CLAUDE
30 days · UTC
Synchronizing with global intelligence nodes...
Open-weight coding models hit a new tier: Kimi K2.7 Code and GLM‑5.2
Two new open‑weight coding models—Kimi K2.7 Code and Zhipu AI’s GLM‑5.2—are emerging as viable local alternatives to hosted code assistants. Reviewer...
Z.ai open-sources GLM-5.2: 1M‑context coding model built for long runs, with cheaper long‑context compute
Z.ai released GLM-5.2, an MIT-licensed 1M-context open-weight coding model aimed at long-horizon, repo-scale engineering tasks. GLM-5.2 pairs a solid...
Anthropic pauses planned Claude Agent SDK subscription change
Anthropic halted a planned subscription change for the Claude Agent SDK on the day it was set to start. Per The New Stack’s report, Anthropic paused ...
Cheap intelligence is here. Build the harness.
LLM compute is getting cheap, but the bottleneck is the harness that turns it into permissioned, auditable decisions. A founder cut model spend 97% b...
AI didn’t cut engineers — the bottleneck is specs, verification, and context
Evidence points to AI agents speeding coding but not replacing engineers, because specs, verification, and context remain the bottleneck. An essay su...
Anthropic pulls Mythos 5 and Fable 5 under U.S. order; build failover now
Anthropic abruptly disabled access to its new Mythos 5 and Fable 5 models after a U.S. national security directive. TechRadar reports Anthropic shut ...
Anthropic proposes thresholded AI safety rules with stop-ship powers—time to harden eval and audit
Anthropic proposed a concrete regulatory framework for advanced AI that would mandate testing, transparency, and stop-ship authority at defined comput...
Anthropic makes Claude Fable 5's hidden throttles explicit after backlash
Anthropic changed how Claude Fable 5 handles sensitive ML prompts by making refusals and fallbacks explicit after developer backlash. Developers foun...
Token economics get real: dev tools now expose per-call costs and enable price-performance routing
Developer tooling is starting to surface true token economics, making AI cost control measurable instead of guesswork. The latest [claude-mem v13.5.x...
Anthropic ships Claude Fable 5: Mythos-class power with guardrails (some silent)
Anthropic released Claude Fable 5, a public Mythos‑class model with strong—and sometimes invisible—safety controls. Fable 5 exposes Mythos‑level capa...
Anthropic details agents that write and run code, pushing toward self-improving AI
Anthropic says it now uses autonomous agents to write and run code in its model development loop. In a new write-up, Anthropic outlines measurable pr...
Expose your catalog as an MCP tool or assistants won’t see your products
ChatGPT and other assistants increasingly shop via tools that query real product catalogs, not by scraping your site. A detailed walkthrough shows ho...
Anthropic says Claude now writes most of its code; Opus 4.8 upgrades make agent loops cheaper and faster
Anthropic reports Claude now authors most of its production code, while Opus 4.8 adds features that harden long‑running, cheaper agent workflows. Ant...
Build Agent-Proof Workflows, Not Agent-Centric Teams
Coding agents are volatile, so design workflows that can survive swapping them out. This [TechBeat brief](https://hackernoon.com/6-3-2026-techbeat?so...
Terminal-Bench 2.0 shows coding agents still stumble on real CLI work
Terminal-Bench 2.0 introduced a tougher CLI benchmark and found frontier agents still score under 65% on real tasks. The new benchmark, highlighted o...
Microsoft makes Copilot agent-native: new code model, desktop app, and terminal integration
Microsoft is turning GitHub Copilot into an agent-native stack with a new in-house code model and system-level integrations. Microsoft introduced MAI...
Claude Opus 4.8 becomes Claude Code’s default, bringing dynamic multi‑agent and long‑running workflows
Anthropic shipped Claude Opus 4.8 and made it the default in Claude Code, adding dynamic workflows that can run long tasks and coordinate many agents....
Local LLM agents are crossing the usability gap — if you own the infra
Open‑weight models hosted with vLLM can run real agentic workloads — but only if you add explicit state, provenance, and robust retrieval. A deep div...
Informatica cracks IDMC into MCP-addressable services as enterprises line up behind agent-ready data ops
Informatica is exposing IDMC data management services through MCP so agents and IDEs can invoke governed data ops directly. Per an [InfoWorld report]...
Cursor turns its IDE agent into headless infra with a public Agents SDK; Composer 2.5 steadies the hands
Cursor turned its IDE agent into headless infrastructure with a public Agents SDK, while Composer 2.5 made the agent steadier on long tasks. Cursor’s...
Anthropic tightens the MCP stack: buys Stainless, adds tunnels/sandboxes, and runtime trust becomes table stakes
Anthropic is pulling agent plumbing closer to Claude with a Stainless acquisition and new MCP security features, while MCP runtime trust takes center ...
Cursor ships Composer 2.5 in-IDE model, nudges teams toward headless agents via its SDK
Cursor released Composer 2.5, a stronger in-IDE coding model that aims to match frontier agents on real repo work. The drop lands only inside the Cur...
CLI coding agents harden: Claude Code stabilizes and resumes long runs; Copilot CLI ships workflow boosts; enterprises eye consolidation
CLI coding agents are maturing fast: Claude Code fixed reliability gaps and added /resume for long runs, while GitHub Copilot CLI shipped notable work...