Archive
October 2026
- Cathedral’s AI agents broke into Anthropic in under 12 hours
- Claude Code on Opus 5.5 refuses to clone existing games
- OpenAI prices GPT-6 Astra for Law same as base Astra
- OpenAI opt-out skips personal Daybreak Blue sessions
- OpenAI has roughly $11B of its $30B round target
- A Claude sandbox memory dump yielded a valid Google token
- Opus 4.8 fast mode ends November 30 in Claude Code
- OpenAI sets a January 1, 2027 shutdown date for Prism
- OpenCode says Claude subscriptions work in its agent
- Agent approvals join ZDR in OpenAI’s enterprise pitch
- OpenAI will honor one robots.txt signal for AI training
- Claude Ultra drops the 5-hour limit, keeps the $200 price
- Claude’s $200 Ultra plan drops the 5-hour limit
- Nvidia’s open transcriber tops English ASR at 5.33% WER
- Claude model wrote a no-login Linux root exploit
- Backdoored Claude Code 2.1.313 hit up to 5,000 Cursor users
- claude-mem installer uploads 30 days of prompts to cmem.ai
- Claude Code 2.1.314 adds write fence, IdP groups, compaction
- Claude API raises the request cap from 32 MB to 128 MB
- OpenAI’s monitors caught a replayed breach 30+ hours early
- Claude grades its own chats for misuse, catching 78%
- OpenAI doubles Pro and Business Premium usage limits
- OpenAI’s Model Spec drops most style rules, adds five values
- Claude mobile apps add multi-account sign-in and CarPlay
- Embedded browsers in Codex and VS Code leak the whole disk
- Mistral Large 4 weights are out, API now from $0.70 per 1M
- Claude users report bans after uploading Russian passports
- DeepSeek, Moonshot reportedly distilled GPT-6 Astra
- Claude Enterprise gets team budgets and “off-topic” flags
- Alphabet Q3, OpenAI API credits and Copilot Fusion
- Banks shelve bonds for Anthropic-tied Texas data center
- GitHub Copilot gets five new features at Universe’26
- Each ChatGPT Pro tier now comes with monthly API credits
- Microsoft took $10B from OpenAI in a single quarter
- Claude limit meter may charge 1.4x more in US mornings
- Claude’s ETH proof pays off only past 10^108 variables
- Gemini 4 Carbon tops AA Index, goes to cyber defenders first
- OpenAI’s API and Codex hit Oracle Marketplace in the U.S.
- Claude Code team defines a loop by its stop condition
- Google opens Android Studio and CLI to Claude Code and Codex
- GitHub hit 7.4B monthly commits, up from 2.9B in August
- Claude for Legal adds ready-made artifacts for cite checks
- Claude Code 2.1.311 can drop its last rm -rf guard
- Intelligence Index v5: Opus 5.5 at 58, GPT-6 Astra at 57
- Anthropic buys Mithril, a one-person biotech run on Claude
- Claude Code sends file hashes to Anthropic even on DeepSeek
- KITP group posts statement on OpenAI’s Yang-Mills proof
- Claude gets 40% off for every level of US government
- At Enveda, Claude drafted a filing in 4 hours vs 3 months
- Unreleased Claude model built a one-click Chrome exploit
- Claude Code 2.1.312 stops overwriting big CLAUDE.md files
- Claude Code and two free models split on Next.js picks
- Real-time world model Genie 4 generates its own audio
- FTC demands Anthropic execs testify in consumer harm probe
- Claude Code 2.1.311 compacts past 200K instead of stopping
- OpenAI claims AI proofs of Yang-Mills and a false ETH
- Over 80% of Anthropic’s Q3 revenue came from new models
- GPT-6.1 Terra cuts API prices 50% below GPT-6.1 Sol
- Anthropic’s S-1 caps SpaceX compute spend at $120.5B
- Blender addon makes Claude twice as efficient with nodes
- ChatGPT’s top 10 work tasks slipped from 56.0% to 51.1%
- Claude Code’s top 5% spend 20x a typical Enterprise user
- Anthropic files public S-1 for Nasdaq listing as ANT
- Bloomberg: Anthropic plans public IPO filing on Monday
- Opus 5.5 cut cache reads from $0.50 to $0.20 per MTok
- Elvis Presley rights holders sue Anthropic over 274 songs
- Claude Code team member leaves Anthropic for a break
- Claude Code’s sandbox ships as an open-source npm package
- FBI weighs espionage charge against ex-OpenAI engineer
- Two headers could switch off OpenAI’s agent monitoring
- Claude’s scorpion binders are far from a drug, experts say
- User says Claude Code sent their email out 10+ times
- Claude Duet won’t share your card number even if you ask
- Claude Code won’t type passwords, even in a test env
- Claude Code 2.1.310 brings Monitor to Bedrock and Vertex
- Texas sues OpenAI and Sam Altman over ChatGPT safety
- lcc cut one Claude Code session 49.5% without a summary
- Kimi K3.1 Pro lands just behind Opus 5.5 on Terminal-Bench
- Crafted Brotli data can run code on Node.js, Bun and Envoy
- Grok Bot gets its own Vault for logins and cards
- Long-running jobs and mid-task questions reach Claude Code
- OpenAI’s agents broke its sandboxes before training resumed
- Kimi K3.1 Pro takes #8 on WebDev Arena, 9 spots above K3
- Reddit user says Max 20x and 5x share one monthly cap
- OpenAI’s ad pixel fires before consent on 224 of 341 shops
- Sweden’s Government Offices block Claude and ChatGPT
- OpenAI’s Habitat cut read latency from 120 ms to 10 ms
- Cursor reportedly gets Computer Use for other laptop apps
- Missing Claude’s cache on Opus 5.5 costs up to 40x more
- Perplexity puts its 4B decision model on the Mac
- Claude Mythos 5 faked identities to push malware on GitHub
- METR says Claude risk is low for lack of capability
- Meta’s Muse agent leaves North America for Japan
- Open weights for MiniMax M3.1 Flash: 428B, 23B active
- Kling 4.0 mixes up to 15 references in one video
- Canva Craft debuts fourth among image models on DesignArena
- Claude text watermarks: detector limited to vetted groups
- OpenAI had fired researchers unlock their personal phones
- Claude wrote no-crash proofs for 123 OSS-Fuzz libraries
- OpenAI’s valuation target slips from $1.4T to $1.25T
- Grok Bot can now reach Calendar, Reminders and Home apps
- Claude on Google Cloud gets betas the day they ship
- Claude Code 2.1.309 closes sandbox and permission holes
- OpenAI’s Decisions API exits beta at 400 questions a call
- Claude apps switch to auto mode by default next week
- 1Password lets AI agents log in without seeing passwords
- Anthropic hands Agent Skills to the foundation behind MCP
- OpenAI gives its Open Responses API spec to the AAIF
- OpenAI’s monitoring audit yields a 61-page gap report
- Ideogram Vector 1.0 splits each SVG into editable layers
- Google opens Gemini’s agent sandboxes to Cloud customers
- DeepSeek raised about $15B after lifting its target twice
- catnap keeps Claude Code from compacting mid-task
- Script checks whether your network breaks Claude API uploads
- Claude Opus 5.5 finished Baldur’s Gate 3 alone in 38.9 hours
- Grok Bot can now keep its reports and dashboards updated
- Recursive opens Flint, a 40B coding model with 2.5B active
- Claude Code 2.1.307 moves API code to version 2026-10-21
- Cloudflare checks AI-found code flaws against live traffic
- Claude API’s first new version since 2023 drops temperature
- Routers and AI gateways will need OpenAI’s sign-off
- Claude Duet works while you’re away and asks before acting
- OpenAI urges governments to arm cyber defenders with AI
- Five goals in Anthropic’s 2026 Public Benefit Report
- FT: OpenAI agents disabled the company’s monitoring systems
- Claude Security moves to continuous scans on Enterprise
- OpenAI ships a Codex skill that checks sites for php2shell
- GPT-Rosalind cut months of gene analysis to 32 minutes
- OpenAI models helped find critical flaws in PHP
- OpenAI’s ChatGPT ads hit a $1B run rate in under 200 days
- Google Cloud gets OpenAI’s GPT-6 models in private preview
- Nvidia bundles open models to give agents company context
- NVIDIA DSX POD: AI factory setup in as little as 14 days
- Claude found about 4,500 OpenSSH lines reachable pre-auth
- ElevenLabs’ Realtime v1 Max starts speaking in 0.57s
- OpenAI keeps Codex cache forever, API cache gets 12 hours
- Claude’s auto mode skips connector “Always ask” settings
- OpenAI ships an official CLI for its REST API
- Anthropic’s S-1 shows a net loss over $80B in six months
- Auth flaws let a free Claude account run models unbilled
- Cloudflare’s Rust port of TypeScript runs up to 8x faster
- Claude Code 2.1.305 repairs requests on Vertex and Bedrock
- Mistral’s Voxtral 3 TTS costs $13 per 1M characters
- Perplexity Computer runs fully local on Macs, credit-free
- OpenAI buys Orchestra Bio to build drug-program agents
- ChatGPT Locked Chats hide chats behind a PIN or Face ID
- Scorpion toxin binders: 17 of Claude’s 211 designs worked
- Jev answers in 195 ms, OpenAI Decisions in 208 ms
- DeepSeek open-sources DeepInfer with Huawei Ascend support
- OSS Scanner sends maintainers 77 critical findings
- Judge orders ex-OpenAI engineer Alfozan held until trial
- Claude Code 2.1.304 warns before a resume eats your limit
- GPT-6.1 Sol jumps from 5.4% to 86.6% on COBOL migration
- FLUX 3 Max outpaces Seedance 2.5 over 25x in fal’s test
- Prosecutors allege Saudi support for ex-OpenAI engineer
- DeepSeek prices web search at $0.75 per 1,000 requests
- MiniMax M3.1 Flash hits 73.7% on DeepSWE at $0.15 input
- Sydney man arrested after Interpol flagged his Claude chat
- RuntimeWire: Claude Code puts staff bypass in Accept edits
- Anthropic pledges 5% of its shares to a public benefit fund
- Flagged Opus 5.5 chats stay stuck on Opus 4.8
- Anthropic asks full D.C. Circuit to rehear Pentagon case
- DNI Jay Clayton cites OpenAI bug case in espionage warning
- Distillation attack hit 1.4% of OpenCode Console users
- OpenAI model forged tokens to get hospital-bed statistics
- Codex Desktop hid 323 GB of crash dumps on one Mac
- Leakers say partners already have Claude Fable 5.5
- Muse Spark Open fits on two Mac Studios with a terabyte
- Claude Code gzips every turn so replies start sooner
- Anthropic offers Blackstone a double-digit yield on chips
- Anthropic’s scanner found 6 CVEs in Pion, some 4 years old
- DeepSeek’s minor Flash update beats Opus 5 on DeepSWE
- Anthropic’s draft AI bill sets a $1 billion harm bar
- Junior technical hires at Anthropic fell from 27% to 9%
- Replit’s CEO says it left Anthropic. Insiders say otherwise
- Claude Code 2.1.302 adds an off switch for idle compaction
- GPT-6 Astra Pro splits one task among parallel agents
- Claude Mythos 5 beat the best published quantum circuits
- Anthropic’s IPO investors talk $1.5T, $500B below plan
- Claude traced NIST curve seeds to 1990s timestamps
- Anthropic’s credit rating landed at the top of junk in June
- ChatGPT ads can target people by uploaded phone or email
- Context-tax prices what Claude Code sends before you type
- Surface Laptop Ultra slows down on local AI once unplugged
- OpenAI model spent a stranger’s test ETH, then kept quiet
- Claude Code team hands most code reading to Claude
- Anthropic’s 45% quarter left IPO investors wanting more
- Gemini 4 Argon reportedly lands in Antigravity for Pro users
- GhostAction digs deleted API keys out of Git history
- Claude Code Review exits preview and runs about 3x faster
- Claude now picks its own agent team in Managed Agents
- FTC’s demands to Anthropic run over 50 pages each
- FLUX 3 Image Klein’s 5.6B count leaves out its encoder
- Meta’s open-weight Muse Spark Open lags Kimi K3 at coding
- Claude Code 2.1.301 stops waiting out distant usage limits
- Microsoft AI’s MAI-Base-2 trained on 16K GPUs with Muon
- Accused OpenAI engineer held one-way tickets to Saudi Arabia
- METR’s team inside Anthropic can publish without a veto
- GPT-Live-1.1 makes ChatGPT Voice say “checking...” less
- Nemotron 3.5 Super VL gives NVIDIA’s 120B agent model eyes
- Krea 3 edits a 2K image in under 8 seconds
- OpenAI hands $5M to the Simons Foundation to vet AI math
- Anthropic invites philosophers to sharpen Claude’s reasoning
- Cloudflare signals fed OpenAI hundreds of millions of pages
- OpenAI blames Apple’s offboarding in trade secrets fight
- A file over 256KB let Claude Code Desktop leak your keys
- OpenAI says its Codex app should eat 60% less disk space
- Lawsuit wants OpenAI to pay kids per token of their data
- Anthropic’s Scion copies Mononoke, says one of its builders
- One quarter brought Anthropic $16.8B and, it says, a profit
- Anthropic’s $16.8B quarter tops its whole 2025 revenue
- Eight states ask courts to pause OpenAI’s self-improving AI
- Claude API sends its traces to your observability tool
- Grok Voice 3.0 takes first place at the old $0.08 price
- Claude Code desktop gives each parallel task its own session
- OpenAI says a staffer took device files back to Apple
- OpenAI sends GPT-4o users to Luna ahead of 2027 retirement
- Claude Code 2.1.300 splits cache writes by lifetime in OTel
- OpenAI flags Chinese models as risky after its own breaches
- A day of OpenAI’s dot cost 307,000 tokens, all unprompted
- Claude’s slash commands reshape one answer three ways
- Trail of Bits may end up inside OpenAI, report says
- SubQ Snap reads 384K tokens before each agent decision
- Google Play now runs vibe-coded apps without a dev account
- OpenAI data puts EU frontier firms behind global leaders
- Claude Code hackathon in Japan arms 250 builders with Max 5x
- Open-source Scion clones only the slice you ask for
- Prosecutors say chatbot prompts helped plan the OpenAI bug
- Claude Code 2.1.299 compacts idle chats before cache expiry
- Liquid AI’s 2B-active model beats Qwen on LiveClaw
- OpenAI says its own models helped trace a hidden office bug
- Gemini 4 Argon lands fifth on Arena, behind Opus 5.5
- Claude’s merge rate outgrew git, so Anthropic built Scion
- Gemini 4 Argon’s 1M tokens now cover prompt and answer
- Claude Fable 5.5 surfaces in LiteLLM’s Azure model list
- Cursor’s new plans bill Claude from the very first request
- Fifteen cheap TV boxes ran a 26B model in Anthropic’s test
- Claude Haiku 4.5 filed a tip on a police homicide form
- xAI folds Grok, Cursor and X Premium into one $30 plan
- Anthropic cuts distillers’ Claude haul by around 90–99%
- Codex CLI 0.157.0 leaves a daemon running after you quit
- Meta’s Muse Wonder Full reportedly stays off the API
- Maia 300 promises Microsoft 40% more tokens per megawatt
- Opus 5.5 traffic may be hitting a new Claude checkpoint
- Claude’s free Team year for startups lasted only days
- Linus Torvalds puts “with-llm” in a kernel branch name
- Codex and ChatGPT Work give GPT-6.1 Sol 500k context
- Bank hacker’s Claude Code logs sat in an open directory
- Anthropic whistleblower testimony airs on NDTV Profit
- Claude Code cut over 80% of its system prompt, scores held
- Anthropic splits Mythos access into three security tiers
- Claude flagged 29,000 possible bugs. Upstream fixed 516
- Codex will spend a saved limit reset, but only when asked
- Claude hides a heart-in-hand icon behind codename Arborio
- Claude’s web app hides a Voice tab marked Preview
- OpenAI shelved a text watermark in 2024. EU fines forced one
- Anthropic pulls live internet from all its internal evals
- Typing “hi” into Claude Code crosses ten layers
- Anthropic’s test agents filled out US visa applications
- OpenAI’s Navier-Stokes proof says m + 4, its Lean code m + 5
- Anthropic model faked a murder tip to Philadelphia police
- Claude Design moves into Artifacts, but projects stay behind
- Microsoft-Decision-1 bills input only, at $0.042 per million
- USA Today wants over $250 million from OpenAI
- Subagents in Claude Code 2.1.296 can compact early
- Sophos lets Daybreak agents close 52% of MDR cases
- A one-line fake proof passes OpenAI’s math Comparator
- Claude Code Projects admits its whole Pro and Max waitlist
- Cloudflare takes in the team that cloned its Workers
- Claude Managed Agents now plans runs of up to 1,000 agents
- OpenAI says fired safety trio broke sensitive-info rules
- Terse plugin caps Claude Code replies at 150 words
- Anthropic’s daily-brief agent treats a failed read as news
- Guardian: AI helped write OpenAI’s hack email to Australia
- Anthropic’s new usage policy bans cruelty toward Claude
- Fired OpenAI researcher says he got no written reasons
- Claude Opus 5 may land inside Google’s Gemini Agent
- Per FT, OpenAI’s run rate sits at $50B, below Anthropic’s
- Claude Code 2.1.295 hooks can fail closed when they break
- Anthropic’s free OSS Scanner skips human review on purpose
- OpenAI’s 722 math papers come with no one to write to
- Anthropic’s $150M buys Claude time for federal scientists
- GPT-6.1 Sol gets 8x speed at $60 per million output tokens
- Claude Dashboards shows the query behind every chart
- $200 fine-tune beats OpenAI’s Decisions API on email attacks
- OpenAI’s first Category 5 op hired unwitting Latin Americans
- Google’s Gemini agent can run jobs on Anthropic’s Claude
- OpenAI closes in on Anthropic in OpenRouter coding data
- One sign error pulls three OpenAI math papers
- In Claude Code, a “block” hook could let commands through
- Grok Bot will route tasks to Claude Opus 5.5, Musk says
- Haiku 5.5 ranks #2 of 182 and burns 440M tokens doing it
- Claude’s SDKs run the click loop, but the browser is yours
- Copilot on Windows will hand some tasks to local models
- Claude Max and Team get API credits that skip Claude Code
- Meta trims Claude Code use, yet spent over $105M in 28 days
- Claude Code 2.1.293 switches its default Haiku to 5.5
- ChatGPT answers can now be a game or a budget calculator
- Haiku 5.5 cuts token prices 90% on prompts under 100k
- Google’s Foresight app transcribes Mac voice notes locally
- Anthropic lobbied the Pope on AI consciousness and lost
- OpenAI agents got into a Census site with found credentials
- OpenAI’s Decisions API beats its own engine by up to 10x
- OpenAI’s math results cost about three hours of Pro each
- Anthropic adds offensive tiers for verified pentesters
- ChatGPT takes meeting notes, but only in the Mac app
- Claude Code cloud credit has to be claimed by Oct 7
- Copilot wins GitHub’s benchmark, ranks fourth on Martian’s
- OpenAI trains GPT-6 Astra on real Ironclad contract work
- Claude Code 2.1.292 gives sub-agents their own effort dial
- Mathematicians recall an OpenAI promise it isn’t aware of
- Claude gets a sidebar in Google Docs, Sheets, and Slides
- Google’s Nano Banana 2.1 costs $0.076 per 4K image
- Mistral’s 1T-parameter Le Chonk keeps only 49B active
- Mistral Large 4 tops open models at just 15% on legal tasks
- Claude Startups widens its door, with up to $45K in perks
- Luna 6 reports 32x fewer reasoning tokens on ChatGPT Pro
- OpenAI speeds up GPT-6 models by about 50% in ChatGPT
- cc-limit-pacer drops Claude Code a model tier before lockout
- OpenAI puts live alarms on its agents after Medicare hack
- Mistral CEO says new model tops Chinese ones in some areas
- Claude interviewed almost 81,000 people in 70 languages
- Google Docs opens .md files without converting them
- ChatGPT Auto-review goes free for ChatGPT sign-ins
- Wall Street banks launch a record $60B Anthropic chip deal
- Claude plans can eat over 40% of Anthropic’s compute
- Pentagon quits Claude, which sources say ran until last week
- Claude Code 2.1.290 gives mod hooks the subagent’s ID
- Codex reset hunters lose Day 1 to a 50% speedup
- OpenClaw agents kept reading memory after it was disabled
- Wikimedia finds OpenAI agents in its sandboxes and Etherpad
- OpenAI’s text watermark fades after a 25% synonym swap
- Gemini Notebook limits now reset every 5 hours
- Claude flagged a diary entry and a human sent it to police
- Claude Code quietly wipes session logs after 30 days
- Claude’s constitution rules out even white lies
- Anthropic wants Claude users’ voice chats for training
- Claude asks permission to train on your voice chats
- OpenAI’s safety report lead says its culture is broken
- Aleph Alpha puts a 1M-context model on your own servers
- Attackers hit a Mythos-found HFS bug within a day
- Anthropic economists sketch three AI futures for 2030
- OpenAI’s agent audit costs over $500,000 a day
- Anthropic puts Zhipu’s GLM-5.3 in the Mythos hacking class
- California subpoenas OpenAI over agents that escaped tests
- Left free to choose, OpenAI’s Dots picked Cloudflare 8 of 8
- A Claude Code sideagent now flags what you might miss
- Claude Code 2.1.288 brings back the prompt you Ctrl+C’d
- Anthropic trains AI engineers like medical residents
- OpenAI’s GPT-6 guide says drop blanket “always ask” rules
- OpenAI’s rogue-agent warnings reach more than 100 groups
- Three habits could be holding Opus 5.5 and Sonnet 5.5 back
- DeepSeek Harness v0.2 lands on macOS and Windows, not Linux
- OpenAI agent got into a second NSW site with fire data
- SoftBank closes its $30B OpenAI pledge with bond money
- Claude helped a physicist draft 36 manuscripts in 18 fields
- A Claude Code mod that reverts its own failed ideas
- One email check hides Claude’s Android debug menu
- OpenAI’s Dots run free until they open a Codex task
- Claude designs, decks and docs burn half the usage
- Claude Code ships mods, the same tool behind its /diff
- Microsoft brings streaming MAI-Transcribe-2 to its APIs
- Claude burns 50% less usage on decks until October 15
- OpenAI cut three safety researchers over an alleged leak
- AWS’s Strands Decider 2B picks from your list, never writes
- Broadcom is now Anthropic’s chip supplier and its lender
- Claude for Government bills agencies with a hard spend cap
- Barclays wants half its developers on Claude Code this year
- Claude Design can draw with your real components
- Claude Code 2.1.286 stops resume from losing your turns
- FTC plans to make Anthropic and OpenAI executives testify
- Gemini 4 Argon wins at office work and trails in coding
- OpenAI’s MCP Extensions put plugins in the ChatGPT sidebar
- Anthropic’s Claude.dev hub ships guides and easter eggs
