped system messages that clear after one turn without invalidating the prompt cache, and confirmed that text from Fable 5.1 and Mythos 5.1 carries Anthropic's text watermark, with supported images/videos generated through the code execution tool carrying C2PA Content Credentials.
Anthropic · Sep 1
On Sept. 1, Anthropic announced Enterprise Frontier Safeguards (EFS), developed with more than 100 customers across financial services, healthcare, manufacturing, telecom, law, retail, and the public sector, plus cloud partners AWS, Google Cloud, and Microsoft Azure. EFS stores customer data in infrastructure the customer controls rather than Anthropic's, aiming to combine zero-data-retention-level privacy with the cross-session monitoring needed to catch sophisticated misuse; it starts rolling out in phases this fall, with eligible customers getting zero data retention on Fable 5 and Fable 5.1 in the meantime.