#33 - Is software eating the world, or is the Star Trek computer not enough?

AlphaFold; Peter Thiel vs. Marc Andreessen

Google’s DeepMind did something extraordinary this week. From its announcement (emphasis mine):

Proteins are essential to life, supporting practically all its functions. They are large complex molecules, made up of chains of amino acids, and what a protein does largely depends on its unique 3D structure. Figuring out what shapes proteins fold into is known as the “protein folding problem”, and has stood as a grand challenge in biology for the past 50 years. In a major scientific advance, the latest version of our AI system AlphaFold has been recognised as a solution to this grand challenge by the organisers of the biennial Critical Assessment of protein Structure Prediction (CASP). This breakthrough demonstrates the impact AI can have on scientific discovery and its potential to dramatically accelerate progress in some of the most fundamental fields that explain and shape our world.

What stands out to me about AlphaFold, one of the most important innovations in biology over the past few decades, is that it wasn’t developed or discovered by biologists in a traditional wet lab. At its core, AlphaFold is software.

Is innovation dead? redux

Earlier this year, I wrote a post titled “Is innovation dead?” It’s time to re-visit.

In 1971, President Richard Nixon declared the “war on cancer.” Congress, in its excitement, declared this war “a national crusade to be accomplished by 1976,” a fitting celebration for our great nation’s Bicentennial. Well, the Bicentennial came and went, and almost fifty years later, where’s the cure? Scientists keep saying that the breakthrough is right around the corner, that we’re so close, we just need to wait a few more years. But let’s not mince words: They’ve been saying this for a while now. “Close” doesn’t quite cut it. At bottom, we don’t yet have a cure for cancer.

Thielian stagnation

Peter Thiel cites the failed war on cancer as evidence of technological stagnation. He quipped nearly a decade ago that “we wanted flying cars. Instead we got 140 characters.” Indeed, innovation in the world of bits (i.e., software) has proceeded at a rapid clip while innovation in the world of atoms (i.e., more physical things like cancer cures) has not. Here’s Thiel talking at the National Conservatism Conference:

Maybe we’ve built the Star Trek computer, but we don’t have anything else from the Star Trek universe. We certainly don’t have warp drive. We don’t even have the Concorde anymore. We don’t have the matter transport technology that can rearrange matter and create cornucopia. Instead, we’ve experienced a few decades of relative stagnation, where for the first time in the history of the United States, the younger generation is finding it quite a struggle to live up to the standards of their parents, the baby boomers, and the people who came before them. This is a shocking, shocking disconnect.

In other of Thiel’s talks and writing, he cites two broad categories of explanations for innovation stagnation. The first, which he doesn’t buy, is that we’ve picked all the low-hanging fruit. That is, good ideas are getting harder to find, incremental innovations are increasingly expensive, and there’s nothing we can do about it. The second, which he does buy, is that our institutions and culture have failed us. While the world of bits is relatively de-regulated, other engineering fields, like mechanical, aerospace, and biomedical engineering, are regulated to death, thereby also discouraging talented people from entering those fields. Venture capitalists began investing less in moonshot ideas and more in cynical, incrementalist technologies. Young smart people in a previous age may have become entrepreneurs, but today, they prefer instead to walk well-trodden paths. Society is stuck in a malaise where politics prevents us from getting anything done, and our lack of ambition / imagination prevents us from even trying to get anything done in the first place.

Peter Thiel’s thesis is that we can and must break out of our stagnation lull. We can’t simply rely on innovation in the world of bits to bring us to the expansive future we had envisioned back in the 1960s and 70s. We need innovation in the world of atoms, too.

Andreessen: Software is eating the world

If Peter Thiel is an innovation stagnationist, Marc Andreessen is an information technology evangelist. In his seminal blog post, Why Software is Eating the World, Andreessen makes the case for software being the industry to transform all other industries, including those in the world of atoms.

Let’s look, for instance, at transportation. Thiel is quick to point out that airplane speeds have actually gone down over the past few decades; that transportation is heavily regulated; that the Golden Gate Bridge took 3 years to build in 1933 but that it’s taken the last 10 years to build a small access road to the bridge; that we don’t have flying cars; and that we haven’t transitioned to more sustainable forms of transportation.

Andreessen responds by saying that software has obviated the need for huge changes in the physical world of transportation. Video-conferencing software like Zoom means we don’t need to travel in the first place. Peer-to-peer traffic networking apps like Waze enable drivers to re-balance their routes in real-time, thereby optimizing traffic networks and reducing the need to build new roads. Ride-sharing apps like Uber and Lyft are reducing the number of cars on the road. Machine learning models for self-driving cars like the ones developed by Google’s Waymo are reducing the need for car ownership in the first place.

So how would Andreessen look at a technology like AlphaFold? At its core, AlphaFold is software that predicts the structure of proteins. AlphaFold could revolutionize our understanding of disease and lead to new, more targeted pharmaceuticals for up-til-now intractable disorders. It’ll also likely accelerate the time it takes to bring new medicines to market, potentially shaving years and hundreds of millions of dollars in costs from drug development. Interestingly enough, one of the crucial pieces in finally concluding the war on cancer may very well be this software. In Andreessen’s words, software is eating biology.

But is software eating other scientific fields as well? Perhaps. As DeepMind wrote in its announcement, “This breakthrough demonstrates the impact AI can have on scientific discovery and its potential to dramatically accelerate progress in some of the most fundamental fields that explain and shape our world.” Indeed, scientific research has traditionally been dichotomized into theoretical and experimental science, but the rise of computational science—using machine learning to model scientific phenomena—is turning science into a trichotomy. Here, for instance, is a Google N-Gram showing computational science’s relatively recent explosion onto the scene of science:

The combination of software with the traditional sciences might be able to help us understand scientific phenomena that we heretofore have been unable to crack.

Thiel’s rejoinder

While Andreessen’s argument is compelling, at the end of the day, I’m still more sympathetic to Thiel’s view and his response. In a debate with Andreessen a few years ago, Thiel said:

The question becomes: Are computers alone enough to save us? It’s very hard for them to break out of this virtual world when the real world remains massively regulated. I think it's a very interesting question what will happen with the self-driving cars. Will you be able to have them? Who’s liable? Will the company doing the computer code be liable for car crashes? All these sorts of questions will be tremendous hang-ups in a super risk-averse society that we live in.

In other words, the Star Trek computer, acting alone, can’t bring us the innovation in the world of atoms we’ve been sorely missing in the last 50 years. Indeed, when it comes to self-driving cars, the algorithms are all well and good, but don’t forget that you’ll still need to build the actual car and clear the regulations to get the cars on the road. Similarly, when it comes to cancer, AlphaFold is great, but let’s not forget that anyone developing a cure for cancer still has to get past the big, bureaucratic FDA. We need something more than just the software. So, AlphaFold may have eaten biochemistry, but how much did it actually bite off? And how much do we still have to do before we get a cancer cure? We can sum it up in an activation energy-like graph:

Under this interpretation, perhaps Andreessen’s and Thiel’s points may actually be compatible. Software is eating the physical world but leaving behind the unsavory scraps for the rest of us humans to clean up. The question is whether we’ll have the appetite to do it.


📚What I’m reading

  1. Facebook’s experimental hate-speech policies seem to be working in Myanmar. Interesting quote: “It is the first time [Facebook] has created country-specific community standards.” (Rest of World)

  2. Hardware and software for general robots. It’s not as easy as you think for a robot to perform a “simple” task like opening a package of dates. (Eric Jang)

  3. A conservative commentator analyzes the 2020 Election fraud claims. (Ag_conservative on Patreon)

  4. Demography is not destiny. Democrats counted on a coalition of minorities to give them permanent electoral dominance. But no one asked the minorities themselves. Interesting quote: “Conservatives should have more faith in the power of their ideas . . . Conversely, Democrats need to understand that their identity categories are not determinative of voter choices and that they, too, need to convince the electorate that they support good ideas.” (American Purpose)

  5. How to save democracy from technology. (Foreign Affairs)

  6. Welcome to the party, Zoom. The backstory: San Francisco State University was planning a Zoom event with Leila Khaled, a member of the Popular Front for the Liberation of Palestine, which the U.S. has designated a terrorist group. The event had 1,500 RSVPs. Zoom, however, canceled the livestream event because it was afraid it’d be criminally prosecuted for “material support” of terrorism. Other internet platforms like Google and Facebook have been dealing with analogous content moderation problems for years now, so it’s interesting to see Zoom joining the fray. (NPR)

  7. Unforgiven. I’m technically watching (not reading) this, but it’s so good. Dave Chappelle weaves a compelling story about how poorly he’s been treated throughout his career. (Dave Chappelle on Instagram)

  8. Van Buren v. United States and the Computer Fraud and Abuse Act. (SCOTUSBlog)

  9. Substack billionaire. Mike Solana, a VP at Founders Fund, writes a hilarious and insightful piece about tech, the media, and his recent run-in with a California politician. (Pirate Wires)

  10. Tony Hsieh’s American Tragedy: The self-destructive last months of the Zappos visionary. (Forbes)