Scholar Critiques OpenAI and Anthropic: AI Math Breakthroughs Lack Scientific Rigor

rbhar90 · x · 2026-08-13

Investor Rahul Bharwan compares the current AI breakthrough landscape with past practices. He notes that when DeepMind released AlphaFold, it generously open-sourced a massive database and published detailed papers, allowing the scientific community to collectively test, reproduce, and ultimately "digest" the breakthrough, leading to a Nobel Prize.

In contrast, recent mathematical breakthroughs from OpenAI and Anthropic are causing "indigestion." The companies are blasting out results without contextualizing them or publishing scientific details. Driven by competitive race dynamics, this marketing-first approach limits true scientific impact and prevents the mathematical community from properly verifying and understanding these advances.

Original post →

More from AGI Musings

AGI Musings channel →