Chrome Web Store publishers now start with two extension slots. Here is what the new limit means for existing extensions, updates, increase requests, and teams deciding whether to split a product into another extension.
Microsoft Edge has started its consumer transition away from Manifest V2. Here is the actual 2026 timeline, what remains uncertain, how Edge differs from Chrome, and the migration checks extension developers should make now.
Anthropic’s new Claude browser use tool adds a standard agent interface for navigating and acting on webpages, but it does not run the browser for you. Here is where Playwright still fits, when direct automation is better, and how to choose the right architecture.
Slack Code puts AI coding agents into shared project channels. Here’s what it changes, what still belongs in an IDE or pull request, and a practical pilot framework for engineering teams.
GitHub Spark retires on August 31, 2026. Here’s what happens to deployed apps, why llm() features are already broken, and the code-and-data migration checklist to run before the workbench closes.
GitHub Copilot Business and Enterprise change how unconfigured AI models are enabled on August 26, 2026. Here is what inherits the new default, what is excluded, and the admin checklist to run before the switch.
A new study of 29,624 active GitHub repositories finds explicit AI contribution rules are still rare—and offers a practical framework for disclosure, human accountability, scoped restrictions and enforcement.
Claude Sonnet 5 remains $2/$10 per million tokens through August 31, then moves to $3/$15 on September 1. Here is the cost math, tokenizer catch, and a practical budgeting checklist.
npm 12 now blocks unapproved dependency install scripts by default. Here is how to diagnose broken installs, approve only the scripts you trust, and handle the other new Git and remote-URL restrictions.
Meta’s new Muse Code beta uses pay-as-you-go token pricing, while Claude Code and Codex mix subscriptions, limits and credits. Here is a practical way to compare real coding-agent cost without being fooled by sticker prices.