Stanford professor: AI-guided DNA synthesis hype is over the top, but tech is solid
anshulkundaje · x · 2026-09-26
Stanford professor Anshul Kundaje critiques an AI-driven DNA synthesis company: blog claims are over the top and the writing esoteric, but the core tech is compelling — models guiding stochastic DNA synthesis to build massive structured random libraries, with compressed sensing plus seq2function models learning from pooled screens in active learning loops. He sees real value for scaled data generation and model training in some well-fitting applications.
Related event: Stanford Professor Weighs In on AI-Driven DNA Synthesis Startup(2 posts)→
More from Research
- Stanford's Noah Goodman uses philosophy to improve LLM pretraining, jokes ASI achieved — xuanalogue · 2026-09-26
- Researchers surface spurious probes across models: Sonnet 5 recommends green tea in evals, oolong in production — jankulveit · 2026-09-26
- Three papers, one warning: 1% synthetic data can trigger strong model collapse — suchenzang · 2026-09-26
- Simulation beats distillation: the real story of synthetic data is post-training worlds — realsohamparekh · 2026-09-26
- Experts Rise Where LLMs Disagree: rationale labeling cuts codebook revision from months to days — windx0303 · 2026-09-26
- Dev hails continual learning paper: AGI defined in 2000, only now is anyone training for it — willcb · 2026-09-26