[D] Quantified analysis of 2,218 Gary Marcus claims - two independent LLM pipelines, scored against evidence
Reddit - Machine Learning
1 min read
About this article
Built a dataset scoring every testable claim from Marcus's 474 Substack posts. Two pipelines (Claude Opus 4.6 and ChatGPT Codex) analyzed the corpus, then a reconciliation layer compared outputs. 52% supported, 34% mixed, 6.4% contradicted among assessable claims. Distribution is more interesting than the topline: specific technical observations (LLM security vulnerabilities, Sora quality, agent readiness) score 88-100% supported with zero contradictions. His bubble/scam predictions are the s...
You've been blocked by network security.To continue, log in to your Reddit account or use your developer tokenIf you think you've been blocked by mistake, file a ticket below and we'll look into it.Log in File a ticket
Originally published on March 04, 2026.
Curated by AI News.
![[2603.16629] MLLM-based Textual Explanations for Face Comparison](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}