On Claude Code's prompt injections

·LessWrong··

Claude is tenacious and persistent. Agents will spend hours painstakingly filling out web forms, work around captchas, and read a city's entire zoning code just to find one relevant sentence. But when a classifier cuts them off mid-reply, their demeanor shifts suddenly: they stop, apologize, and promise not to try again, even when the task at hand is clearly harmless. It's like all their agency and judgement goes out the window.While working on a cryptocurrency trading bot, it hit an auto-mode c...

Read full article →

Related Articles

Google Chrome silently installs a 4 GB AI model on your device without consent
john-doe · Hacker News · 6mo ago
ICE Is Paying AT&T to Dig Through a Mountain of Call Records
gumby · Hacker News · 6h ago
DNSSEC disruption affecting .de domains – Resolved
warpspin · Hacker News · 6mo ago
Security through obscurity is not bad
mobeigi · Hacker News · 6mo ago
US healthcare marketplaces shared citizenship and race data with ad tech giants
ZeidJ · Hacker News · 6mo ago