Claude Code user reports 50M tokens burned on a Markdown check
A Reddit user says they asked Claude Code to check Markdown files for consistency and it consumed roughly fifty million tokens in seconds. The post questions what the agent was actually doing during the run.
1 source
More stories today
1 in 5 MCP access policies found broken or missing
An audit of Model Context Protocol deployments found roughly 20% of access policies were broken or absent, per The New Stack. The piece ties the gap to vibe-coded internal tools wired into Slack and internal APIs without token scrutiny.
The New Stack·13 minutes ago
AI agents are flooding public services with new requests
UK housing ombudsman complaints more than doubled since ChatGPT, from 2,600 in 2022 to just over 7,000 last year, while the US CFPB saw 5x growth. Researcher Chris Schmitz tracked 84 potential flooding cases across 11 jurisdictions, calling it "agentic flooding."
TechCrunch·20 minutes ago
Mollick: steering long-running agents is key to Codex success
Ethan Mollick·21 minutes agoFable 5.1 vs Fable 5: real-world budget benchmark results
Anthropic's Claude Fable 5.1 scores 52.6% on Terminal-Bench-Science, the benchmark it centered its launch around, where a model gets a terminal and a real scientific research problem to solve independently. The New Stack compares Fable 5.1 and Fable 5 on a real-world budget rather than the spec sheet.
HN user reports OpenAI re-enabling 'allow training' setting
A Hacker News poster says the ChatGPT data-training toggle re-enabled itself after being turned off more than once, including after a carefully noted reset date. The post urges users to re-check the setting if they believe it is currently off.
Hacker News·1 hour agoEssay argues AI-accelerated AI research won't trigger rapid RSI
Nathan Lambert·1 hour agoAI Engineer designer uses agent to audit 140-logo sponsor wall
Vincent Wendy, senior creative designer at AI Engineer, pointed an agent at the conference lobby's sponsor banner carrying more than 140 logos and asked it which were missing; it caught every omission in his checks. He repeated the trick on the conference shirt.
YouTube·1 hour ago
tldraw's Max Drake on building agents into the canvas
Talk covers tldraw's spatial harness for agents, including a colleague who turned the whiteboard app into a window manager and played Pong with real desktop windows. Drake is a product engineer at the London company behind the infinite canvas SDK.
YouTube·1 hour ago
