FEATURED
06:15 UTC
Reports of OpenAI Codex instability and silent GPT-5.5→5.4 downgrades
release problems outages controversies
medium
Treat Codex and GPT-5.5 xhigh as unstable right now—verify models, sandbox edits, and enforce guardrails before letting agents touch prod.
claude-code
06:16 UTC
Claude Code adds live Artifacts; GitHub Action v1.0 cleans up CI
new feature deep dive
medium
Claude Code now turns in-progress work into live shareable pages and gives you a cleaner, GA GitHub Action to wire AI into CI.
google
06:18 UTC
ARD lands: a common layer for agent tool discovery across enterprise silos
new product launch
medium
ARD solves discovery; you still own auth, lifecycle, and guardrails.
anthropic
06:19 UTC
Anthropic shows Claude Opus 4.7 autonomously solving a robotics integration ~20x faster than last year’s human team
data benchmark study
medium
Agentic LLMs just crossed an autonomy threshold—start testing them on small, verifiable runbooks before they start running whole workflows.
zhipu-ai
06:20 UTC
Open-weight coding models hit a new tier: Kimi K2.7 Code and GLM‑5.2
new product launch
medium
Open‑weight coding models are now good enough to trial as local, private code assistants—run your eval and see if they beat your hosted setup.
perplexity-ai
06:22 UTC
Perplexity Brain puts agent work-memory in production, and research backs the approach
new product launch
medium
Agent-centric memory is moving from research to production—pilot it on repeat workflows and measure cost, accuracy, and explainability gains.
model-context-protocol-mcp
06:25 UTC
Zep Graphiti shows a practical path to real-time agent memory—and a nudge toward portable skills
workflow use case
medium
You can ship faster agent memory now—and keep your skills portable—by pairing Graphiti’s self-hosted graph+vector approach with MCP-style skill definitions.