Kundaje updates ChromBPNet preprint to refute seq2PRINT's 'better bias correction' claim
anshulkundaje · x · 2026-09-23
Following his public dispute with PRINT/seq2PRINT, Anshul Kundaje announces an updated ChromBPNet preprint (out within days) that he says will 'unequivocally show' that seq2PRINT's claim of better bias correction than ChromBPNet is 'outright false.' He reiterates his team first demonstrated learning biases and in-network bias correction before PRINT's preprint, criticizing the authors for avoiding prior-work context.
More from Fun
- CuiMao, rumored linked to Dario, hires AI-obsessed assistant in Shenzhen — dotey · 2026-09-23
- An 'absolutely awesome mistake': Metaballs via Metal compute shaders, circa 2019 — dreamwieber · 2026-09-23
- One Voice Note in, a Working JEV Prototype Out: AI Agent Builds Demo Mid-Shower — NathanWilbanks_ · 2026-09-23
- Developer: Sometimes You Have to Be a Mean Boss to Get Claude to Do It Right — chrisfirst · 2026-09-23
- Claude Opus 5.5 one-shots a stunning 3D mini-map of the Bay to Sea Ranch — kyleturman · 2026-09-23
- Why Not Hire Opus 5.5? Viral Joke Notes the Role May Not Exist in 12 Months — ordax · 2026-09-23