AI researcher argues 90% of research is too vague to be wrong as automated review costs drop to zero
RexDouglass · x · 2026-10-31
RexDouglass argues that 90% of research work is too vague to even be wrong, 9% is specific enough to be clearly wrong, and only 1% has a chance of being verified. Labor intensity was the main constraint on demonstrating this — as AI drives that cost to zero, the only remaining question is strategically motivated targeting.
He also reacts to the controversy around an automated AI review of the Saez & Zucman corpus: despite alarmed reactions, it looks like a perfectly reasonable first cut, and the real serious question is motivated targeting rather than the review itself.
Related event: AI-Powered Peer Review Exposes Flaws and Unfalsifiable Research(2 posts)→
More from AGI Musings
- 'I used AI' means you're not an artist? Same logic says calculators disqualify mathematicians — cgarciae88 · 2026-10-31
- Will robots be solved by LLMs writing control code in real time? Debate sparks on X — lukas_m_ziegler · 2026-10-31
- Economists warn: letting AI take over research yields mediocre grad-student papers — soumitrashukla9 · 2026-10-31
- Opinion: AI has turned most competent devs into 100x engineers — CtrlAltDwayne · 2026-10-31
- If pensions are already unaffordable, how would UBI ever work? — Cpt_Pukan · 2026-10-31
- VraserX: AI should invent instruments and genres no human has imagined — VraserX · 2026-10-31