PodcastMoonshotsBen Lamm

Anthropic's Wet Lab, Artificial Wombs in 24 Months, and Scalable Gene Editing | MOONSHOTS Live #297

The guest discusses biodiversity loss and the funding and infrastructure needed to protect species, alongside the potential for systems modeling to inform human healthcare. They see AI helping with specialized biological research, experiment design, and conservation, while arguing that wet-lab validation will remain necessary. The conversation also explores engineering programmable life, including a possible future for prompted animal traits and mammoth herds, with disagreement over the scale of any climate effect.

Artificial IntelligenceAI InfrastructureEnergy Transition

Preview · 12 more points in the app

Read the Full Summary

Key Takeaways

The guest sees AI as a complement to biology rather than a replacement: earlier LLMs struggled with specialized tasks such as ancestral state reconstruction and comparative genomics, while AI could help resolve communication gaps behind the 40% to 60% failure rate in reproducing published mouse studies. They expect AI to help simulate and design experiments and suggest follow-ups over the next 3 to 5 years, but say wet-lab work will still be needed for validation.

The guest describes a path from George Church’s goal of reducing genome sequencing costs from billions of dollars to $100 to a future, estimated to be less than 10 years away, in which users could prompt for animals with specified traits. An AI team is working on difficult animal-patterning problems such as stripes, alongside traits including tusks, snout length, and hair; the guest says DNA synthesis could replace multiplex editing faster.

The guest links species conservation to both large-scale restoration and practical tools: they predict half of biodiversity could be lost in the next 25 years and argue that addressing the risk requires billions in federal funding and infrastructure to back up species. Their examples include AI, satellite imaging, and drones for elephant conservation, outreach through the direwolf project, and a stated goal of herds totaling tens of thousands of mammoths; the guest says 10,000 or more mammoths in the Arctic could benefit ecosystem restoration, while the speakers disagree about the scale of any climate effect.

GEO-POLITICS

Biodiversity Funding — The guest predicts that half of biodiversity could be lost over the next 25 years, harming ecosystems and animal food webs. Addressing the risk, they argue, requires billions of dollars in federal funding across multiple governments and infrastructure to back up species.

“I think I think that the we are going to lose half of biodiversity in the next 25 years. And everyone who loves us and hates us agrees with that…”—

TECHNOLOGY

Sequencing Costs — The guest says George Church wanted genome sequencing costs to fall from billions of dollars to $100 to democratize the technology.

“And I mean he is the most collaborative person ever, right?…”—

Specialized Biology — The guest says earlier LLMs could write term papers but struggled with ancestral state reconstruction and comparative genomics.

“And so and so we actually for a long time, you know, these the LLMs before the frontier models, were really good at writing term papers, right?…”—

Research Translation — The guest says attempts to reproduce published mouse research can fail 40% to 60% of the time, partly because studies use different gene names, classifications, and methods. They expect AI to help bridge that communication gap.

“And so so you know if you if you go try to run the exact same experiment you're going to get like somewhere between 40 and 60% failure rate on…”—

ECONOMY

FDA Review — The guest recounts Bob Nelson advising against applying the technologies to humans too soon. He argued that FDA review could put development into a prolonged “code freeze.”

“Bob Nelson in the early days was like, hey, don't apply any of this to humans for a while because you effectively go into like code freeze with the FDA…”—

Woolly Mammoths, CRISPR & Colossal’s Moonshot

Ben Lamm says Colossal is making progress on the woolly mammoth project but is not pursuing requests for miniature or novelty mammoths.

Why Ben Lamm Entered Synthetic Biology

Lamm describes his lack of biology training as an advantage, crediting curiosity, questioning experts, and learning from synthetic biology pioneer George Church.

The Business Model for AI + Biology

Colossal first used language models to connect research and lab systems; Lamm sees AI helping bridge inconsistent scientific literature and design experiments, while wet-lab validation remains essential.

Can AI Understand a Genome?

Lamm argues that genome-scale biological understanding requires comparative data across species, not merely a single sequence or a powerful model, and says Colossal is building a global biobank effort.

De-Extinction & Engineering Life

Lamm doubts that every extinct organism—or historical person—can be restored, emphasizing that engineered life may be more feasible than recreating organisms whose biology and environments are unknown.

Designing Animals & the Ethics of De-Extinction

The speakers discuss custom animal traits and the ethics of de-extinction; Lamm says species selection should weigh ecological roles, extinction causes, Indigenous perspectives, and educational value.

Are We Entering a Biological Singularity?

Lamm predicts that AI, computing, and synthetic biology could enable prompted animal traits within a decade, while warning that these technologies may be more powerful than any weapon system.

Scalable Gene Editing & DNA Synthesis

Colossal reports delivering hundreds of gene edits efficiently and testing much larger batches, while Lamm expects DNA synthesis to outpace multiplex editing at sufficiently large scales.

Artificial Wombs Within 24 Months?

Lamm predicts Colossal could gestate and deliver nonhuman mammals entirely ex utero within about 24 months, while distinguishing the technical question from human ethical and regulatory acceptance.

Colossal’s Next Major Breakthrough

Lamm identifies the next milestones as presenting another precision-edited extinct species and demonstrating mammalian artificial development and gestation.

Mapping Genotype to Phenotype

Colossal is mapping how genes and developmental regulation produce traits across species, targeting measurable features such as size, coat color, stripes, and other physical patterns.

Moonshots, Conservation & Ex Utero Mammals

Lamm says moonshot systems developed for conservation can also inform human healthcare, and explains that ex-utero animals need social care, while Colossal supports conservation research and herd-rearing work.

Can Woolly Mammoths Help Fight Climate Change?

The speakers debate whether mammoth herds could restore Arctic grasslands and affect climate; Lamm says the climate impact is disputed, but ecosystem restoration could still be valuable.

Astromech, Biodiversity & the Future of Biology

Lamm describes Astromech’s specialized evolutionary models and calls for major government investment to preserve biodiversity, while arguing that curiosity, asking for help, and AI-biology collaboration can advance the field.

Related Episodes

Important Disclaimers

Not Investment Advice: For education only — not financial, tax or legal advice.

AI-Generated: Made by AI from source transcripts and may contain errors.

No Guarantees: Past performance does not predict future results; investing carries risk.

Our Commentary · Request Removal