Peter Thiel is worth paying some attention to. And it’s not just because he’s a wealthy venture capitalist – his views on pharma and biotech research are worth noting because he’s an excellent example of an intelligent, motivated outsider, someone with a strong technical background who’s approaching drug discovery de novo. Thiel is willing (and very able) to make large investments in technologies that he thinks are worthwhile, so what does he think of drug research?
“Not much” is one possible answer, although those views are, to be sure, more directed at “Big Pharma” than at drug research in general. But a phrase like “Pharma companies are bets against innovation” does stir the blood a bit. This was part of a larger point about how many technology companies end up as “antitechnology” companies, because they’re trying to hold on to their positions, and not have done to them what they probably once did to their predecessors. When I think about that, I can’t help but be reminded of the disappearing Old Bolsheviks in the various Soviet group photos of the 1920s and 1930s – Lenin and Stalin were far from the only political revolutionaries to have decided that, now that the revolution has succeeded, the last thing they needed around the place were people with some experience in overthrowing governments. “Surely we’re past that now, eh, comrades?”
Thiel’s book “Zero to One” had many of us wondering what other paint-filled balloons would be tossed. And although there was some more of that sort of thing (the phrase “high-salaried, unaligned lab drones” comes to mind), you could also see him grappling with the differences between drug research and the types of technology that he was more familiar with. The “Andy Grove fallacy” is always a hazard in such cases – the feeling that the pace of change in computing hardware and software are the natural pace of scientific advancement, which means that any fields that aren’t moving along so briskly must be lazy or complacent. I have no desire to offend anyone by saying this, but the reason drug research moves more slowly than Silicon Valley thinks it should is because it’s harder, both scientifically and practically (see that last link for more).
So now here’s a new interview with Thiel in Technology Review, and you can see that the grappling process is continuing. Asked about his Founders Fund investing in Stemcentrx, he says “I don’t think that we would be seen as particularly sophisticated biotech investors”, this to explain why the company’s high valuation must be a result of their own technological promise rather than the fact that Founders Fund is backing them). But he also says that he invested because “it was a biotech company that looked like a software company,” and that makes you wonder.
It may well be that biotech companies that remind you of software companies are exactly what you don’t want, because it may not even be possible for that hybrid to exist in the real world. It depends on what exactly it is that reminds you of the software business – a willingness to take on big, bold projects is probably good, but a conviction that everyone else is doing it wrong (and so slowly!) is probably not. But another line from the interview is a good sign: “A big difference between biology and software is that software does what it is told, and biology doesn’t.”
That’s exactly right – software is designed by humans, and biology isn’t, and I’m very happy to see that as a feature of Thiel’s thinking. And then there’s this, on biotech startup valuations:
You have to get through basic research, preclinical, Phase I, II, and III, and then marketing. So approaching it analytically, the question is how do you discount [the risk of failure at each step]. If you do half on each step, and there are six steps, that’s 2 to the 6th, or 64. So something worth a billion at the end means you start at [a value of] $16 million.
The thing I don’t like about this as an investor is that the numbers are totally arbitrary. They are just made-up numbers. And our feeling with many biotechs is that people understate these probabilities. They say it’s half, but maybe it’s just one in 10. And if even if just one of these steps is one in 10, you are really screwed.
Well. . .that’s all true. And there are some numbers that can be attached to that – averages, at least. The clinical success rate (across Phase I, II, and III) really is one in ten. Now, those are three of the six steps he’s talking about, and three 50% chances in a row would give you 12.5% success, which is higher than any therapeutic area has shown. So a fifty-fifty chance on each of those is a high estimate, and (in my experience), it’s a high estimate for the two earlier steps as well. Thiel’s back-of-the-envelope calculation isn’t quite into “really screwed” territory, but it does suggest that a starting value of $16 million is very much a lowball figure.
Thus Stemcentrx. Thiel says that “we felt the whole company was designed to get these probabilities as close to one as possible at every step, to get rid of as much of this randomness or contingency as possible. That is something that we found deeply reassuring.” And I’ll bet that they did, but is that reassurance justified? As that last post on the company (and the comments to it) show, the whole cancer-stem-cell idea is not without controversy. Not everyone even believes that there are such things in many (or most?) types of cancer, and their general usefulness as a therapeutic target is solidly in the “unproven” category. Similarly unproven are the best ways to target them, even stipulating that there’s something to target and that it’s a good idea to do so.
And that, folks, is what innovative drug discovery looks like. This is exactly the situation in many another therapeutic area; you find yourself stepping off into the unknown very, very quickly. This interview makes it clear that Thiel believes that Stemcentrx’s approach has significantly cut down the potential risks, and he basically has to believe that, because here’s his take on the alternative:
But if biotech companies tend to invest money in ways that are pseudo random, then a lot of it must get wasted. You end up doing things where you say, “I am not sure it’s going to work.” Well, that sounds like a wasteful thing to do. The standard excuse that biotech companies have is that, “We don’t know if it’s going to work, so we have to do it this way.” That has to be inefficient.
It sure is. I’m just not sure yet how much we can do about it. And the incentives to believe that someone has done something about it are so strong, and so pervasive (and have wiped out so many times in the past) that I hope an outside observe can be forgiven for wondering if things really are different this time.
Update: some thoughts on the same interview from Wavefunction.