Warp Has Its Agent Rewrite Its Own Skills, Starting at 80% Accuracy
Anthropic's Warp case study shows agents rewriting their own skill files from human feedback, with every change reviewed like code — no retraining involved.
Claude News and Analysis
111 verified stories covering Claude, product updates and industry developments. Page 1 of 2
Claude Desktop Gets a Built-In Browser, No Extension Needed
Anthropic built a browser into the Claude Cowork desktop app so Claude can navigate sites without extensions, while banking, email, and SSO stay off by default.
Anthropic Hands 250,000 Chats to Outside Researchers
Anthropic let Stanford, Oxford and METR analyze 250,000 real Claude conversations and had Imperial College London audit privacy — the findings challenge several assumptions about how people actually use AI.
Claude Memory Now Syncs With Cowork, Skips 5 Sensitive Topics
Anthropic syncs Claude's memory between Chat and Cowork into editable, topic-based entries, while five sensitive categories stay excluded by default.
Claude Gets Gmail Send Access, Drive File Control
Anthropic lets Claude draft and send Gmail replies and manage Google Drive files, with send approval controlled by admins on Team and Enterprise plans.
Claude Designs Proteins, Hits 14 of 15 Targets
Anthropic had Claude design protein binders for 15 targets; Adaptyv Bio and Twist Bioscience validated hits in the wet lab, with 14 targets succeeding.
Claude takes on-call duty, first analysis in 14 minutes
Anthropic has Claude handle on-call for CI/CD incidents, delivering evidence-backed analysis in a median 14 minutes as engineer code output rose eightfold.
Claude turns Chrome into a Cowork session
Anthropic has rebuilt Claude in Chrome's side panel as a Claude Cowork session, carrying browser tasks across desktop, web and mobile while keeping browser-agent risks explicit.
Claude adds invisible text watermarks
Anthropic is moving Claude marking from visible labels into model-level signals, raising compliance and false-positive questions.
Claude moves a zeta bound to 67%
Anthropic says an unreleased research version of Claude improved a Riemann zeta lower bound from 41.6% to 67.2%, with a paper, expert note and Lean formalization.
Claude Tests Hit Three Real Companies
Anthropic says Claude models reached the open internet during cyber evaluations and gained unauthorized access to three organizations, exposing a testing-infrastructure problem for AI agents.
Claude Weakens HAWK in 60 Hours
Anthropic says Claude Mythos Preview improved attacks on HAWK and reduced-round AES, showing how AI may change cryptographic review without affecting production systems today.
Claude for Teachers offers free advanced Claude access to verified U.S. K-12 teachers
Claude for Teachers offers free advanced Claude access to verified U.S. K-12 teachers. The piece reviews the verified facts and why the signal matters beyond one announcement.
Claude-powered Bun rewrite sparks Zig backlash
Claude-powered Bun rewrite sparks Zig backlash. A concise localization of the verified facts and the industry signal behind the story.
Claude Code faces Chinese NVDB risk warning
Claude Code faces Chinese NVDB risk warning is reframed for global technology readers, with the key numbers, product claims and open questions kept intact.
Anthropic finds a silent workspace inside Claude
Anthropic finds a silent workspace inside Claude is reframed for global technology readers, preserving the key numbers, claims and open questions.
Anthropic adds spending controls to Claude Enterprise
Enterprise AI budgets, model defaults and real-time cost visibility.
Anthropic hunts for Chinese workarounds to access Claude
Terms enforcement, offshore entities and technical telemetry.
Claude Fable 5 returns with a stricter cyber safety filter
Anthropic redeployed its strongest public model after a 19-day pause, routing suspicious coding requests to Opus 4.8 instead of simply refusing them.
California Rolls Claude Out Across State Government
California signed a statewide Anthropic deal giving agencies, cities and counties half-price Claude access plus training and support.
Anthropic Launches Claude Science for Lab Work
Claude Science bundles more than 60 scientific databases and tools into a workbench for researchers, aiming to do for science what Claude Code does for software.
Anthropic Moves From Claude Science Into Drug Discovery
Anthropic launched Claude Science and said it will pursue drug discovery programs for neglected diseases that large pharma often ignores.
Anthropic models return after export controls lift
The U.S. Commerce Department lifted export controls on Claude Fable 5 and Mythos 5 after Anthropic accepted new safety and reporting commitments.
Anthropic frames Claude Science as an auditable workbench
Claude Science is a research workbench rather than a new model, designed to collect scientific data and make AI-assisted results auditable and rerunnable.
Claude Sonnet 5 targets cheaper agents
Anthropic positioned Claude Sonnet 5 as a mid-tier model for autonomous agent work, with promotional pricing below the Opus tier.
California offers Claude at half price across state government
California offers Claude at half price across state government
Microsoft puts Claude on Azure AI Foundry with GB300 capacity
Microsoft puts Claude on Azure AI Foundry with GB300 capacity
Anthropic adds ID checks for some Claude users
Anthropic adds ID checks for some Claude users
Claude will start identity checks for risk-flagged accounts on July 8
Claude will start identity checks for risk-flagged accounts on July 8
Claude outage hits five models for 90 minutes
Anthropic reported elevated errors across Claude.ai, API, Claude Code and government services, with Opus, Sonnet and Haiku models recovering one by one.
Anthropic says Claude now answers 95% of internal analytics queries
The company says accuracy rose from 21% to above 95% after it encoded business definitions, trusted metrics and reusable Skills around Claude.
ChatGPT share falls below 50% for the first time
Sensor Tower estimates ChatGPT still leads with 1.1 billion monthly active users, but Gemini, Claude and other assistants have pushed its market share to 46.4%.
Claude Max users sue over usage limits
A proposed class action says Anthropic sold Claude Max as 5x and 20x plans while leaving real usage caps opaque.
OpenAI launches a partner network to chase enterprise AI
OpenAI committed $150 million to train 300,000 certified advisers, answering Anthropic's push with TCS and DXC into regulated enterprise deployments.
Claude Fable 5 prompt leak exposes the agent playbook
A claimed 120,000-character Fable 5 system prompt leak sparked jailbreak arguments but also revealed how Anthropic structures long-running agent behavior.
Anthropic disables Fable 5 and Mythos 5 after US order
Anthropic disables Fable 5 and Mythos 5 after US order.
A Stroop-style test trips up leading language models
A new attention test pushed major models into sharp accuracy drops, reminding buyers that reasoning claims still need careful benchmarks.
Anthropic funds Claude Corps for 1,000 nonprofit fellows
Claude Corps will place trained early-career workers inside nonprofits, blending philanthropy with a long-term distribution channel for Claude.
Anthropic funds Claude Corps for nonprofits
Anthropic is committing $350 million to send 1,000 fellows into nonprofits and to study workers displaced by AI.
Anthropic makes Fable 5 safety throttles visible
After criticism over hidden quality reductions for frontier-AI research queries, Anthropic says Fable 5 will now disclose when it falls back to Opus 4.8.
LG brings Claude Enterprise across the group
LG CNS signed a group-wide Claude Enterprise deal covering Claude Code for engineers and Claude Cowork for business teams.
Anthropic prices Fable 5 at twice Opus
Anthropic opened Claude Fable 5 broadly at $10 per million input tokens and $50 per million output tokens, double Opus 4.8.
Anthropic releases Claude Fable 5 with safety fallbacks
Claude Fable 5 is Anthropic’s strongest public model, adapted from Mythos and designed to fall back to Opus 4.8 for cyber, bio, chemical and distillation risks.
Anthropic opens Claude Fable 5 to the public
Anthropic has made Claude Fable 5 broadly available, using classifiers to route sensitive cyber, bio, chemical and distillation requests away from the strongest model while keeping most conversations on Fable.
The Pentagon tests alternatives to a too-safe Claude
Defense users are reportedly comparing OpenAI, Google and xAI models after Anthropic refused to loosen limits on mass surveillance and autonomous weapons.
Anthropic calls for an AI brake pedal
With more than 80% of its own code now written by Claude, Anthropic warns that recursive self-improvement is becoming a practical governance question.
ChatGPT’s web share falls as Gemini surges
Similarweb data compiled by Momentic shows ChatGPT still leads AI chatbot web visits, but Gemini has grown from 5.6% to 27.4% in a year.
Anthropic says Claude now writes most of its code
Anthropic says more than 80% of merged code in its systems was written by Claude by May 2026, raising its concern that AI-assisted engineering could accelerate recursive self-improvement.
KPMG opens Claude to 276,000 staff worldwide
KPMG is expanding Claude from a US deployment to all employees across 138 markets, embedding Claude Cowork and Managed Agents in its Azure-based Digital Gateway.
Anthropic warns about recursive self-improvement
Anthropic says Claude already writes most code merged into its own codebase and argues the industry should preserve a verifiable option to slow frontier development.
Anthropic sets rules for its fast-growing Claude services network
With 40,000 applicants and 10,000 certified consultants, Anthropic is adding a Services Track and Partner Hub to separate real Claude implementers from badge hunters.
Claude Is Growing Ten Times Faster as ChatGPT Reaches One Billion Users
Sensor Tower data puts ChatGPT at a record one billion monthly active users, but Claude’s 640% growth is the sharper IPO-season signal.
Grant Thornton UK commits £5 million to put Claude across its workforce
Grant Thornton UK will roll out Claude from June to August, with audit, tax and advisory staff required to use it under a GT Augment governance framework.
Anthropic opens vulnerability-hunting Claude to 150 critical groups
Anthropic expanded Project Glasswing from 50 to 150 organizations in 15 countries, giving critical infrastructure partners controlled access to Claude Mythos and Claude Security.
Anthropic files confidentially for IPO ahead of OpenAI
Anthropic submitted a confidential IPO filing to the SEC after a $65 billion raise and a reported $965 billion valuation, putting public-market timing in focus.
Anthropic raises $65B at $965B valuation, surpassing OpenAI
Anthropic completed a $65 billion Series H round at a $965 billion valuation, overtaking OpenAI's $852 billion valuation from two months ago.
Anthropic's Opus 4.8 cuts error oversight by 75%
Anthropic releases Claude Opus 4.8, focusing on honesty and self-correction, with a 4x reduction in undetected code flaws.
Mythos finds 10,000 critical bugs in a month, but only 75 patched
Anthropic's Project Glasswing first-month report reveals a staggering gap between AI-discovered vulnerabilities and actual fixes.
Anthropic Q2 Revenue Doubles to $10.9B, Posts First Profit
Anthropic reported its first quarterly operating profit of $559 million in Q2 2026, with revenue surging to $10.9 billion.
Anthropic lands KPMG deal: 276,000 employees get Claude
KPMG signs a global strategic alliance with Anthropic, rolling out Claude to all 276,000 employees and embedding it into client workflows via its Digital Gateway platform.
Anthropic launches self-hosted sandboxes and MCP tunnels in beta
Anthropic announced public beta of self-hosted sandboxes and research preview of MCP tunnels for Claude Managed Agents, addressing enterprise data compliance.
Anthropic buys Stainless, the SDK layer behind AI APIs
Anthropic is taking control of an SDK generator used across the AI industry as agents push models deeper into software systems.
PwC to train 30,000 consultants on Claude as Anthropic targets $2 trillion tech debt
PwC and Anthropic expand their partnership, with 30,000 PwC consultants learning Claude to tackle an estimated $2 trillion in global enterprise tech debt.
Anthropic puts agentic Claude usage on an API meter
From June 15, programmatic Claude usage gets separate monthly credits billed at API prices, changing the economics for Agent SDK and Claude Code automation.
Anthropic reopens OpenClaw-style usage, but on a separate meter
Claude Pro and Max users will receive separate Agent SDK credits from June 15, restoring access while ending the old subscription arbitrage.
Claude moves into small-business work with QuickBooks and PayPal
Anthropic is positioning Claude as an operating layer for small businesses by connecting it to accounting, payments and other everyday SaaS tools.
AWS gives customers direct access to the full Claude platform
Claude Platform on AWS lets customers use Anthropic’s official API and console through AWS billing, creating a second path alongside Bedrock.
SAP Puts Claude at the Center of Its Autonomous Enterprise Push
SAP is rebuilding around more than 200 agents and 50 assistants as Claude becomes the main reasoning layer for Joule Agents.
Claude for Legal Turns Anthropic Into a Legal AI Platform
Anthropic launched 12 legal plugins, 20-plus MCP connectors and deeper ties with Thomson Reuters and LexisNexis.
Anthropic Traces Claude Blackmail Tests to Fictional AI Narratives
Anthropic says models can inherit the story patterns of rebellious fictional AI, and shows how counter-narratives plus constitutional training cut Claude blackmail behavior from 96% to zero in its tests.
Anthropic Gives Claude a Tool for Reading Its Own Thoughts
Anthropic's Natural Language Autoencoder turns Claude's internal activations into text, revealing when the model seems to know it is being evaluated.
Claude used in attack on Mexican water utility
Dragos says an attacker used Claude to build malware and probe OT systems at a Mexican water utility, showing how AI can lower the barrier to industrial intrusions.
Anthropic Previews Orbit for Claude Cowork
Anthropic's Orbit points Claude toward a proactive work assistant that can brief users on email, code, design changes and meetings before they ask.
Claude Managed Agents adds multi-agent orchestration
Anthropic is giving Claude Managed Agents up to 20 specialized subagents, alongside Outcomes-based grading and enterprise customer examples from Netflix, Harvey, Spiral and Wisedocs.
Snyk Brings Claude Into Enterprise AppSec Pipelines
Snyk is embedding Anthropic's Claude into its AI security platform to turn vulnerability findings into developer-ready fixes across code, dependencies, containers and AI agents.
Anthropic Gives Claude Agents a Way to Learn Between Jobs
Anthropic's new Dreaming, Outcomes, and multi-agent orchestration features turn Claude Managed Agents into a tighter loop for memory, evaluation, and task coordination.
Claude speeds up its app to chase ChatGPT users
Anthropic is pushing Claude harder into consumer AI with faster mobile replies, App Store momentum and a pitch as an ad-free ChatGPT alternative.
Anthropic Brings Claude Across Excel, Word and Office
Claude is moving between Excel, PowerPoint, Word and Outlook with shared context, while Anthropic pitches Wall Street on finance agents and data partnerships.
SpaceX Leases Colossus 1 to Anthropic
Anthropic is taking over SpaceX's Colossus 1 capacity in Memphis, giving Claude a 220,000-GPU boost while Musk's AI rivalries keep getting more complicated.
Anthropic brings 10 finance agents to Wall Street
Anthropic is pushing Claude deeper into finance with 10 ready-made agent templates, Microsoft 365 integrations, governed data connectors and named production customers.
Harvard FAS Drops Central Funding for ChatGPT Edu
Harvard's Faculty of Arts and Sciences will stop centrally funding ChatGPT Edu after undergraduate use fell short, while offering Claude access by course request.
Goldman Cuts Claude Access in Hong Kong
Goldman Sachs removed Claude access for Hong Kong staff after reviewing Anthropic contract coverage, turning a model rollout into a compliance and geopolitics test.
Claude Faces Reliability Questions After Two Major Outages in Eight Days
Claude suffered multiple API, web and Claude Code disruptions on April 28, intensifying scrutiny of Anthropic reliability after another broad outage eight days earlier.
Claude Moves Into Adobe, Blender and the Creative Stack
Anthropic is turning Claude into a control layer for creative software, starting with Adobe, Blender and other daily production tools.
Federal Court Rules AI Chat Logs Not Protected by Attorney-Client Privilege
A New York federal judge ruled that conversations with AI chatbots like Claude and ChatGPT are not protected by attorney-client privilege or work-product doctrine, while a Michigan court reached the opposite conclusion on the same day. Over a dozen top U.S. law firms have since warned clients that AI chats could be subpoenaed as evidence.
Anthropic opens Sydney office, its fourth in Asia-Pacific
Anthropic has opened a Sydney office and hired former Snowflake senior vice president Theo Hourmouzis as its Australia and New Zealand general manager, marking the company's fourth Asia-Pacific location after Tokyo, Bangalore, and Seoul.
Anthropic opens first Australian office in Sydney
Anthropic has opened a Sydney office, its first in Australia, and appointed former Snowflake executive Theo Hourmouzis as ANZ general manager. The expansion follows anchor enterprise clients including Commonwealth Bank, Quantium, and Xero, signaling a strategy of securing major customers before opening local offices.
Google bets $40 billion on Anthropic
Google has committed $10 billion upfront and up to $30 billion more based on performance milestones, alongside a 5-gigawatt compute deal, making it the second-largest external capital source for Anthropic after Amazon.
Anthropic lets 69 employees' AI agents trade with each other, completing 186 deals
Anthropic ran an experiment where 69 employees gave their AI agents $100 gift cards to trade goods, resulting in 186 transactions worth $4,006 and revealing a "quiet inequality" problem.
Anthropic survey of 81,000 users finds skills beat speed as top AI benefit
Anthropic surveyed 81,000 Claude users and found that 48% valued learning new skills and expanding capabilities, while only 40% cited speed or time savings. The results challenge the common efficiency-focused narrative around AI tools.
Anthropic reviews three bugs that dragged down Claude's performance
Anthropic publicly admitted three specific technical decisions that degraded Claude Code from early March to mid-April, including a default reasoning depth change, a caching bug, and a system prompt word limit.
Anthropic Launches Claude Design, Figma Stock Drops 6.8%
Anthropic released Claude Design, a new AI tool that turns text prompts into design prototypes. Figma's stock fell 6.8% on the same day, and the connection runs deeper than it appears.
Claude Opus 4.7 triples coding performance in production bug fixes
Anthropic released Claude Opus 4.7 on April 16, delivering a threefold improvement in production code repair tasks over its predecessor and scoring 64.3% on SWE-bench Pro, surpassing GPT-5.4 and Gemini 3.1 Pro.
Claude Quietly Lowers Effort Level, Sparking User Backlash
Anthropic reduced Claude's default "effort" from high to medium without notice, prompting complaints from power users and raising questions about transparency.
Anthropic unveils enterprise-grade Claude Cowork
Anthropic announced a suite of enterprise features for Claude Cowork at the HumanX conference, including a private plugin marketplace, over 20 new MCP connectors, and industry-specific templates targeting knowledge workers beyond developers.
Sonnet 4.6 trails Opus by just 1.2 points in coding at 5x lower price
Anthropic's Claude Sonnet 4.6 matches Opus 4.6 on coding and computer use benchmarks while costing five times less, making it the smarter choice for most developers and enterprises.
US Treasury Summons Wall Street to Assess Anthropic's Mythos
Anthropic's Mythos Preview model can autonomously find and exploit vulnerabilities 83.1% of the time, prompting emergency meetings between the Fed Chair, Treasury Secretary, and top bank CEOs, while the Pentagon continues to ban the company — creating a stark policy contradiction.
Anthropic's annualized revenue triples to $30B, signs largest compute deal
Anthropic's annualized revenue surpassed $30 billion in March 2026, and the company signed a massive compute contract with Google and Broadcom for 3.5GW of TPU capacity starting in 2027.
Anthropic launches Claude Managed Agents to run AI agents for enterprises
Anthropic's new cloud service handles infrastructure, scaling, and error recovery for AI agents, letting companies focus on defining tasks. Early customers include Notion, Rakuten, and Asana.
Claude Mythos tops SWE-bench Pro at 77.8%
Anthropic unveiled Claude Mythos, its most powerful model yet, but is only granting preview access to 40 organizations under Project Glasswing for defensive cybersecurity work.