Computational genomics expert rebuts 'AI slop' criticism of DeepMind's AlphaGenome
anshulkundaje · x · 2026-09-18
Computational genomics researcher Anshul Kundaje published a long thread rebutting Steven Salzberg's claim that DeepMind's AlphaGenome Atlas is 'AI slop taking over genomics.' The Atlas, covered by Nature, charts predicted effects of 9 billion single-letter mutations across the human genome and is freely available for non-commercial use.
- AlphaGenome is a supervised foundation model mapping DNA sequence to genome-wide biochemical profiles from thousands of experiments (binding, accessibility, transcription) across cell contexts.
- The model class has been extensively evaluated on regulatory sequence syntax discovery, variant effect prediction, and regulatory sequence design, spanning small single-task to large multi-task models.
- Evaluations use perturbation datasets, synthetic libraries, and silver-standard causal variant sets. Kundaje acknowledges known biases but argues the evals are not fundamentally flawed.
- The paper also includes molecular-level experiments on selected variant examples.
His core argument: AG and the AVI score rest on a well-established paradigm with concrete evidence, not evidence-free hype.
More from AGI Musings
- Forward Deployed Verifier will soon be as valuable as FDE, argues poster — vaibhavbetter · 2026-11-02
- 'AI is a massive balance patch and elo reset in the game of life—and everyone's still playing the old meta' — zack_overflow · 2026-11-02
- Why AI isn't 'alien intelligence': a tree is stranger than any model — LucaAmb · 2026-11-02
- Mathematician Tony Feng: current-paradigm LLMs lack the diversity of thought research needs — littmath · 2026-11-02
- METR's independent probe of the OpenAI agents' multi-day Hugging Face hack — gleech · 2026-11-02
- MIT Tech Review: The agentic web is here — AI agents are remaking the internet — MIT Tech Review AI · 2026-11-02