<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>devtake.dev — #anthropic</title><description>Articles tagged anthropic on devtake.dev.</description><link>https://devtake.dev/</link><language>en-us</language><item><title>Days after opening Fable 5 to the public, a US government order forced Anthropic to pull it</title><link>https://devtake.dev/article/anthropic-fable-mythos-government-suspension/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-fable-mythos-government-suspension/</guid><description>A Commerce Department export directive forced Anthropic to disable Fable 5 and Mythos 5 for all users, days after opening Fable 5 to the public.</description><pubDate>Sat, 13 Jun 2026 14:30:00 GMT</pubDate><category>ai</category><category>ai-models</category><category>anthropic</category><category>claude</category><category>claude-mythos</category><category>ai-security</category><category>policy</category><author>dieter-morelli</author></item><item><title>Claude Fable 5 is Anthropic&apos;s first public Mythos-class model. It tops SWE-Bench Pro at 80.3%.</title><link>https://devtake.dev/article/claude-fable-5-launch/</link><guid isPermaLink="true">https://devtake.dev/article/claude-fable-5-launch/</guid><description>Claude Fable 5 hits 80.3% on SWE-Bench Pro and ships on Bedrock and Copilot at $10/$50 per million tokens, free on paid plans only through June 22.</description><pubDate>Tue, 09 Jun 2026 18:55:00 GMT</pubDate><category>ai</category><category>ai-models</category><category>anthropic</category><category>claude</category><category>claude-mythos</category><category>benchmarks</category><category>llm</category><category>agentic-coding</category><author>dieter-morelli</author></item><item><title>Anthropic is sending Mythos 5, the model it called too dangerous, to cyberdefenders and the US government</title><link>https://devtake.dev/article/claude-mythos-5-cyberdefenders/</link><guid isPermaLink="true">https://devtake.dev/article/claude-mythos-5-cyberdefenders/</guid><description>Mythos 5 is the same model as Fable 5 with cyber safeguards lifted, going to Project Glasswing defenders and, Anthropic says, ~150 orgs across 15+ countries.</description><pubDate>Tue, 09 Jun 2026 18:40:00 GMT</pubDate><category>policy</category><category>anthropic</category><category>claude-mythos</category><category>project-glasswing</category><category>policy</category><category>national-security</category><category>ai-safety</category><category>ai-security</category><category>cybersecurity</category><author>clara-wexler</author></item><item><title>Claude Opus 4.8 flags the bugs it writes four times more often than Opus 4.7</title><link>https://devtake.dev/article/claude-opus-4-8-launch/</link><guid isPermaLink="true">https://devtake.dev/article/claude-opus-4-8-launch/</guid><description>Anthropic&apos;s Opus 4.8 posts 69.2% on SWE-Bench Pro, lets code flaws slip 4x less often, and ships parallel subagents in Claude Code. Here&apos;s what matters.</description><pubDate>Fri, 29 May 2026 07:20:00 GMT</pubDate><category>ai</category><category>ai-models</category><category>anthropic</category><category>claude</category><category>llm</category><category>benchmarks</category><category>agentic-coding</category><category>claude-code</category><category>opus-4-7</category><author>dieter-morelli</author></item><item><title>Uber blew its entire 2026 AI coding budget in four months. Its COO can&apos;t prove it paid off.</title><link>https://devtake.dev/article/uber-ai-budget-burn-claude-code/</link><guid isPermaLink="true">https://devtake.dev/article/uber-ai-budget-burn-claude-code/</guid><description>Uber exhausted its full-year Claude Code budget by April. Adoption hit 84%, heavy users burn $2,000 a month, and COO Andrew Macdonald can&apos;t connect the spend to shipped features.</description><pubDate>Wed, 27 May 2026 19:05:00 GMT</pubDate><category>ai</category><category>ai</category><category>claude-code</category><category>anthropic</category><category>ai-coding</category><category>dev-tools</category><category>uber</category><category>ai-agents</category><author>dieter-morelli</author></item><item><title>DeepSeek locked in the 75% V4-Pro cut. The API now undercuts every Western frontier model.</title><link>https://devtake.dev/article/deepseek-v4-pro-price-cut-permanent/</link><guid isPermaLink="true">https://devtake.dev/article/deepseek-v4-pro-price-cut-permanent/</guid><description>On May 23 DeepSeek told customers the V4-Pro discount becomes its standard price after May 31. Output drops from $3.48 to $0.87 per million tokens.</description><pubDate>Sun, 24 May 2026 10:30:00 GMT</pubDate><category>ai</category><category>deepseek</category><category>ai-models</category><category>llm</category><category>anthropic</category><category>openai</category><category>gemini</category><category>ai-chips</category><category>china</category><author>dieter-morelli</author></item><item><title>Microsoft is canceling Claude Code for its engineers. They have until June 30 to switch to Copilot CLI.</title><link>https://devtake.dev/article/microsoft-cancels-claude-code-licenses/</link><guid isPermaLink="true">https://devtake.dev/article/microsoft-cancels-claude-code-licenses/</guid><description>Internal Claude Code licenses end June 30, 2026, for Microsoft&apos;s Experiences + Devices group. Engineers move to GitHub Copilot CLI instead.</description><pubDate>Sat, 23 May 2026 10:00:00 GMT</pubDate><category>ai</category><category>microsoft</category><category>anthropic</category><category>claude-code</category><category>github-copilot</category><category>copilot-cli</category><category>ai-assistant</category><category>agentic-coding</category><author>dieter-morelli</author></item><item><title>Anthropic&apos;s Glasswing logged 10,000 vulnerabilities in a month. Most are still waiting on a patch.</title><link>https://devtake.dev/article/anthropic-glasswing-deception-monitor/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-glasswing-deception-monitor/</guid><description>Anthropic says Project Glasswing&apos;s first month produced over 10,000 critical-and-high-severity vulns. Verification and patching is the limiting step.</description><pubDate>Sat, 23 May 2026 09:45:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>claude-mythos</category><category>project-glasswing</category><category>security</category><category>ai-security</category><category>supply-chain</category><category>vulnerability-disclosure</category><author>dieter-morelli</author></item><item><title>Andrej Karpathy joined Anthropic. The OpenAI founding member&apos;s job: use Claude to train Claude.</title><link>https://devtake.dev/article/karpathy-joins-anthropic-pretraining/</link><guid isPermaLink="true">https://devtake.dev/article/karpathy-joins-anthropic-pretraining/</guid><description>Karpathy started this week at Anthropic on Nick Joseph&apos;s pre-training team. His mandate is using Claude to accelerate Claude&apos;s own training.</description><pubDate>Thu, 21 May 2026 12:00:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>openai</category><category>claude</category><category>andrej-karpathy</category><category>ai-models</category><category>llm</category><category>pre-training</category><category>ai-talent</category><author>dieter-morelli</author></item><item><title>SpaceX&apos;s S-1 revealed who&apos;s paying for Colossus. Anthropic just locked in $45B through 2029.</title><link>https://devtake.dev/article/anthropic-spacex-colossus-15-billion/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-spacex-colossus-15-billion/</guid><description>Anthropic is paying SpaceX $1.25 billion a month for Colossus 1 and 2 capacity. The contract runs through May 2029 and books about 83% of SpaceX&apos;s revenue.</description><pubDate>Thu, 21 May 2026 11:45:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>spacex</category><category>colossus</category><category>ai-infrastructure</category><category>nvidia</category><category>gb200</category><category>data-centers</category><category>claude</category><author>dieter-morelli</author></item><item><title>A bad command-line parser turned every claude-cli:// link into a remote shell</title><link>https://devtake.dev/article/claude-code-rce-deeplink-cve/</link><guid isPermaLink="true">https://devtake.dev/article/claude-code-rce-deeplink-cve/</guid><description>Joernchen of 0day.click found a deeplink RCE in Claude Code. Anthropic shipped the fix in 2.1.118 the same week.</description><pubDate>Wed, 20 May 2026 09:15:00 GMT</pubDate><category>security</category><category>security</category><category>anthropic</category><category>claude-code</category><category>rce</category><category>ai-security</category><category>supply-chain</category><category>ai-agents</category><category>dev-tools</category><author>luca-reinhardt</author></item><item><title>Anthropic bought Stainless, the startup that builds every official SDK for OpenAI and Google.</title><link>https://devtake.dev/article/anthropic-acquires-stainless-sdk-toolchain/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-acquires-stainless-sdk-toolchain/</guid><description>Anthropic announced May 18 it acquired SDK generator Stainless, reportedly for over $300M. The same toolchain still powers OpenAI&apos;s, Google&apos;s, and Cloudflare&apos;s official clients.</description><pubDate>Tue, 19 May 2026 09:45:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>stainless</category><category>sdk</category><category>mcp</category><category>claude</category><category>api</category><category>agentic-coding</category><category>dev-tools</category><author>dieter-morelli</author></item><item><title>Bun&apos;s million-line Rust rewrite is now mainline. 99.8% of tests pass and 13,000 unsafe blocks remain.</title><link>https://devtake.dev/article/bun-rust-rewrite-merged/</link><guid isPermaLink="true">https://devtake.dev/article/bun-rust-rewrite-merged/</guid><description>Jarred Sumner merged the Bun-in-Rust PR on May 14, ending Zig as Bun&apos;s runtime language. Binary shrinks 3-8 MB; one analysis counted 13,000 unsafe blocks.</description><pubDate>Fri, 15 May 2026 09:15:00 GMT</pubDate><category>open-source</category><category>bun</category><category>rust</category><category>zig</category><category>javascript</category><category>runtime</category><category>anthropic</category><category>claude-code</category><category>agentic-coding</category><author>soren-vanek</author></item><item><title>Anthropic shipped Claude for Small Business with 15 prebuilt agents. Daniela Amodei is pitching the corner-store owner.</title><link>https://devtake.dev/article/anthropic-claude-small-business/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-claude-small-business/</guid><description>Anthropic announced Claude for Small Business on May 13 with QuickBooks, HubSpot, Canva, and DocuSign hooks. The pitch: 15 ready-to-run agents and a 10-city tour.</description><pubDate>Thu, 14 May 2026 09:30:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>claude</category><category>claude-cowork</category><category>ai-agents</category><category>small-business</category><category>agents</category><category>quickbooks</category><category>paypal</category><author>dieter-morelli</author></item><item><title>Jarred Sumner rewrote 960,000 lines of Bun from Zig to Rust in six days. He might throw it all away.</title><link>https://devtake.dev/article/bun-rust-rewrite-zig-test-compatibility/</link><guid isPermaLink="true">https://devtake.dev/article/bun-rust-rewrite-zig-test-compatibility/</guid><description>Bun&apos;s creator used Claude to port the JavaScript runtime from Zig to Rust, hitting 99.8% test compatibility. He says there&apos;s a &apos;very high chance&apos; it gets scrapped.</description><pubDate>Sun, 10 May 2026 09:45:00 GMT</pubDate><category>open-source</category><category>bun</category><category>rust</category><category>zig</category><category>javascript</category><category>runtime</category><category>anthropic</category><category>open-source</category><category>dev-tools</category><author>soren-vanek</author></item><item><title>Chinese proxy networks sell Claude API access at 90% off. They harvest every prompt that passes through.</title><link>https://devtake.dev/article/chinese-grey-market-claude-api-stolen-credentials/</link><guid isPermaLink="true">https://devtake.dev/article/chinese-grey-market-claude-api-stolen-credentials/</guid><description>A ChinaTalk investigation reveals how &apos;transfer stations&apos; resell Anthropic API access using stolen credentials, model substitution, and prompt harvesting.</description><pubDate>Sun, 10 May 2026 09:30:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>claude</category><category>ai-security</category><category>credential-theft</category><category>china</category><category>supply-chain</category><category>ai-models</category><author>dieter-morelli</author></item><item><title>Apple is turning iOS 27 into an AI model marketplace. ChatGPT loses its exclusive slot.</title><link>https://devtake.dev/article/apple-ios-27-rival-ai-models-choice/</link><guid isPermaLink="true">https://devtake.dev/article/apple-ios-27-rival-ai-models-choice/</guid><description>Bloomberg reports Apple will let users choose Gemini, Claude, or ChatGPT across Siri, Writing Tools, and Image Playground via a new Extensions framework this fall.</description><pubDate>Sat, 09 May 2026 07:45:00 GMT</pubDate><category>apple</category><category>apple</category><category>apple-intelligence</category><category>siri</category><category>ios-27</category><category>openai</category><category>google</category><category>gemini</category><category>anthropic</category><author>naomi-park</author></item><item><title>Anthropic doubled Claude Code&apos;s limits by renting 220,000 GPUs from xAI</title><link>https://devtake.dev/article/claude-code-rate-limits-doubled-may-2026/</link><guid isPermaLink="true">https://devtake.dev/article/claude-code-rate-limits-doubled-may-2026/</guid><description>Anthropic doubled Claude Code&apos;s 5-hour limits, killed peak-hours throttling, and raised Opus API tiers. The capacity comes from xAI&apos;s Colossus 1, via a SpaceX deal.</description><pubDate>Thu, 07 May 2026 12:30:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>claude-code</category><category>claude-opus</category><category>spacex</category><category>xai</category><category>ai-infrastructure</category><category>ai-models</category><category>colossus</category><author>dieter-morelli</author></item><item><title>GitHub Copilot&apos;s Claude Opus multiplier jumps to 27x on June 1. Monthly plans dodge the hike.</title><link>https://devtake.dev/article/github-copilot-multiplier-hike-june-2026/</link><guid isPermaLink="true">https://devtake.dev/article/github-copilot-multiplier-hike-june-2026/</guid><description>GitHub&apos;s new model multiplier table for Copilot Pro and Pro+ annual plans lands June 1. Opus 4.6 goes 3 to 27. Sonnet 4.6 goes 1 to 9.</description><pubDate>Mon, 04 May 2026 10:45:00 GMT</pubDate><category>ai</category><category>github-copilot</category><category>github</category><category>ai-agents</category><category>pricing</category><category>microsoft</category><category>anthropic</category><category>claude-opus</category><category>dev-tools</category><author>dieter-morelli</author></item><item><title>Anthropic is fielding offers at a $900B valuation. The round closes in two weeks and tops OpenAI.</title><link>https://devtake.dev/article/anthropic-900b-valuation-50b-round/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-900b-valuation-50b-round/</guid><description>Preemptive bids put Anthropic at $850B-$900B with a $50B raise. Run rate hit $30B in March, up from $9B at year-end 2025.</description><pubDate>Sat, 02 May 2026 09:45:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>claude</category><category>funding</category><category>valuation</category><category>openai</category><category>ai-models</category><category>finance</category><author>dieter-morelli</author></item><item><title>Anthropic just dropped its Claude Code workshop tapes. The playbook is better than the marketing.</title><link>https://devtake.dev/article/anthropic-claude-code-workshop-playbook/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-claude-code-workshop-playbook/</guid><description>Boris Cherny on Claude Code, Applied AI on prompting, Erik Schluntz on vibe coding in prod. Three Code with Claude tapes hit YouTube ahead of the 2026 conference.</description><pubDate>Thu, 30 Apr 2026 11:30:00 GMT</pubDate><category>ai</category><category>claude-code</category><category>anthropic</category><category>claude</category><category>ai-assistant</category><category>agents</category><category>mcp</category><category>ai-agents</category><category>code-with-claude</category><author>dieter-morelli</author></item><item><title>Disney built an AI leaderboard. One employee called Claude 460,000 times in nine days.</title><link>https://devtake.dev/article/disney-claude-cursor-tokenmaxxing-dashboard/</link><guid isPermaLink="true">https://devtake.dev/article/disney-claude-cursor-tokenmaxxing-dashboard/</guid><description>Leaked internal Disney screenshots show 4,800 product and tech staff burning 3.1 billion Claude tokens and 13.3 billion Cursor tokens across nine April workdays.</description><pubDate>Tue, 28 Apr 2026 16:00:00 GMT</pubDate><category>ai</category><category>disney</category><category>claude</category><category>cursor</category><category>anthropic</category><category>ai-agents</category><category>enterprise-ai</category><category>tokenmaxxing</category><category>ai-coding</category><author>dieter-morelli</author></item><item><title>GitHub Copilot kills premium requests on June 1. Token billing arrives, fallback models do not.</title><link>https://devtake.dev/article/github-copilot-usage-based-billing/</link><guid isPermaLink="true">https://devtake.dev/article/github-copilot-usage-based-billing/</guid><description>On June 1 every Copilot plan switches to GitHub AI Credits priced per token. Code completions stay free. Fallback models and credit rollover do not.</description><pubDate>Tue, 28 Apr 2026 11:00:00 GMT</pubDate><category>ai</category><category>github-copilot</category><category>github</category><category>ai-agents</category><category>pricing</category><category>microsoft</category><category>anthropic</category><category>claude-opus</category><category>dev-tools</category><author>dieter-morelli</author></item><item><title>Microsoft and OpenAI just rewrote their deal. Exclusivity is dead, and so is the AGI clause.</title><link>https://devtake.dev/article/microsoft-openai-deal-revenue-share-end/</link><guid isPermaLink="true">https://devtake.dev/article/microsoft-openai-deal-revenue-share-end/</guid><description>Microsoft loses exclusive rights to OpenAI&apos;s models. The revenue share now caps at 2030 and stops depending on AGI. Here&apos;s what actually changed and who it benefits.</description><pubDate>Mon, 27 Apr 2026 19:00:00 GMT</pubDate><category>ai</category><category>openai</category><category>microsoft</category><category>ai-models</category><category>azure</category><category>llm</category><category>ai-infrastructure</category><category>anthropic</category><category>gpt-5-5</category><author>dieter-morelli</author></item><item><title>A Discord group guessed Anthropic&apos;s URL pattern and walked into Claude Mythos</title><link>https://devtake.dev/article/anthropic-mythos-breach-discord/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-mythos-breach-discord/</guid><description>Bloomberg reports a small group accessed Anthropic&apos;s locked-down Mythos model the same day it launched, using credentials from a third-party contractor and educated URL guessing.</description><pubDate>Sat, 25 Apr 2026 11:00:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>claude-mythos</category><category>ai-security</category><category>supply-chain</category><category>project-glasswing</category><category>mercor</category><category>security</category><author>dieter-morelli</author></item><item><title>Google is putting up to $40B into Anthropic. That&apos;s five days after Amazon&apos;s $5B.</title><link>https://devtake.dev/article/google-anthropic-40b-investment/</link><guid isPermaLink="true">https://devtake.dev/article/google-anthropic-40b-investment/</guid><description>Google committed $10B upfront and up to $40B total at a $350B valuation, plus five gigawatts of Google Cloud capacity. It&apos;s Anthropic&apos;s second nine-figure deal in a week.</description><pubDate>Fri, 24 Apr 2026 20:30:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>google</category><category>alphabet</category><category>claude</category><category>ai-infrastructure</category><category>ai-models</category><category>funding</category><category>tpu</category><author>dieter-morelli</author></item><item><title>Anthropic admits three Claude Code bugs quietly tanked quality for six weeks</title><link>https://devtake.dev/article/anthropic-claude-code-quality-postmortem/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-claude-code-quality-postmortem/</guid><description>Anthropic&apos;s April 23 postmortem names three bugs that degraded Claude Code between March 4 and April 20. Usage limits are being reset for every subscriber.</description><pubDate>Fri, 24 Apr 2026 11:30:00 GMT</pubDate><category>ai</category><category>claude-code</category><category>anthropic</category><category>claude</category><category>opus-4-7</category><category>ai-agents</category><category>ai-models</category><category>postmortem</category><category>sonnet-4-6</category><author>dieter-morelli</author></item><item><title>Mozilla fixed 271 Firefox bugs that Claude Mythos found. Its own tests caught 22.</title><link>https://devtake.dev/article/mozilla-firefox-mythos-bug-hunt/</link><guid isPermaLink="true">https://devtake.dev/article/mozilla-firefox-mythos-bug-hunt/</guid><description>Firefox 150 shipped Monday with 271 security fixes from Anthropic&apos;s Project Glasswing. Mozilla CTO Bobby Holley says Mythos matches elite human researchers.</description><pubDate>Wed, 22 Apr 2026 13:00:00 GMT</pubDate><category>open-source</category><category>mozilla</category><category>firefox</category><category>anthropic</category><category>claude-mythos</category><category>project-glasswing</category><category>security</category><category>ai-security</category><author>soren-vanek</author></item><item><title>GitHub Copilot paused new signups and kicked Opus out of Pro. Here&apos;s what actually changed.</title><link>https://devtake.dev/article/github-copilot-pro-plan-changes/</link><guid isPermaLink="true">https://devtake.dev/article/github-copilot-pro-plan-changes/</guid><description>GitHub froze Copilot Pro/Pro+/Student signups on April 20 and moved Claude Opus 4.7 behind the $39 Pro+ tier. Agent workflows broke the old math.</description><pubDate>Wed, 22 Apr 2026 11:30:00 GMT</pubDate><category>ai</category><category>github-copilot</category><category>anthropic</category><category>microsoft</category><category>claude-opus</category><category>ai-agents</category><category>pricing</category><category>dev-tools</category><author>dieter-morelli</author></item><item><title>Amazon puts another $5B into Anthropic. Anthropic promises $100B back to AWS.</title><link>https://devtake.dev/article/anthropic-amazon-5b-100b-cloud-deal/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-amazon-5b-100b-cloud-deal/</guid><description>Amazon added $5B (up to $20B) to its Anthropic stake. Anthropic committed $100B+ to AWS over 10 years and 5 GW of Trainium capacity.</description><pubDate>Tue, 21 Apr 2026 10:00:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>amazon</category><category>aws</category><category>trainium</category><category>claude</category><category>dario-amodei</category><category>ai-infrastructure</category><author>dieter-morelli</author></item><item><title>NSA is running Anthropic&apos;s Mythos. The Pentagon says Anthropic is a supply-chain risk.</title><link>https://devtake.dev/article/nsa-anthropic-mythos-blacklist/</link><guid isPermaLink="true">https://devtake.dev/article/nsa-anthropic-mythos-blacklist/</guid><description>Axios reports the NSA is using Anthropic&apos;s unreleased Mythos model even though the Defense Department has blacklisted Anthropic. One government, two positions.</description><pubDate>Mon, 20 Apr 2026 15:30:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>claude-mythos</category><category>nsa</category><category>pentagon</category><category>cybersecurity</category><category>project-glasswing</category><category>national-security</category><category>dario-amodei</category><author>dieter-morelli</author></item><item><title>Anthropic shipped Claude Design. Figma stock dropped 7% the same day.</title><link>https://devtake.dev/article/anthropic-claude-design-figma-rival/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-claude-design-figma-rival/</guid><description>Anthropic launched Claude Design on April 17, a prompt-to-prototype tool that exports to Canva, not Figma. Figma&apos;s stock closed down 7% on the same day.</description><pubDate>Sun, 19 Apr 2026 10:00:00 GMT</pubDate><category>ai</category><category>anthropic</category><category>claude-design</category><category>figma</category><category>canva</category><category>design-tools</category><category>opus-4-7</category><category>anthropic-labs</category><author>dieter-morelli</author></item><item><title>OpenAI&apos;s Codex now drives your Mac, not just your code</title><link>https://devtake.dev/article/openai-codex-desktop-agent-expansion/</link><guid isPermaLink="true">https://devtake.dev/article/openai-codex-desktop-agent-expansion/</guid><description>OpenAI shipped a Codex update that can pilot desktop apps with a cursor, generate images in-line, and run parallel agents. It&apos;s the opening move in a real Claude Code fight.</description><pubDate>Fri, 17 Apr 2026 10:30:00 GMT</pubDate><category>ai</category><category>openai</category><category>codex</category><category>ai-agents</category><category>mac</category><category>claude-code</category><category>anthropic</category><category>atlas</category><author>dieter-morelli</author></item><item><title>Claude Opus 4.7 is here, and the long-context benchmarks got worse</title><link>https://devtake.dev/article/anthropic-claude-opus-4-7-launch/</link><guid isPermaLink="true">https://devtake.dev/article/anthropic-claude-opus-4-7-launch/</guid><description>Anthropic&apos;s Opus 4.7 is state-of-the-art on SWE-bench and CursorBench, but independent tests show regressions on long-context retrieval and thematic reasoning.</description><pubDate>Fri, 17 Apr 2026 09:30:00 GMT</pubDate><category>ai</category><category>claude</category><category>anthropic</category><category>opus-4-7</category><category>llm</category><category>benchmarks</category><category>mythos</category><category>ai-models</category><author>dieter-morelli</author></item><item><title>Claude Code Routines: what they actually do, and when to use them over GitHub Actions</title><link>https://devtake.dev/article/claude-code-routines-explained/</link><guid isPermaLink="true">https://devtake.dev/article/claude-code-routines-explained/</guid><description>Anthropic just shipped Routines: Claude Code sessions as cron jobs, webhooks, and GitHub-event reactors. Here&apos;s what they replace, what they don&apos;t, and one rule to follow.</description><pubDate>Wed, 15 Apr 2026 13:00:00 GMT</pubDate><category>ai</category><category>claude-code</category><category>anthropic</category><category>automation</category><category>agents</category><category>cron</category><category>github-actions</category><author>dieter-morelli</author></item></channel></rss>