Oct 18, 2017 · 24m · 20vc
20VC: Lux Capital's Scientist-In-Residence on Why We Cannot Just Be Specialists Today, The Benefits of Interdisciplinary Thinking & Computational Creativity That Makes Man and Machine Partners with Sam Arbesman
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The 20VC, Harry Stebbings interviews Sam Arbesman, Scientist-in-Residence at Lux Capital, about the value of radical interdisciplinarity, human-AI collaboration, and the convergence of biology and computer science in deep tech investing.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 33.3% of the talking time here. How this is scored →
Harry as informed peer 2.7 Guest teaching 4.0 Guest disagreement 1.6 Harry pushing back 2.4
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 19:30 Rejecting the cleantech comparison
Sam directly dissents from Elad Gil's premise raised by Harry, arguing that frontier tech is far too broad and domain-differentiated to be lumped into a single cleantech-style bubble narrative.
Hardest push from Harry ▶ 9:30 Challenging interdisciplinary focusHarry pushes back on Sam's interdisciplinary framework by invoking the foundational venture capital belief that successful startups must focus monomaniacally.
Biggest teaching moment ▶ 17:27 Explaining biological non-linearitySam dismantles the common software engineering assumption that DNA is simple code, explaining how non-linear biological complexity frustrates naive tech approaches.
Harry holds his own ▶ 19:30 Citing Elad Gil on deep tech hypeHarry demonstrates domain expertise by citing prominent angel investor Elad Gil's warning about frontier tech hype to test Sam's confidence in VC readiness.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Sam Arbesman's Background and Role at Lux Capital | 1 | 3 | 0 | 0 | Harry invites Sam to share his background and how he became scientist in residence at Lux Capital. Sam gives a detailed overview of his trajectory from computational biology to complexity science and his role as connective tissue at Lux. | |
| Defining Radical Interdisciplinarity and Its Competitive Advantages | 2 | 4 | 0 | 1 | Harry prompts Sam on radical interdisciplinarity in startups. Sam explains how combining disparate fields creates competitive moats, using examples like computational gastronomy and tax code machine learning. | |
| Structuring Teams with Generalists and T-Shaped Individuals | 4 | 4 | 3 | 5 | Harry asks if interdisciplinary thinking contradicts the classic VC requirement for monomaniacal focus. Sam reframes the argument, noting that mission focus can coexist with interdisciplinary execution without turning a founder into a dilettante. | |
| Computational Creativity and Human-Machine Collaboration | 2 | 4 | 1 | 2 | Harry brings up popular anxieties regarding AI and human capability. Sam frames AI as an essential collaborative tool for exploring complex systems that exceed human cognitive bandwidth. | |
| Pace of Innovation and Technological Augmentation | 3 | 4 | 1 | 2 | Harry asks if society can truly appreciate AI-assisted creative work given the rapid pace of change. Sam introduces the Yiddish term 'nachos' and references screensavers to explain vicarious pride in human-programmed generative output. | |
| Parallels Between Biology, Computer Science, and AI | 3 | 5 | 2 | 3 | Harry asks about the dangers of using computer science analogies in biological domains. Sam explains why treating biology as simple code fails due to its massive non-linear complexity. | |
| Venture Capital Readiness for Deep Tech and Frontier Science | 4 | 4 | 4 | 4 | Harry references investor Elad Gil's warning that deep tech resembles the 2000s cleantech bubble. Sam dissents, arguing frontier tech is too broad and domain-specific to be painted with such a broad brush before transitioning into quickfire questions. |
Statements from this episode (12)
Arbesman: Combining disparate ideas creates a defensible moat for startups
“So I think if a startup can bring together disparate ideas and carve out a niche and unique way of thinking and provide something of value to the world, that's something that is both differentiable and also a defensible moat because it's hard to jump across di…”
Arbesman: Combining specialists, T-shaped people, and generalists builds optimal organizations
“Bring all those different people together, specialists, teachers, individuals, people who are a little bit more general minded into an organization is going to make it the best version possible.”
Arbesman: The startup ecosystem is far less siloed than academia
“The startup ecosystem is certainly in a much better shape than for example, like academia, where academia is incredibly siloed.”
Arbesman: Computational tools are necessary to navigate increasingly complex systems
“As the world around us is becoming more and more complex, it's becoming more difficult to fully understand either the technologies that we ourselves have built, or the systems that we're trying to understand in science, like a biological system, for example. A…”
Arbesman: Machines will increasingly join human teams in creative collaboration
“But ultimately, I think creativity, it's always been, to a certain degree, a collaborative act, and now we're going to increasingly count machines as part Of our collaborative teams now.”
Arbesman: Human advances will require AI partnerships due to cognitive limits
“And so we're going to need these kind of machine partners in order to continue making advances, and so rather than viewing it as humans are being displaced, it's really essentially that humans are only through this partnership are humans going to be able to co…”
Arbesman: Humans will feel vicarious pride in AI-generated creations
“We are going to increasingly need to have a certain amount of nachos for our technologies, where these technologies might be doing things, might be doing activities that they might be making creations that are not our own, but we were the ones who ultimately p…”
Arbesman: Data can now be stored directly within DNA sequences
“We're now able to store data in DNA, which has a certain similarity to computation.”
Arbesman: Modularity is a core hallmark shared by biology and technology
“Modularity is a hallmark of both biology and technology.”
Arbesman: Computer malware encoded in DNA successfully compromised a sequencing machine
“We've actually seen in the news recently this that malware was encoded in DNA and then it compromised and took over a machine involved in analyzing the sequences which is pretty wild stuff.”
Arbesman: VCs can succeed in deep tech by avoiding naive cross-domain analogies
“Yes, I do think so. As long as we're really just not stepping too naively from one domain to another, and we recognize essentially both the promise and perils of analogizing from one complex system to another. So as long as we do this with open eyes, I think w…”
Arbesman: Frontier tech is too broad to be treated as a single category
“While there might be certain similarities, I think frontier tech overall, and it's really this, it's very broad and diffuse space, and so in many ways I'm almost hesitant to group it all under kind of that umbrella term. And because the knowledge you need to d…”