Fast AI Progress Forces Predictions to Be Tested, Exposing Fashionable Skepticism
jeremiecharris · x · 2026-08-05
The author points out that an underrated virtue of rapid AI progress is that it forces intellectual honesty. When predictions can actually be tested by reality in real-time, it becomes hard to keep arguing based on cope.
In a follow-up tweet, the author lists some high-confidence, now-known-wrong "fashionable AI skepticism" takes that discouraged people from getting ahead of real problems:
- "AI won't discover real cyber exploits"
- "Rogue AI is just sci-fi"
- Arguably, the 2025 "infra bubble" narrative
However, the author jokes that getting wiped out by rogue AIs is probably a bigger deal than having your predictions proven wrong.
More from AGI Musings
- Security Expert Rebuts: Fully Reasonable Tech Stacks Matter More Than Ever in AI Era — craigbalding · 2026-08-05
- Sam Altman at Stanford: AI Agents Will Soon Outwork Human Interns — TansuYegen · 2026-08-05
- Stop Drowning in Tutorials: 10 Must-Read Books for AI Engineers in 2026 — mdancho84 · 2026-08-05
- Google DeepMind and Partners Fund Multi-Agent AI Safety Research with $10M — typewriters · 2026-08-05
- Intelligence Co-founder Envisions a Marketplace for AI Model Capabilities — illscience · 2026-08-05
- Indie Dev Insight: Replace the Perfect Co-Founder with 2-3 AI Agents — yihui_indie · 2026-08-05