Monday, August 16, 2021

Marc Andreessen: An existential crisis in science [Talent search]

Richard Hanania interviews Marc Andreessen, Flying X-Wings into the Death Star: Andreessen on Investing and Tech, August 16, 2021:

On basic research:

That said, something has gone very wrong in the basic research complex that we do have in the universities and in the federal funding system. We know something has gone very wrong because of the replication crisis. John Ioannidis at Stanford wrote this classic paper where he said “50% of published research is non-reproducible” which means you basically can’t reproduce the same results meaning you can’t do anything with the research, it’s basically fake. Interestingly he was studying biomedical research, which is an area of research you think would be very focused on getting things right. It subsequently turned out he might have been underestimating the problem. Fields like biomedicine might be as fake as 70%.

There’s this other amazing study that maybe we can link to for your listeners. It was a study of medical trials funded by a branch of the federal apparatus, heart and lung research. It’s this great chart that shows that new medicines stopped working at the point when the funders of the research require the researchers to register their experiments ahead of time. It’s basically this chart that shows for 30 years or something you had all of these new medicines coming out and the research results were like “wow this really works we should give it to patients,” and then there’s this point where the people running medical trials were forced to pre-register their hypothesis with the funding agency. They were forced to basically say upfront “this is exactly what we’re testing for; here’s exactly how we’re going to measure the results.” Subsequent to that point, new medicines have stopped working.

There’s two possible implications to this, right? The low-hanging fruit argument is that it used to be easier to make medicines that work and now we’ve harvested the low hanging fruit in science and technology, now it’s harder to do it and it’s just a coincidence. The other explanation is that these old medicines didn’t work either. They were basically fake research results based on data mining or p-hacking in such a way that the results were fraudulent. The implication is that there are a lot of drugs on the market today that don’t actually work.

Anyway, science right now is in an existential crisis. This is a real, real issue, and there’s now a generation of scientists who specialize in pointing this out and analyzing it. Andrew Gelman and others. I’ll give you an example: I had a conversation with the long-time head of one of the big federal funding agencies for healthcare research who is also a very accomplished entrepreneur, and I said, “do you really think it’s true that 50-70% of biomedical research is fake?” This is a guy who has spent his life in this world. And he said “oh no, that’s not true at all. It’s 90%.” [Richard laughs]. I was like “holy shit,” I was flabbergasted that it could be 90%.

He’s like “well look, 90% of everything is shit”, which is literally this thing called Sturgeon’s law which says that 90% of everything is bad. 90% of every novel written is bad, 90% of music, 90% of art… 90% of everything is bad. So his analysis was, anything you get in the field of medical experimentation, biomedical development, and this is going to be true of any field, there’s like five labs total in the world that are really good at what they’re doing and doing really cutting edge work. And this is true of quantum computing; pick any field for advanced technology you want.

So those five labs have a pretty good shot at doing interesting work, even some of that is going to reproduce and some of it isn’t. But once you get out of those top five labs, it’s pretty much make-work, incremental, marginal improvements at best, and a complete waste of time otherwise. And I said “good God, why does the other 90% continue to get funded if you know this?” And he said, “well, there are all these universities and professors who have tenure, there are all these journals, there are all these systems and people have been promised lifetime employment.” Anyway, a longwinded way of saying that we have pretty serious structural and incentive problems in the research complex.

Finding talent:

Richard: When Peter Thiel was doing his Thiel fellowship for people, it was seen as this huge eccentricity. That was 10, 20 years ago? You pay people to drop out of school. That’s fascinating that it’s now a better credential to have dropped out of Harvard in Silicon Valley than to have actually finished.

Marc: The Thiel fellowship is super interesting. I’m super open on these things, and even I was like “really? 19-year-olds?” Basically he paid a bunch of 19-year-olds to drop out of college, they all moved to a group house in San Francisco, and he told them all “go nuts.” On the surface you would say “oh my god! What are you doing? Let’s hope that 100% of them survive.” And if you look at what that cohort has done, it’s a very small number of kids… that’s a point Peter makes, it wasn’t that many kids! People really freaked out given the fact that it was 20-year-olds or something.

That tells you how important the social signal and status stuff is. He calls the existing universities the Catholic Church; part of the Reformation, sells indulgences. They know full well what they’re doing… they know they’re corrupt, they know their system is broken. So of course they’re going to disproportionately panic at the indication… they see Martin Luther walking down the street of course they’re going to freak out.

But you look at that cohort of kids, they’ve been successful beyond belief! Ethereum! Ethereum is a $300 billion outcome in ten years; a Thiel Fellow. Sigma, another company we’re invested in is extraordinarily successful… they’re revolutionizing computer design; Thiel Fellow. And there’s another dozen of these that are like these really spectacular breakthrough… these people are just doing incredibly fundamental work. Centered in tech, but of course that’s where people can do breakthrough work these days.

There it is. That worked. It begs this question, “OK, why are we not all doing that? There’s something there.”

Richard: I think Peter Thiel being able to select 20 people is probably the point. [laughs] I went to law school at the University of Chicago. It just showed me you could have so many smart people… and I just felt like law school was for people who didn’t know what to do with their lives. That was me too, I didn’t go to a great college, I didn’t know what to do. That's what they were there for. Selection is everything, so you’re right. Not everybody can found Ethereum, but people can think about starting a hardware store instead of getting some worthless degree, I think that’s realistic for people.

Marc: Selection is a big part of everything, I can see the general point. I will say I’m sure it’s just not 20. I’m sure it’s not 2 million, but I’m sure it’s not 20.

Richard: Well even if it’s 2,000…

Marc: Yeah! Well how about that? It’s not like the Harvard undergraduate class, how many kids per year? Probably a couple thousand freshmen a year. That’s not that big. We load onto those kids the responsibility apparently for determining every aspect of how our society works. We’re making some selection there already. What's the total reporting staff for the New York Times? 800?

Here’s the way we’re thinking about it from a tech standpoint. You can’t just go flat up against the bundle, you can’t do a frontal assault on the death star. Even if you could, and people have tried to full on create new universities, you run up against the accreditation cartel. It’s one of these government regulatory capture things. Universities are a government supported cartel, and quite literally they have access to federal student lending through the accreditation process, the accreditation process is formed by the incumbent. The government has delegated to these institutions the ability to decide who gets to compete against them, and the answer is nobody gets to compete against them. These are nominally non-profits but you would never know it looking at their financial statements.

It is this idealized, stagnant entrenchment of the status quo which is extremely powerful. You’re not going to frontal assault it. What you can do and what we’re doing is you can slowly strip away pieces of it; you can pull pieces out. You can for sure pull out the actual skills training part of it. You can pull out the dating part of it. You can pull out a lot of the logistical components of it. You can pull out housing, you can pull out the food component.

And then you get into the serious stuff. Can you re-credentialize? Can you create new forms of signal and status that are not just based on these legacy institutions? It’s the X-wing death star thing. Any individual effort probably won’t work, but if we run enough experiments over time and strip off enough pieces I think we’ll be able to start having an impact.

There's more at the link.

No comments:

Post a Comment