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

42m 138K views 30 Sep 2026

Is it worth your time?

Worth your time if you want to understand how artificial intelligence and advanced gene editing are reshaping synthetic biology and commercial de-extinction projects.

The conversation explores how artificial intelligence, scalable gene editing, and synthetic biology are accelerating de-extinction efforts and altering the future of conservation. It examines the technological convergence driving these shifts and the practical realities of engineering life outside natural biological limits.

Scalable gene editing and artificial wombs. Lamm argues that the convergence of synthetic biology, compute, and artificial intelligence will create tools far more powerful than any weapon system ever created. He details that while biopharma typically celebrates delivering twenty edits, his team consistently delivers over three hundred edits at greater than ninety percent efficiency and predicts mammalian artificial wombs will birth animals fully ex utero within twenty-four months.

Artificial intelligence in biology. Lamm explains that while frontier models initially struggled with messy biological literature—leading to a forty to sixty percent failure rate when replicating published studies—current systems assist with simulation experiments and laboratory automation. He expects AI to bridge this communication layer by unifying disparate nomenclature, though wet lab experiments remain necessary to validate hypotheses over the next three to five years.

De-extinction and biodiversity conservation. Lamm warns that humanity is set to lose half of all biodiversity in the next twenty-five years, arguing that society must invest in infrastructure to back up species. He explains that Colossal weighs species selection based on environmental contributions, extinction causes, and indigenous spiritual connections, aiming to restore degraded ecosystems like the Siberian tundra with herds of woolly mammoths.

Woolly Mammoths, CRISPR & Colossal’s Moonshot

Ben Lamm recalls backstage anxiety over whether a screen prompt would prematurely leak Colossal's plans. Recalling Elon Musk's request at a previous Abundance 360 conference for a pet woolly mammoth, Lamm notes that teacup mammoths rank as their second most common public request. While joking about engineering new pink melanin for Paris Hilton, he confirms they are not working on miniature versions, though the actual mammoth project is making strong progress.

Why Ben Lamm Entered Synthetic Biology

Entering synthetic biology with a background in gaming, defense, artificial intelligence, and mobile was a massive advantage, according to Ben Lamm. Rather than needing a biology degree, he relies on surrounding himself with the smartest minds, regularly meeting with experts like Beth Shapiro to learn ancient DNA extraction and exploring multiplex editing boundaries. He frames this curiosity-driven approach as a lesson for entrepreneurs, noting that Harvard geneticist George Church calls him his best student due to relentless hours of questioning.

Describing George Church, Lamm details the pioneer of synthetic biology as the head of genetics at Harvard whose lab originated next-generation reading technologies and numerous multi-billion-dollar spinouts. Church combines a prolific startup record with a collaborative, open-source ethos aimed at democratizing genome sequencing by driving costs from billions to one hundred dollars.

The Business Model for AI + Biology

Ben Lamm explains that while frontier models like those from OpenAI and Anthropic are powerful, using them in biology initially required building middleware to connect lab notebooks and Jira—noting that retraining scientists to use Jira proved harder than stem cell reprogramming. Early large language models excelled at writing term papers rather than comparative genomics, but today they assist with running simulation experiments and laboratory automation.

Addressing the challenge of messy literature, Lamm points out that published studies on mice often use different nomenclature or run experiments slightly differently, leading to a forty to sixty percent failure rate when trying to replicate published work. He expects AI to bridge this communication layer by unifying disparate nomenclature, though wet lab experiments remain necessary to test and validate hypotheses over the next three to five years.

Can AI Understand a Genome?

Ben Lamm addresses whether a raw gene sequence fed into Anthropic can yield useful biological insights, arguing that a single genome alone is insufficient. Instead, Colossal is rolling out a Noah's Arc 2.0 model, working with global governments to establish localized bio-repositories and perform deep sequencing. By analyzing thousands of genomes across diverse evolutionary lineages—such as examining p-53 genes and immortal jellyfish—researchers can study species immune to certain diseases and backtrack those traits down the tree of life. Concluding that the data set ultimately proves more valuable than the model itself, Lamm stresses that comparative genomics at scale is what bridges the gap.

De-Extinction & Engineering Life

Ben Lamm discusses de-extinction, noting that while a 19th-century Russian philosopher named Nikolai Fedorov envisioned using technology to revive every human who ever lived, Colossal focuses on species rather than humans. He points out that very few organisms leave a fossil record, citing that fewer than a hundred T-Rexes have ever been found despite billions existing. Because of this, he argues it is more likely that humanity will engineer life to advantage rather than successfully bring back every past organism. Lamm notes that even cloning historical figures like Mozart would not guarantee the same outcome, since environment and epigenetics play massive roles, though genetic predispositions and aptitudes under the right simulated environments can be understood.

Ben Lamm addresses designing animals through AI systems that can design genome sequences to reflect specific phenotypes, comparing public openness toward custom creations like a Pikachu versus human genetic engineering. He confirms that the hesitation stems from social, political, regulatory, and ethical concerns rather than technical ones. Discussing the ethics of de-extinction, Lamm explains that because bringing back everything is not yet plausible, Colossal weighs species selection based on environmental and food web contributions, extinction causes, and indigenous spiritual connections. He notes that factoring in pop-culture appeal, such as bringing back direwolves, helps bridge sci-fi fans into conservation and genome engineering education. Lamm warns that converging synthetic biology, compute, AI, and eventually quantum will create technologies far more powerful than any weapon system ever created, predicting that a complete knowledge base and software for biology where one can prompt and design a desired animal is less than 10 years away.

Ben Lamm outlines the rapid scaling of gene editing, explaining that while biopharma and academia typically celebrate delivering twenty edits, Colossal consistently delivers over three hundred edits at over ninety percent efficiency and is currently testing a thousand. He notes that DNA synthesis and machines generating desired gigabase code will likely replace multiplex editing faster. Addressing artificial wombs, Lamm states that while human applications face societal and regulatory hurdles, Colossal will birth mammalian animals fully ex utero from gestation through delivery without a surrogate within twenty-four months. Looking toward the next major inflection points, he highlights showing the world the next extinct species through precision gene editing, mammalian artificial development and gestation, and unannounced projects in development.

Designing Animals & the Ethics of De-Extinction

Ben Lamm discusses designing animal genomes with artificial intelligence to achieve specific physical traits, noting that people often find the idea of custom animals like a Pikachu more acceptable and approachable than human genetic engineering. He explains that objections to this work are not technical, but rather stem from social, political, regulatory, and ethical concerns.

Addressing the broader philosophy of de-extinction, Lamm references Stewart Brand's famous remark that humans are as gods and might as well start acting like them. He points out that bringing back every species is not yet plausible, requiring a thoughtful selection process. When considering which species to revive, criteria include their historical contribution to the environment and food web, the reasons for their extinction, and the perspectives and spiritual connections of indigenous peoples. For instance, in working on direwolves, his team weighed pop-culture familiarity to help draw fans of fantasy into supporting wolf conservation and learning about genome engineering. He acknowledges that society will likely make mistakes in these choices as technologies proliferate and governments deploy them.

Lamm asserts that the convergence of synthetic biology, compute, artificial intelligence, and eventually quantum technology will create tools far more powerful than any weapon system ever created. He adds that society does not yet fully grasp this shift, though science fiction works like Jurassic Park correctly anticipated the profound weight of these powers.

Looking toward the future, Lamm estimates that within a decade, people will be able to use software to design and prompt custom animals and engineer key phenotypes on base-level organisms, synthesizing them and growing them ex utero, outside the womb.

Scalable Gene Editing & DNA Synthesis

Detailing recent progress in gene delivery, Lamm notes that while most of the biopharma industry and academia celebrate delivering twenty edits, his team consistently achieves over three hundred edits at greater than ninety percent efficiency. They are currently testing a thousand edits, aiming to triple their base edits year over year, though he notes that DNA synthesis will likely replace multiplex editing faster once it reaches a certain scale.

Turning to the development of artificial wombs, Lamm explains that while human applications face complex societal, ethical, and regulatory hurdles, the technological timeline for mammals is much closer. He predicts that within twenty-four months, Colossal will successfully birth mammals fully ex utero from gestation through delivery, without ever entering a surrogate mother.

Asked about upcoming milestones, Lamm points to revealing the next de-extinct species through precision gene editing, achieving mammalian artificial development and gestation, and eventually unveiling several surprise projects currently kept under wraps.

Mapping Genotype to Phenotype

Lamm discusses the mapping of genotype to phenotype, noting the rapid breakthrough of protein folding as a major milestone for biotechnology. When asked if accumulated data will eventually allow researchers to predict precisely what a sequence will produce without needing to build it, he explains that Colossal currently approaches this on a species basis while working to extrapolate across broader clades of animals. He cites sizing as a major area of study, pointing out that transitioning the Tasmanian tiger from a marsupial mouse to its larger form involves a fifteen-hundred-fold scale, and researchers study how developmental timing and gene regulation control these changes.

While scaling an animal beyond certain biological limits is impractical, variations within specific functional ranges are achievable, and groups like dogs and birds demonstrate massive scale functions. Beyond strict de-extinction goals focused on specific genes, the team attempts clade-level extrapolations to understand how sizing might be adjusted by twenty to fifty percent in other species. Lamm notes that traits like coat color, sizing, skin, scales, and feathers will become highly measurable and inducible very quickly.

To tackle patterning and stripes—which present difficult computational challenges alongside tusks, snout length, and hair—the company employs a dedicated artificial intelligence team. Showing the public the first generation of their work in Dallas, Lamm reveals that the initial public display involved eight total gene edits accomplished in a single delivery.

Moonshots, Conservation & Ex Utero Mammals

Ben Lamm responds to a question about whether a moonshot must tackle one massive problem within a single vertical or can operate horizontally across everyday needs, explaining that cross-vertical work is possible because his own attention deficit disorder provides a broader systems-level perspective across multiple categories. He points out that extinction is a systems problem, meaning the same modeling used to preserve and revive species also applies to human healthcare, provided the foundational technology remains applicable to broad use cases rather than becoming too narrow.

Turning to the ethical issues surrounding ex utero mammals and their emotional and social needs, Lamm notes that his organization runs a foundation to open-source all conservation technologies for free, partnering with seventy-five global groups and funding specific welfare projects. Because mammoths and elephants are social creatures, the team plans to return herds rather than isolated individuals, working across sixteen different parallel lines with the ultimate goal of bringing back tens of thousands of Colombian and woolly mammoths. He describes funding Elephant Havens in Botswana to assist orphaned and abandoned elephants, utilizing artificial intelligence and various tools to study how matriarchal societies function and how synthetic herds can be reared. Additionally, the team supports research into satellite imaging, drones, and migratory corridors to understand elephant social dynamics and hierarchies, ensuring that any future ex utero animals grow up within the proper social structure.

George Church previously shared a study at MIT proposing the woolly mammoth as a potential response to global warming.

Can Woolly Mammoths Help Fight Climate Change?

The woolly mammoth's native habitat spans the tundras of Siberia and northern Canada. Historically, these animals roamed and knocked down trees, much like forest elephants in Africa do today. Without trees, grass takes over, and grass sequesters more carbon than trees while offering a greater albedo effect through light reflection back into space.

Lamm explains that high-conviction scientists like George Church calculate that a sufficient mammoth density in the tundra could impact permafrost by lowering summer temperatures six to eight degrees. While other scientists within the broader community and at Colossal debate whether the impact will reach that specific level, Lamm emphasizes that the animals still provide a net positive benefit by helping restore a degraded ecosystem. He notes that while PhDs often disagree on the exact degree of impact, bringing them together to recognize that both sides hold valid points helps reconcile their views, ensuring that regardless of whether mammoths solve carbon climate issues, they deliver positive ecological value to the tundra.

Astromech, Biodiversity & the Future of Biology

Ben Lamm describes his company Astromech as building internal inflection models rather than relying solely on massive foundational models that cannot instantly process the entire tree of life. He explains that these specialized models study how genome sequences evolved, when they evolved, and why they did not evolve in related clades to act as a traffic control cop for global bio-vaults and massive data sets.

Addressing the greatest unaddressed moonshot, Lamm warns that humanity is set to lose half of all biodiversity in the next twenty-five years, noting that zoos and nonprofits alone cannot fix the crisis without billions in coordinated government funding. Arguing that society backs up texts and photos, he insists people must invest in infrastructure to back up species because healthy ecosystems benefit biodiversity, the food web, and human health. He compares this to placental cell banking with LifeBank USA, which stores a newborn baby's stem cells and natural killer cells rather than throwing away the 3D printer that created the child. He notes that Colossal is currently partnering with the US Fish and Wildlife Service and other unannounced governments to back up natural resources through an open-source framework available to any nonprofit, academic group, or foundation.

Asked how to develop a creative mindset for applying technology, Lamm attributes his approach to simple curiosity and a willingness to admit what he does not know, comparing it to childlike wonder. He advises people to ask for help, noting that he often secures meetings with world-class experts simply by cold-emailing them when his teams explain concepts he does not understand, because people generally want to help.

Responding to a question from Kathy Wood regarding Anthropic's announcement of wet labs capable of CRISPR-like creation, Lamm calls the move massively validating for the biotechnology industry. He rejects the idea that AI giants will destroy traditional biotech, arguing instead that health and biology will become one of the most accepted deployments of AI technology. Asked if he worries about existential doom from artificial intelligence, Lamm describes himself as an optimist who believes in human ingenuity, comparing potential scary moments to ordinary turbulence on a flight that still lands safely.

Mentioned

People

Elon Musk Palmer Luckey George Church Bob Nelson Nikolai Fodorov Stuart Brand William Shatner Kathy Wood Alex Wisner-Gross

Companies

Anthropic Colossal Biosciences OpenAI Ginkgo Bioworks Cellularity