The Great AI Slowdown Pivot: Veteran Commenter Calls Labs' Pause Call Pure PR Spin
DavidLinthicum · x · 2026-09-15
Enterprise tech veteran David Linthicum published a scathing long-form critique of this week's coordinated 'slow down for humanity' call from AI labs, calling it the most cynical PR play he's seen in 30 years.
Key arguments:
- No concrete mechanism: None of Amodei, Altman, Hassabis, or Musk lays out an actual failure scenario, timeline, or falsifiable claim — just apocalyptic imagery around 'recursive self-improvement' that resembles faith more than engineering.
- Progress is stalling anyway: Frontier models of the past couple of years offer dazzling demos but underlying capabilities have been roughly flat for two years, with curves driven by scaling budgets, not breakthroughs.
- A convenient narrative: The capital burn is unsustainable and the next model won't deliver the promised leap; calling a pause that was happening regardless lets labs reframe 'our progress stalled' as 'we chose to protect humanity.'
- Entrenching incumbents: The slowdown burdens startups and open source with compliance they can't afford while the biggest players write the rules.
He still takes real risks seriously — misuse, bioweapon uplift, security exploits, data leakage — as engineering problems with engineering solutions.
More from AGI Musings
- 'The intelligence explosion is starting' — optimistic take says curated data keeps models safe — djcows · 2026-09-15
- AI safety researcher Krueger calls for immediate indefinite halt to frontier AI — DavidSKrueger · 2026-09-15
- Critique of CEA's community health team: it balances 'community interest', not victim advocacy — JMannhart · 2026-09-15
- Blogger calls on frontier labs to own product behavior: 'Claude is Anthropic's product, not magic' — tekbog · 2026-09-15
- Researcher: don't slow frontier models' healthcare capabilities, lives at stake — iScienceLuvr · 2026-09-15
- Dev: everyone's too busy discussing doom to remember tech abundance is likelier — djcows · 2026-09-15