Episode Summary
Executive Summary: The conversation centers on Colossal’s vision for de-extinction, synthetic biology, and AI-driven biology. Ben Lamm argues that outsider entrepreneurship, large datasets, and AI-assisted tooling can accelerate breakthroughs in species restoration, conservation, and eventually mammalian gestation ex utero. He frames the work as both scientifically ambitious and ethically complex, emphasizing conservation, open-source access, and governance over human applications.
Main Topics: De-extinction as a platform company (Priority: 5/5): Lamm positions Colossal not just as a mammoth project, but as a systems company for restoring extinct species, conserving living ones, and building reusable biological technologies. AI as infrastructure for biology (Priority: 5/5): He argues that current foundation models are most useful as communication, workflow, and experiment-design layers in biology, especially for literature mining and data organization, rather than as immediate end-to-end genome solvers. Data over model in genomics (Priority: 5/5): The discussion emphasizes that large comparative genomic datasets, especially across clades and species, are more valuable than a single clever model for predicting genotype-to-phenotype relationships. Ethics, governance, and species selection (Priority: 4/5): The speakers debate how to decide which species should be brought back, weighing ecological role, cause of extinction, Indigenous views, educational value, and societal acceptability. Ex utero mammalian development (Priority: 5/5): Lamm claims Colossal expects to birth mammals fully ex utero within roughly two years, presenting artificial gestation as a major technical inflection point. Conservation and climate applications (Priority: 4/5): The mammoth project is discussed as potentially benefiting Arctic ecosystems and climate, while also advancing elephant conservation, orphan care, and wildlife monitoring today. Entrepreneurial mindset and curiosity (Priority: 3/5): Lamm credits his success to curiosity, asking good questions, surrounding himself with experts, and being willing to enter fields outside his original expertise.
Key Arguments: Being an outsider to synthetic biology is an advantage because it enables curiosity, better questions, and access to world-class experts without preconceived limits. AI will first help biology by bridging nomenclature, literature, and workflow gaps; wet-lab validation will still remain necessary. A single genome sequence is not enough to infer a phenotype; large-scale comparative datasets across species and evolutionary lineages are essential. The most plausible near-term use of synthetic biology is not reviving every historical organism, but engineering life and approximating extinct traits. Ethical decisions about de-extinction should consider ecological function, food-web role, cause of extinction, and Indigenous perspectives. Open-source conservation technologies can benefit existing species today while also building the infrastructure for future de-extinction. Species restoration may have broader environmental benefits, including ecosystem repair and potential climate effects, even if precise causal claims remain debated. The long-term opportunity includes designing organisms by phenotype, scaling traits like coat color, size, feathers, stripes, and other developmental patterns. Human applications should be governed separately from conservation and animal work, with different investors, rules, and regulatory pathways. Colossal’s work is framed as a response to biodiversity loss, which Lamm says will otherwise remove a large share of species within decades.
Data Points: Company valuation: Over $10 billion - Described as Colossal’s current valuation after starting from synthetic-biology outsider status. Founding track record: Four successful companies before Colossal - Referenced as prior entrepreneurial background across gaming, defense, AI, and mobile. Published-work failure rate: 40% to 60% failure rate - Lamm says reproducing published mouse experiments often fails because of nomenclature and methodological differences. Current multiplex edit delivery: 300+ edits at 90%+ efficiency - Used as evidence of rapid progress in gene-editing throughput. Testing scale: 1,000 edit deliveries - He says the team is currently testing thousand-level edit delivery experiments. First woolly mice edits: 8 base edits - The publicly shown first generation of woolly mice required eight edits in one delivery. Large-scale genome comparison: 1,500x fold difference - Used to describe the developmental distance between a marsupial mouse and a marsupial wolf in Tasmanian tiger work. Time to ex utero mammalian birth: Within 24 months - Lamm says Colossal will likely birth mammals fully ex utero within two years. Time to broader biological design capability: Less than 10 years - He says the world of prompting desired animal phenotypes via biology is likely under a decade away. Projected biodiversity loss: Half of biodiversity in 25 years - Lamm claims humanity could lose about half of biodiversity in the next 25 years. Global partnerships: 75 global partners - He says the company has a broad conservation partnership network. Mammoth lines in development: 16 different lines - He says multiple mammoth-related lines are being worked on simultaneously.
Pivotal Quotes: "I think that the data set is more valuable than the model." — Ben Lamm: On why comparative genomics data matters more than any single AI model for biology. "We are as gods, we might as well start acting like that." — Stuart Brand (referenced by speaker): Used in the ethics discussion about whether humans should revive species. "We will, Colossal, will birth mammals fully ex utero." — Ben Lamm: His forecast for artificial gestation and ex utero mammalian development within 24 months.
Implications: The episode suggests de-extinction and conservation are converging with AI, making biology more programmable. Near-term winners will be teams with large datasets, wet-lab capacity, and strong ethics/governance, not just powerful models.