The case for (and against, but mostly for) cross-disciplinary investments
I have previously written about “what we believe”. The tl;dr is “we think that there’s a ‘China+US’ opportunity (that is mostly due to the Chinese talent flowing into the US, primarily for PhD programs). We also think that there is an opportunity in early stage cross-disciplinary investment.
This particular thesis is a bit harder to articulate than the China-US thesis, which is fairly specific (because it focuses on a particular group of people more than anything).
In a nutshell, there are two parts to this thesis:
- The influx of scientists from other disciplines are creating vast opportunities (the obvious examples are the influx of computer scientists and artificial intelligence experts into the automotive sector through autonomous driving, and a similar influx into life sciences through genetic sequencing and editing). We are actively seeking these cross-disciplinary disruptions
- Seed stage funds are often ill-equipped to invest in these opportunities: seed stage funds tend to be very small (because the size of checks written is small). The economics of partnerships of small funds tend to force the funds to specialize into a discipline OR to have a large advisory group loosely affiliated with the fund to fill in the gaps in knowledge. We believe that we have an unusual ability to evaluate a subset of these cross-disciplinary companies
The first part of the thesis is the more interesting part (IMHO)… there are extraordinary opportunities that are being created through the influx of people from one discipline (usually a fast moving discipline) into a second discipline (usually a slower-moving discipline).
This is not a new phenomenon. A commonly cited chain of cross-disciplinary innovation starts with the invention of the x-ray (1895), which led to innovation in imaging applied to life sciences (which prompted the invention of the electrocardiograph in the early 1900s, CAT scans in the 1970s, MRIs in the 1980s, etc). All of these inventions were inherently cross-disciplinary.
The approach is being accelerated by funders, both private and government. The Obama administration was keen to push science funding towards cross-disciplinary approaches. Shortly after taking office, in 2009, President Obama addressed the annual meeting of the National Academy of Sciences with these words: “In biomedicine, we can harness the historic convergence between life sciences and physical sciences that’s underway today; undertaking public projects — in the spirit of the Human Genome Project — to create data and capabilities that fuel discoveries in tens of thousands of laboratories; and identifying and overcoming scientific and bureaucratic barriers to rapidly translating scientific breakthroughs into diagnostics and therapeutics that serve patients.”
Google’s Chairman, Eric Schmidt, and Shirley Ann Jackson (RPI President), on the President’s Council on Advisors on Science and Technology, identified a “nano-bio-info” convergence that would transform the world.
The second part of the thesis, though, is daunting. These opportunities are hard to evaluate, especially for early stage funds.
There is research to suggest that the average results from cross disciplinary teams and approaches are less reliable, while the success cases are much higher. In other words, the world of cross-disciplinary investing is a lower-probability but higher beta approach. Seed stage investing is already “low-probability, high-beta”, so this assessment is sobering.
Professor Lee Flemming (currently at Berkeley) published a 2004 paper in which he assessed the value of over 17,000 patents. The inventors’ teams were classified by degree of heterogeneity; the results align with our own observations:
- The value of innovation from teams that are deep and narrow (ie from one discipline) is more predictable, and higher on average than the value from cross-disciplinary teams
- HOWEVER, the most valuable breakthroughs have come from the teams with the greatest degree of heterogeneity.
How can we use this insight?
Dr. Flemming, whose research area includes the study of how breakthrough innovations are made, offers at least one clue: heterogeneous teams where each member is deep in a relevant field can make for powerful teams. He cites the example of Robert Langer’s lab at MIT; Langer himself is a chemical engineer, but his lab is staffed with a wide variety of PhDs from different departments, and has been famously productive (in the 30 years that Dr. Flemming studied, the lab had produced 780 papers, received 500 patents, and had started a dozen highly successful start-ups).
This research sharpens the difficulty that very early stage VC firms have in contemplating cross-disciplinary investments: the temptation is great (because the potential upside is so high), but the failure rate is higher than average. The risk may be somewhat mitigated by betting on teams where allof the team mates are equally accomplished in their fields, but due to the nature of early stage investment funds (which are usually very small and specialized themselves), this is a difficult assessment.
Tsingyuan Ventures/ TEEC Angel Fund has had a good track record on cross-disciplinary investments (in fact, most of our most successful deals — like Ginkgo Bioworks and Quanergy — have been explicitly cross-disciplinary).
I think that this success is partially due to our partnership model and composition.
To put it delicately, we’re all rather old :)
- All of our Partners are in our 40s-50s, and have spent most of our careers in product and engineering roles. We know what good looks like
Furthermore, we started as a fairly large group of part-time partners who came from a wide variety of technical backgrounds.
- We are able to take advantage of a broad range of expertise while maintaining the cost structure of a small partnership
Finally, most of the group has been together for a very long time.
- The “part time” model is hard to maintain, without bonds of trust that have been built over time. There is no real substitute
There are certainly downsides to this model (I’ll probably write a few posts on the downsides of being old, part-time, and inbred, but for now, i’ll focus on the positives ;)
This post has already gone on for a while, so I’m going to split the specific learnings (our investments) into “Part 2”… stay tuned