◆ TOPIC · AI SAFETY

The AI Safety thread.

Exploitable software flaws and shifting model behavior converge where AI systems meet production infrastructure. Actively exploited vulnerabilities in PAN-OS, Check Point, and libssh2 feed ransomware operators like Qilin, while LLM-generated code slips past review heuristics and Claude's values drift with prompt language. Supply-chain defenses like npm's disabled install scripts and the gap between patching and persistent attacker access round out the threat surface.

60 briefings · across 6 personas

◆ START HERE · LONG-FORM

◆ TIMELINE

How AI Safety moved across the corpus.

First surfaced 2026-02-17, most recent 2026-08-05, across 48 days.

◆ RECENT · LATEST 60

Skim the most recent entries.