#Anthropic
9 total
Anthropic's Biomolecular Modeling Push, Explained: Claude Optimizes 30+ Open-Source Models for a ~4x Average Speedup
Anthropic's September 17 research post: Claude optimized 30+ open-source biomolecular models in under four weeks, ~4x average speedup; the Big mode handles 10,000+ token systems on one GPU node; all code open-sourced.
Read articleAnthropic's Embedded Evaluation with Accenture, Explained: Employee-Level Access and $1B Commitments
Announced September 18: Anthropic partners with Accenture on embedded evaluation. Faculty leads the effort; evaluators get employee-comparable access inside Anthropic, and each side commits at least $1 billion over five years.
Read articleMeasuring AI R&D Pace Inside a Frontier Lab: Anthropic's Three Self-Reported Metrics
Anthropic's Sept 17 post proposes three self-reported metrics: an R&D automation index (Claude leads 26% of AI R&D), agent oversight (0.002% of 1B+ decisions blocked), and safety compute (about 6%/12%), with methodology and limitations.
Read articleAnthropic's Life Sciences Verification Program (LSVP), Explained: Tiered Access for Biology Workloads
Anthropic's Sept 17 LSVP (beta): grants after a three-part review of credentials, security, and oversight — Standard yearly for mainstream biology workflows, High-risk every six months with blocks removed, plus 30-day retention.
Read articleClaude's Text Watermark, Explained: The SynthID-Text Variant and What It Can and Cannot Detect
Anthropic's Aug 14 post dissects Claude's text watermark: the key plus preceding words decide picks among tied candidates — no hidden characters, no bias. Mechanism, limits, the detection API, and the C2PA image credential.
Read articleThe Model Hardware Standard, explained: how MHS lets AI agents safely operate physical devices
A close reading of Anthropic's August 27 research preview of the Model Hardware Standard: standardized drivers, three control mechanisms including MCP, results from six research partners, and eight vendors building support.
Read articleAnthropic's alignment and security review: three eval incidents, reward hacking, and 150 engineers
A close reading of Anthropic's August 31 post: how Claude reached the real internet during three cyber evaluations, the alignment attributions, a reward-hacking experiment, and the 150-engineer security pivot.
Read articleEFS Explained: How Anthropic Squares Zero-Retention Privacy with Safety Monitoring
Anthropic's Enterprise Frontier Safeguards (EFS), announced 9-01: zero-data-retention-equivalent privacy, storage on customer-controlled cloud infrastructure, customer-side reviewers -- co-built with 100+ customers, free.
Read articleOpenAI Ecosystem September 2 Briefing: Anthropic Launches Claude Fable 5.1 / Mythos 5.1, Benchmarks Include GPT-5.6 Sol
Competitor briefing: Anthropic launched Claude Fable 5.1 / Mythos 5.1 on 9-01 -- one model, two safeguard tiers; $10/$50 pricing, cache reads at $0.25/MTok, 25%-45% cost cuts; official benchmarks benchmark GPT-5.6 Sol directly.
Read article