Anthropic replaces Git with Scion after 30,000 merges a day
2 sourcesSeven engineers spent five months building a server-centric, API-first system to remove infrastructure bottlenecks.
New publication notifications
All Anthropic news
All 159 AlexTech.ai stories on Anthropic (the company that builds the Claude models) since June 9, 2026, newest first, each with its original sources.
Anthropic warns small AI models could self-replicate on IoT devices
7 sourcesAnthropic's red team found small models can run on cheap hardware like Android TV boxes, though no wild worm exists yet.
Anthropic bans abuse of Claude: users who bully the chatbot can be banned
3 sourcesEffective Nov 12, the policy targets extreme cruelty, alongside new restrictions on propaganda and weapons development.
AI companies prepare response plans for a major hacking attack
4 sourcesExecutives at OpenAI and Anthropic reportedly plan responses if a catastrophic AI incident sparks public and political backlash.
Anthropic offers free AI scans to open-source projects
2 sourcesEligible maintainers can nominate GitHub repositories and receive automated reports with reproduction steps and potential fixes.
Google gives Gemini a work account so it can act like a colleague
9 sourcesThe agent can plan tasks, create sub-agents and choose between Google's Gemini models and Anthropic's Claude.
Claude Haiku 5.5 costs less and lets developers tune its effort
9 sourcesAnthropic sets rates at $0.10 per million input tokens and adds five adjustable effort levels for developers.
Anthropic may have picked Samsung for a 1,300-watt AI chip
8 sourcesBroadcom would handle the design, targeting 4,000 TOPS and potentially up to 288 GB of HBM3E memory.
Anthropic gives startups a free year of Claude Team
8 sourcesEligible companies also get $1,000 in API credits and partner offers worth up to $45,000.
Claude can edit Google files without leaving Docs, Sheets or Slides
7 sourcesAnthropic's new Google Workspace beta allows Claude to write formulas, build pivot tables, and rewrite text via a sidebar.
Anthropic reported threats written in Claude to the police
8 sourcesAfter a human reviewer flagged credible risks, Florida deputies arrested Carli Michelle Heller for felony violent intent.
Anthropic needs to cut prices for models used by AI agents
5 sourcesA Reddit post compares GPT-6 Luna and Haiku 4.5 pricing, arguing that cached input deepens the cost gap.
Info
AI-Generated Content Notice: The reports, articles, headlines, and images published in this section are produced with the aid of artificial intelligence tools and are subject to human supervision. Despite reviews and checks, content may contain errors, omissions, or missing updates. Materials are for informational purposes only and do not replace official sources or professional advice.
Follow our channels