π Faster, Please! Week in Review #53
Europe's Doom Loop; the new Anthropocene Epoch; the CRISPR Revolution; and much more!
My free and paid Faster, Please! subscribers: Welcome to Week in Review. No paywall! Thank you all for your support! For my free subscribers, please become a paying subscriber today.
Melior Mundus
Essay Highlights
π Learning from Europe's Doom Loop of Decline
There is a growing gap between living standards enjoyed in the United States and those across the pond in Europe. Over the past 15 years, the International Monetary Fund shows, the eurozone economy has grown just 6 percent, compared to 82 percent in the US. So what's Europeβs problem? There are a few obvious answers: lack of access to venture capital, inflexible labor markets, and a heterogeneous home market of distinct languages, cultures, and regulations. The bit about culture strikes me as pretty important. The churn of American society β companies starting and dying, workers switching firms β is also key to America's innovative and productive capacity. I think the macro lesson here is bad things happen when you stop valuing innovation and the benefits of creative destruction. (Of course, it would be even worse if the US went down the EU path since at least the EU can benefit from American innovation.) Weβre already seeing Europe embrace the precautionary principle with AI, with plenty of American policy activists and policymakers suggesting the same here. I hope the US continues to show that its competitive advantage is mostly avoiding the anti-progress mistakes that Europe continues to make.
π Fear not the Anthropocene Epoch!
According to a panel of earth scientists from around the world, humanity in the 1950s seems likely to have entered a new geological era marked by human harm to the planet. Some say humanity should consume less, use fewer resources, and have less of a presence on this planet. But if we want a healthier and more prosperous humanity living on a healthier planet, then humanity must advance. A successful demonstration by Fervo Energy, reported earlier this week, is just the kind of advance we need. Fervoβs βenhanced geothermal systemβ involves drilling deep wells, and then using fracking techniques to create an artificial geothermal reservoir. The company then pumps water into that reservoir, the water gets really hot, and then the water is pumped back to the surface to power a turbine and generate electricity. What this all adds up to is the prospect of clean, limitless, reliable energy. Hereβs what geothermal expert Wilson Ricks from Princeton University says of Fervoβs technical milestone, according to CNBC: βThere is still more development to be done on the path to large-scale and cost-competitive commercial systems, but the significance of this achievement shouldnβt be understated.β
π€ How do we know if the AI Revolution is real?
If the latest period of enthusiasm about artificial intelligence turns out to be more hype than reality, it wouldnβt be the first time. And you donβt need to go back too far to find an example of a coming AI boom that never arrived. Of course, it would be great to know the future in advance, or at least know what signs to look for if GenAI is really living up to what many experts think is its long-term potential. How to do that? For starters, we can look at whatβs happening now in real time. Venture funding in generative-AI startups has grown from $4.8 billion in 2022 to $12.7 billion in the first five months of 2023, according to analysts at research firm PitchBook. A good sign for the near term. And what about longer-term? A combo of looking at past trends β such as the post-1980 software boom β and GS modeling suggest total AI software and harware investment could a peak of around 2-2Β½ percent percentof US GDP in the next decade. Anyway, before we drive ourselves crazy with scenarios about mass tech unemployment or existential risk β or get crazy-go-nuts optimistic about the wonders of warp-speed exponential growth, let's start seeing more good AI investment data.
Best of the pod
𧬠My chat (+ transcript) with author and geneticist Kevin Davies on the accelerating CRISPR Revolution
Kevin Davies is executive editor of The CRISPR Journal and author of the excellent 2020 book, Editing Humanity: The CRISPR Revolution and the New Era of Genome Editing.
When people talk about AI, for instance, they might be talking about different versions or applications of AIβmachine learning being one. So when we talk about CRISPR, are we just talking about one technique, the one they figured out back in 2012? Are there different ones? Are there improvements? So it's really a different technique. So how has that progressed?
You're right. CRISPR has become shorthand for genome editing. But the version of CRISPR that was recognized with the Nobel Prize three years ago in 2020 to Jennifer Doudna and Emmanuelle Charpentier was for one, we can call it the traditional form of CRISPR. And if I refer to it again, I'll call it CRISPR-Cas9. Cas9 is the shorthand name for the enzyme that actually does the cutting of the DNA. But we are seeing extraordinary progress in developing new and even more precise and more nuanced forms of genome editing. They still kind of have a CRISPR backbone. They still utilize some of the same molecular components as the Nobel Prizeβwinning form of CRISPR. But in particular, I'm thinking of techniques called base editing and prime editing, both of which have commercial, publicly funded biotech companies pushing these technologies into the clinic. And I think over the next five to 10 years, increasingly what we refer to as βCRISPR genome editingβ will be in the form of these sort of CRISPR 2.0 technologies, because they give us a much broader portfolio of DNA substitutions and changes and edits, and give the investigators and the clinicians much more precision and much more subtlety and hopefully even more safety and more guarantees of clinical efficiency.
Right. That's what I was going to ask. One advantage is the precision, because you don't want to do it wrong. You don't want mutations. Do no harm first. A big advantage is maybe limiting some of the potential downsides.
In the ideal gene-editing scenario, you would have a patient with, say, a genetic disease that you can pinpoint to a single letter of the genetic code. And we want to fix that. We want to zero in on that one letterβA, C, T, or G is the four-letter alphabet of DNA, as I hope most of your listeners knowβand we want to revert that back to whatever most normal, healthy people have in their genetic code at that specific position. CRISPR-Cas9, which won the Nobel Prize, is not the technology to do that sort of single base edit. It can do many other things, and the success in the clinic is unquestionable already in just a few years. But base editing and, in particular, prime editing are the two furthest developed technologies that allow investigators to pinpoint exactly where in the genome we want to make the edit. And then without completely cutting or slicing the double helix of DNA, we can lay up the section of DNA that we want to replace and go in and just perform chemistry on that one specific letter of DNA. Now, this hasn't been proven in the clinic just yet. But the early signs are very, very promising that this is going to be the breakthrough genome-editing technology over the next 10 to 20 years.