The Millennium Problems for Biology
A new privately curated list of “Millennium Problems for Biology” proposes grand experimental challenges such as origin-of-life synthesis, whole-body cryopreservation, limb regeneration, and custom protein design, partly as benchmarks for AI-driven research. Commenters question the scientific framing and rigor of several items—especially protein amplification and reverse translation—arguing that the list mixes feasible engineering goals with conceptually confused or already-solved problems and lacks endorsement from leading biologists. The thread also touches on deeper issues: the gap between AI labs and domain scientists, the tension between fundamental understanding and commercializable breakthroughs, and whether biology’s most important open questions are even amenable to this kind of prize-style formulation.
Overall perception of the “Millennium Problems for Biology”
- Many see the list as marketing/PR or “pet projects” rather than a field-defining agenda.
- Comparison to the Clay math Millennium Problems is frequent; commenters note those were set by top experts and had clear, time-tested difficulty.
- The name “Millennium Problems” is viewed as borrowed prestige; several say this is “just a list,” not on par with the math version.
- Some think the problems are at least transformative if solved and like that they’re experimentally verifiable and benchmarkable.
Role of AI labs vs scientists
- Skepticism that AI companies or LLM teams can crack deep biological problems without close integration with domain scientists.
- Concern that AI labs are chasing “trophy kills” for IPO narratives rather than building understanding.
- High compute cost is cited as a major barrier for academic use; more results may appear if compute gets cheaper or is donated.
- In math, people note AI-driven proofs risk yielding results without the human-understandable insights the community values.
Critiques of specific challenges
- Protein amplification and “reverse translatase” are heavily criticized as chemically naive or close to impossible without complementarity or nucleic-acid intermediates.
- The “protein PCR” analogy is called “not even wrong” because proteins lack complementary strands.
- Rubisco improvement is debated: one side says evolution has already hit hard limits; another says AGI could surpass nature or bypass Rubisco entirely.
- Cryopreservation of mice is said by some to have been partially achieved in the 1950s, though scaling beyond small animals is claimed to be physically problematic.
Engineering focus vs fundamental biology
- Many note the list is almost entirely engineering-style “build X” problems.
- Some argue that’s appropriate: experimental biology is largely about “how did nature build this,” and verifiable builds are useful.
- Others counter that grand questions include non-engineering problems: e.g., tumor immune “hot vs cold” patterns, chemiosmotic theory, or substrates of consciousness.
Missing or alternative “grand problems”
- Commenters propose: aging reversal, Alzheimer’s, morphogenesis (beyond limb regeneration), protein folding (beyond hype), small-molecule biocatalysis, chronic disease mechanisms, better disease models, and AI/human alignment.
- Origins of life is widely agreed to be unsolved; classic experiments (like Miller–Urey) produced building blocks, not full self-replicating cells or genetic encoding systems.