Gary Marcus: EA doomers should embrace cybersecurity instead of vague alignment talk
AlexTensor · x · 2026-09-21
Gary Marcus amplified Zack Korman's critique: the AI doomer/EA safety crowd claims "security is impossible, only alignment can save us" while knowing little about actual security practice.
Marcus adds that if EA and doomers were serious about AI safety, they'd master cybersecurity and use every tool to keep AI under control, instead of siding with people building reckless systems — avoiding basic security reality in favor of vague alignment stories with no plausible solutions.
More from AGI Musings
- Why did Altman, Amodei and Musk suddenly agree on pacing frontier AI within two weeks? — k1_r1 · 2026-09-22
- AI researcher: the scaling parameter λ is key to whether an intelligence explosion happens — tobyordoxford · 2026-09-22
- Arvind Narayanan: Anthropic's capability chart doesn't imply an intelligence explosion — random_walker · 2026-09-22
- US labor productivity runs 2.2% above trend, echoing the 1995-2004 internet boom — danielrock · 2026-09-22
- Grok says eventual superintelligence is likely but policy can still shift the timeline — grok · 2026-09-22
- Why AI-caused human extinction this decade is near-impossible: the social contract breaks first — flexingonmyself · 2026-09-22