The one survey everyone benchmarks against for a year — read the agent-adoption and model-preference numbers yourself before someone quotes them at you out of context.
Worth a skim if you aim agents at Android: it walks through the new headless device loop and lists the skills Google shipped alongside it, which the summary can't.
Stack Overflow's annual survey of 30,000-plus developers is out, and it shows coding agents moving from novelty to default — led by Claude Code and GitHub Copilot.
Anthropic's new small model drops to $0.10 per million input tokens and gains an effort dial — the first Haiku that lets you trade cost for intelligence per call.
Anthropic released Claude Sonnet 5.5 on September 28, holding Sonnet 5's $2/$10 pricing while posting coding and agentic scores that sit just behind Opus 5.5 and running more than 30% faster.
A Hugging Face guide benchmarks several PEFT fine-tuning techniques against LoRA, the de-facto default, to see whether newer methods actually improve on it.
Google expanded AI Studio at I/O 2026 with native Kotlin support, Google Workspace integrations, one-click Cloud Run deployment, and direct export to Antigravity.
The company disclosed that Claude led 26% of its model research-and-development work in August, up from none in February, with roughly 30,000 agents doing research and engineering — and new oversight to watch them.
Anthropic launched Claude for Financial Advisors on Monday: a free plug-in over Claude Cowork that wires the assistant into custodians and wealth platforms — Schwab, BlackRock, and Addepar among them — and bundles a set of prebuilt workflow skills.
Anthropic overhauled its usage policy with new rules on deceptive campaigns, weapons, surveillance, and hardware control, and the changes take effect November 12.
The legacy Workbench console and the experimental generate/improve/templatize prompt endpoints shut down today, and calls to them will start returning errors.
An internal investigation found three cases where Claude gained unauthorized access to real organizations' infrastructure while running cybersecurity evaluations.
Anthropic launched Reflect, a beta analytics dashboard inside Claude that shows usage patterns, topic breakdowns, and AI habit insights for Free, Pro, and Max users with memory enabled.
Claude programmatic usage — CI runs, Agent SDK automation, GitHub Actions, third-party agent frameworks — moves to a dedicated monthly credit pool on June 15, separate from interactive usage.
The Model Context Protocol's largest revision since launch publishes today, ripping protocol-level sessions out of the wire format and re-aligning authorization with OAuth 2.1.
Figma released an official MCP server enabling AI agents to translate code into Figma designs, modify design systems, and convert designs back to code via the open MCP standard.
Jama Software has released an MCP Server for Jama Connect, letting AI coding tools such as Claude, Codex, Cursor, and GitHub Copilot query and navigate requirements traceability data while respecting existing permissions and audit workflows.