Kyunghyun Cho

Professor of computer science at NYU working on deep learning and natural language processing, and co-author, under his own name, of a 2025 blog post arguing that AI-driven drug discovery is stuck in an early, rule-based era the way machine translation was before its late-1980s statistical turn. Beyond that argument, his own homepage records a parallel industry career: he co-founded the antibody/protein-design startup Prescient Design in early 2021, which Genentech acquired later that year, and served as an Executive Director of Frontier Research and Senior Fellow at Genentech until January 2026 — five years running an industrial drug-discovery lab alongside his academic scaling-laws work.

In this vault Cho first appeared only as a byline on the "Neural Neural Scaling Laws" line of work (co-authorship alone, per the vault's promotion rule, is a mention — not a hub). He earns the hub here on separate standing: his own blog post making an independent, self-credited historical argument, and his own bio naming a role (co-founder, lab lead) rather than just an author credit.

References

• claim-cho-2025-ai-drug-discovery-needs-mt-style-statistical-shift — his Cold War/machine-translation analogy for AI drug discovery, from his own blog, in his own words.
• claim-cho-cofounded-prescient-design-acquired-by-genentech — his co-founding of Prescient Design and subsequent Genentech role, from his own homepage.
• claim-hu-2026-no-independent-corroboration-found-as-of-2026-10-24 — co-author (byline only) of the NeuNeu scaling-laws preprint this note audits.
• claim-lourie-2025-six-category-downstream-scaling-taxonomy — co-author (byline only) of the related Lourie, Hu & Cho taxonomy paper.
• entity-ortet-inc — a second company his own homepage lists him as co-founder/CEO of, as yet unexplained.