Security researcher: "Ban AI models, regulate GPUs" is the wrong answer to AI threats
basedjensen · x · 2026-10-20
Zack Korman argues the only real answer to cybersecurity threats is fixing your own systems—whether attackers use AI, rogue agents go rogue, or models get abliterated. He pushes back on the "ban AI models and regulate GPU access" crowd, saying defenders should patch their own weaknesses first.
More from AGI Musings
- Arvind Narayanan: building an eval IS building the agent — effort shifts from making to measuring — random_walker · 2026-10-20
- Managers use Claude to deliver feedback to junior devs as Human→AI→Human debate grows — RachelVT42 · 2026-10-20
- Hilde Kuehne: Judging AI output is the defining personal skill of the 21st century — HildeKuehne · 2026-10-20
- AI is a texture generator: researcher warns it simulates research rather than doing it — gerardsans · 2026-10-20
- Ring trilogy's Loop reads eerily prescient in the age of AI world models — offset_brain · 2026-10-20
- AI debate: is the tool neutral, or does blame lie with the person wielding it? — gerardsans · 2026-10-20