DeepMind Researcher Clarifies Controversial "LLMs Can't Jump" Paper
CsabaSzepesvari · x · 2026-07-30
Following the widespread debate sparked by his paper "LLMs Can't Jump," DeepMind researcher Csaba Szepesvari issued a clarification to address misinterpretations.
He emphasized that some have framed the research as "DeepMind throwing cold water on AI for science" or arguing that "LLMs can never make real scientific discoveries." He clarified that this is a personal position paper aimed at exploring the limitations of LLMs at specific points in time and under certain methodologies, rather than outright denying their future potential for scientific breakthroughs.
More from AGI Musings
- AI Math Proofs: Is Lean the Ultimate Arbiter or Just Another Black Box? — RexDouglass · 2026-07-30
- Future 100T Param Model to Cost >$250B to Train, Says Joseph Jacks — JosephJacks_ · 2026-07-30
- AI to Disrupt Math in 5 Years? Researcher Warns Against Underestimating LLMs — prof_g · 2026-07-30
- AI Accelerates Discovery: Vulnerability Findings Surge, Algorithmic Efficiency Unchanged — soumitrashukla9 · 2026-07-30
- CoreAutoAI's Jerry Tworek: Inference-Time Compute is the New Scaling Law — agihouse_org · 2026-07-30
- Prediction Market: OpenAI Leads Odds for #1 AI Model by Year-End, Meta Trails at 12% — Polymarket · 2026-07-30