Summary
- 30k+ associates trained on Claude
- Flowsource runs Claude Code in spec-driven development
- Client deployments: 40% faster contract reviews; ~8 hours/week saved
Links
Resources
Recent
Latest Release Posts
Summary
- Anthropic opposes a blanket ban on open weights
- Prioritize chip export controls and anti-smuggling
- Mandate pre-release safety testing for capable models
Summary
- State-of-the-art on coding benchmarks
- Near Fable 5 performance at ~50% cost
- Stronger self-verification and alignment
Summary
- $200M fund for ambitious AI economic research
- Focus on large RCTs, scalable pilots, and public sharing
- Five priorities: workplace, retraining, income support, worker stakes, public investment
Summary
- Connector available in claude.ai
- Queries return Index-backed answers
- No install — enable via connectors menu
Summary
- Additional $20M to Public First Action (total $40M)
- Claude Mythos found thousands of high‑severity vulnerabilities
- Calls for enforceable transparency, independent evaluation, export controls
Summary
- Up to $50,000 in Claude credits
- Two tracks: basic science and biotech
- Apply by Aug 2, 2026 11:59 PM PST
Summary
- Free access for verified US K‑12 teachers
- Standards‑aligned lesson drafting and differentiation
- Student data excluded from model training (FERPA DPA)
Summary
- C$10M in Claude credits to Canadian institutions
- Amii, Mila, Vector added to Anthropic for Startups
- Canada ranks 2nd by per‑capita Anthropic Usage Index
Summary
- Claude Code automates test generation
- 50–70% faster chip validation in iDEC
- 20,000 UST staff trained on Claude
Summary
- Ben Bernanke joins Anthropic LTBT
- Adds macroeconomic and crisis-management expertise
- LTBT advises board and can appoint members
Summary
- Public call for AI’s hardest questions
- Surveys of 52k Americans and 81k Claude users worldwide
- Commitment to publicly track actions and report shortcomings
Summary
- Beta reflection dashboard in Settings
- Built on 4D AI Fluency Framework
- Excludes incognito & health data
Summary
- 466M lines scanned in 20 hours
- Automated fixes, tests, and rebuilds using Claude Code
- Continuous red/blue agent security reviews (95 controls)
Summary
- Fable 5 redeployed globally
- Four classifier categories; larger safety margin
- Draft jailbreak severity framework + HackerOne
Summary
- Beta on macOS/Linux for Pro/Max/Team/Enterprise
- Reproducible artifacts with code, environment, and provenance
- Managed compute across laptop, SSH/HPC, and Modal GPUs
Summary
- Agentic performance near Opus 4.8 at lower cost
- Intro pricing $2/$10 per million tokens through 2026-08-31
- Available as default on Free/Pro and via claude-sonnet-5 API
Summary
- Fable 5 redeployed globally July 1
- New classifier blocks reported jailbreak >99% of cases
- Blocked requests fallback to Opus 4.8; expect more false positives
Summary
- Beta on Slack for Enterprise and Team
- Channel-scoped memories and permissions
- Proactive asynchronous task automation
Summary
- Seoul office opens
- Large-scale enterprise Claude deployments
- NAIRL research access for 60 researchers