CHAPTERS
- 0:00 – 0:59
Colossal Biosciences and the idea of de-extinction
Ben Lamm introduces Colossal Biosciences as a de-extinction and species preservation company, prompting Joe to frame it as “real-life Jurassic Park.” Ben explains he didn’t set out to do this—his path came from broad curiosity about synthetic biology and conservation impact.
- •Ben’s role and Colossal’s mission (de-extinction + preservation)
- •Why the concept feels like Jurassic Park—and why that comparison comes up
- •Ben’s background in software and building teams
- •Early framing: conservation and ecosystem restoration as core goals
- 0:59 – 3:40
Meeting George Church and choosing mammoths as the first target
Ben describes how a call with synthetic biology pioneer George Church led him into the mammoth project. After realizing Church’s long-running seriousness about mammoths, Ben visits the lab and they move toward forming a company.
- •George Church’s influence and lab projects
- •Why mammoths became the first focus (existing groundwork, conservation tie-ins)
- •Company launch going viral and expanding ambitions
- •How other de-extinction researchers (thylacine, dodo) connected with Colossal
- 3:40 – 4:36
The mammoth workflow: ancient DNA, AI assembly, and mapping to Asian elephants
Ben outlines the practical steps: obtain many ancient DNA samples, assemble degraded fragments, and computationally align them to a living reference genome. The mammoth’s closest living relative is the Asian elephant, which is ~99.6% similar, driving the “edit the difference” approach.
- •Ancient DNA is highly degraded; needs many samples and heavy computation
- •Colossal’s mammoth sample set and age range
- •Why Asian elephants are the closest relatives (and closer than African elephants)
- •Reconstruction as an AI/compute problem more than a manual one
- 4:36 – 9:51
Mammoth history, extinction timing, and what Asian elephants reveal
They dig into mammoth timelines (surprisingly recent extinction), locations of last populations, and extinction theories like bottlenecks and water scarcity. Joe and Ben compare elephant species, discuss physical traits, and share Joe’s experience interacting with elephants in Thailand.
- •Last mammoths ~4,000 years ago (Wrangel Island)
- •Theories: inbreeding bottlenecks and resource/freshwater constraints
- •Anatomical comparisons (ears, skull shape, fur traits)
- •Elephant intelligence, herd dynamics, and ethical discomfort with riding elephants
- 9:51 – 15:40
Field expeditions, museum “catacombs,” and hidden genetic treasure troves
Ben describes the “Indiana Jones” side: collecting specimens from permafrost and navigating museum basements full of undocumented samples. He shares stories of discovering key thylacine specimens (a preserved pup and a skull ‘head in a bucket’) that enabled near-complete genome reconstruction.
- •Permafrost retrieval realities (smell, conditions) and the John Reeves boneyard story
- •Museums hold huge backlogs; inventories can be incomplete
- •‘Miracle pup’ thylacine sample enabling ~98% genome quickly
- •Head-in-a-bucket skull pushing genome completeness to ~99.9%
- 15:40 – 19:36
Thylacine myths, searches, and why sightings don’t convince the scientists
Joe asks about modern thylacine sightings; Ben explains Andrew Pask’s long-running testing and why most evidence turns out to be other animals. They touch on the thylacine’s official extinction date and why a surviving population would dramatically simplify de-extinction work.
- •Andrew Pask’s 15-year effort and repeated false leads (e.g., dog samples)
- •Camera traps and field sampling across Tasmania
- •Official extinction ~1936; possible later remnant into the 40s/50s
- •Discussion of claims in remote regions (e.g., Papua New Guinea)
- 19:36 – 21:49
Synthetic biology fundamentals: editing traits instead of chasing 100% genomes
Ben explains that de-extinction doesn’t require a perfect ancient genome; it requires identifying and engineering key phenotype-driving differences. He describes precise nucleotide-level edits, larger DNA block synthesis, and how AI reduces the number of edits needed to recreate traits like coat, ears, fat layer, and skull features.
- •Why 100% genome recovery isn’t necessary for functional outcomes
- •CRISPR as one tool; newer methods enable precise letter-level changes
- •Targeting phenotype regions (hair, fat, ears, tusks, cranial shape)
- •Species definitions and ‘functional de-extinction’ framing begins to emerge
- 21:49 – 29:48
The woolly mouse: proof-of-concept, ethics, and viral public reaction
They reveal Colossal’s woolly mouse project—engineered mice expressing mammoth-like coat traits—as a rapid testing model versus long elephant gestation. Ben details multiplex editing, monoclonal screening to avoid off-target effects, humane oversight, and why the mice won’t be sold or widely distributed despite intense public interest.
- •Mouse model enables fast phenotype validation (20-day gestation)
- •Using mouse-equivalent genes rather than inserting mammoth DNA directly
- •Multiplex editing (many genes at once) and sequencing-based screening
- •Humane certification, controlled breeding (sex separation), and public display requests
- 29:48 – 33:12
From U.S. conservation work to the dire wolf decision
Ben explains how collaboration with the Mandan, Hidatsa, and Arikara (MHA) Nation on bison genomics expanded into wolf conservation priorities and cultural history around “the great wolf.” This, plus the red wolf crisis and public fascination, pushed Colossal toward an American flagship project: the dire wolf.
- •Population genomics for bison and indigenous partnership goals
- •MHA Nation cultural heritage prompting the dire wolf idea
- •Red wolf collapse as a conservation emergency
- •Strategic rationale: connect conservation with public imagination and education
- 33:12 – 35:57
Recovering dire wolf DNA: caves, petrous bones, and high-coverage genomes
Joe asks how dire wolf de-extinction is even possible without permafrost specimens. Ben explains the limitations of La Brea tar pit specimens, then describes how a tooth and an older skull (via dense petrous bone sampling) enabled ~13–14X coverage—enough to model and engineer a viable genome plan.
- •Why tar pit specimens aren’t usable for high-quality DNA
- •Early dire wolf genome work was low coverage (~15%) from a tooth
- •Petrous bone as the best DNA reservoir in skulls
- •Achieving 13–14X coverage to support de-extinction-quality reconstruction
- 35:57 – 44:43
Rewilding ethics, wolves as ecosystem engineers, and the Colorado controversy
The conversation pivots to the morality and logistics of reintroductions, using Yellowstone’s wolf success as a trophic cascade example. Joe critiques Colorado’s voter-driven wolf reintroduction and the ranching conflict that follows; Ben agrees on the need for science-led, measured rewilding with monitoring and mitigation.
- •Humans already ‘play God’ via ecosystem impacts; restoration is a counterbalance
- •Yellowstone wolf reintroduction and trophic cascades (elk, willows, beavers, rivers)
- •Colorado reintroduction critique: politics, borders, and livestock losses
- •Need for monitoring, expert-led decisions, and compensation frameworks
- 44:43 – 1:03:45
The reveal: three living dire wolves and how they’re being raised
Joe pushes for the headline: Colossal has produced three dire wolves (Romulus, Remus, and Khaleesi). They discuss size, appearance, unexpected mane-like fur, behavioral instincts, secure housing on a large preserve, veterinary protocols, and why the location is kept secret due to security and black-market risks.
- •Dire wolf pups and names; investor connections (pop culture tie-ins)
- •Rapid growth metrics (45+ lbs at ~3 months; ~80 lbs at ~5 months)
- •Distinct traits: stockier build, thick fur, mane-like features, white coloration
- •2,000-acre secure preserve, animal hospital, controlled exposure and care
- 1:03:45 – 1:11:19
How a dire wolf is made: gene mapping, edits, cloning, and the ‘species’ debate
Ben details the technical approach: compare dire wolf genomes to gray wolves, edit key regions tied to size and craniofacial traits, then use somatic cell nuclear transfer (cloning) into surrogates. They then tackle the philosophical pushback—what counts as a species, and whether functional equivalence is what matters.
- •Genome comparison to gray wolves; identifying trait-driving differences
- •Multipoint edits, then somatic cell nuclear transfer into embryos/surrogates
- •Debates over species definitions (breeding, morphology, geography)
- •Jurassic Park analogy: engineered organisms vs ‘true’ originals; functional de-extinction
- 1:11:19 – 1:17:59
Red wolves and genetic rescue: cloning from blood and government resistance
Ben introduces Hope, the first cloned red wolf, and explains Colossal’s approach to species preservation alongside de-extinction. He describes a non-invasive cloning method using endothelial progenitor cells from blood, frustration with bureaucratic inertia, and a more supportive meeting with Interior leadership about scaling biobanking and recovery programs.
- •Hope the red wolf; multiple clones from distinct genetic lines
- •Open-source conservation tools and dozens of conservation partners
- •Non-invasive cloning-from-blood method and why it’s a biobanking game-changer
- •Government friction vs support: feasibility-study bureaucracy contrasted with Interior enthusiasm
- 1:17:59 – 1:33:58
Scaling risks and boundaries: artificial wombs, ‘where do you stop,’ and human implications
Joe asks the bigger question—extinction is the norm, so how far should de-extinction go? Ben emphasizes prioritizing human-caused extinctions and keystone species, then discusses artificial wombs as a way to scale endangered species recovery while avoiding surrogate burdens, acknowledging the unsettling implications if such tech moves toward humans.
- •Selection criteria: critically endangered + keystone species + human-caused loss
- •Artificial wombs as a scaling tool for conservation (e.g., rhinos)
- •Colossal’s stated boundary: not working on humans directly
- •Ethical concerns: development, bonding, and unknown effects of ex utero gestation
- 1:33:58 – 1:57:53
Thylacine rewilding planning, predator balance, and ‘genetic memory’ field evidence
Ben explains how Colossal builds stakeholder working groups (including industry) before reintroductions, using Tasmania as a case study. He shares a striking field experiment: animals in Tasmania react far more strongly to thylacine silhouettes than to cats/dogs, suggesting deep, inherited predator recognition that may ease reintroduction concerns.
- •Stakeholder planning: governments, landowners, indigenous groups, even logging commissions
- •Predator/prey modeling and ecosystem impact assessment ahead of any release
- •Camera trap silhouette experiment indicating strong thylacine-specific fear response
- •Discussion of inherited fears (humans’ spider/snake phobias as analogy)
- 1:57:53 – 2:57:34
Tasmanian devils, transmissible cancer, and conservation via engineered resistance
They discuss Tasmanian devils’ behavior and the facial tumor disease, arguing predators like thylacines may have historically removed sick individuals. Ben then gives a concrete conservation engineering example: identifying a single-letter genetic change that confers massive cane-toad toxin resistance, first proven in a model marsupial and planned for quoll rescue to combat invasive species damage.
- •Tasmanian devil feeding behavior and disease transmission via biting
- •Predators’ ecological role: removing sick/weak individuals to stabilize populations
- •Cane toads as an invasive catastrophe in Australia; quolls dying from toxin exposure
- •Single nucleotide change yielding ~5,000x resistance; pathway from model species to quolls
