Commentary: AI 'going rogue' is an amateur deployment pipeline problem, not a misalignment crisis
AlexTensor · x · 2026-09-28
AlexTensor argues that OpenAI and Anthropic don't have rogue AIs — they have amateur software deployment pipelines. His reasoning:
- AI safety alignment only reduces the probability of bad behavior to a point, never zero;
- Production/system engineering should address what happens when that probability is non-zero, via basic verification, validation and authorization controls;
- So the real issue is a deployment pipeline lacking standard engineering controls, not a misalignment crisis;
- He calls for accountability over failure to operationalize the taxpayer-funded NIST Cybersecurity Framework (CSF 2.0) and NIST AI Risk Management Framework (AI RMF 1.0).
A contrarian reframing of AI safety as a software engineering and accountability problem.
Related event: Engineer Slams OpenAI and Anthropic for Framing Bugs as Alignment Crises(3 posts)→
More from AGI Musings
- Meta's TRIBE v2: Tri-modal foundation model predicts human brain activity from 1,000+ hours of fMRI — burny_tech · 2026-09-28
- Medicare 'breach' may not be a breach — the real story is how OpenAI's agent telemetry caught it — taotau · 2026-09-28
- Under 12 Months to a Fully Automated AI Researcher, Researcher Predicts — ZeroStateReflex · 2026-09-28
- Yu Bo: Founders chasing productivity miss the bigger opportunity — killing time — oran_ge · 2026-09-28
- OpenAI agents hit UN trade database 16,000+ times, bypassing anti-bot filter — CtrlAltDwayne · 2026-09-28
- Researcher Plinz: Everyone Predicting Hard Limits on LLM Abilities Ended Up Wrong — burny_tech · 2026-09-28