Generally, the content on this Substack is evergreen. Given the infrequency of posts, as well as personal constraints, it’s better for my sanity and for the wandering internet (little e) explorer to keep it this way.
But sometimes your humble blogger cannot help but get caught in the whirlwinds of the news cycle. There are a few stories, here in August 2023 (and which may be irrelevant almost immediately after this email is sent), whose timing, or convergence, is particularly interesting.
Here we go.
(1) OPPY, baby!
Oppenheimer came out this past month, and though not flawless, it is undoubtedly a masterpiece. I’ll resist my amateur movie critic urges, but to summarize why you should see it: it’s like The Dark Knight, but with better acting and the best cinematography ever. (If that doesn’t convince you to see it, don’t bother to continue reading, as whatever arguments I make henceforth will be weaker than this one.)
It has the potential to be the most important film of our lifetime if it accomplishes one thing: nudge us to be more physics-curious.
If we were to use annual federal spending as a proxy for our curiosity (bear with me), we could start to see how little, relatively speaking, we collectively care about physics research. The most recent federal spending bill, a light 4,115-page read that was condensed by the kind souls at Science, highlights a few spend categories of interest:
Fusion energy sciences: $510 million
High energy physics: $868 million
Research in nuclear physics: $755 million
These numbers are modest by Saudi Arabian sports investing standards, let alone those of US government spending.1
(2a) LK-99
From GRIT:
LK-99 has social media and the science community in a frenzy - what is it? LK-99 is rock. But not just any rock. It’s a rock made from lead, oxygen, phosphorus, and copper. What makes it special is that it has the potential to conduct electricity with zero resistance at room temperature. Currently, all known superconductors require either extreme cooling or pressure to function. This makes many applications impractical. Applications for room-temperature superconductors, on the other hand, would be virtually endless. Right now, scientists around the world are rushing to verify breakthrough claims made by a paper published out of South Korea. With that said, any real-world implementations (assuming claims are confirmed) are still years (or even decades) away.
The Base Rate Times – a truly diamond-in-the-rough Twitter/X handle2 I recommend checking out – reported that some prediction markets had the odds of the study being replicated by 2024 at ~50% last week, before coming down to (a still very exciting!) ~15%:
…which led Twitter Mt. Rushmore inductee Paul Graham to go viral with this:3
Might it be a Twitter-propelled, science-y version of Tulip Mania or COVID NFTs? Absolutely. But as Graham points out, a low probability, super duper high upside idea is worth chasing. To use a sports analogy, if prime Michael Jordan declares himself a free agent, even the lowly Charlotte Hornets would be acting rationally in attempting to recruit him.
We now have an out-of-the-money call option on room temperature superconductivity happening way sooner than anyone expected.
(2b) Major Lasers
Remember when scientists in California successfully got 2.5 megajoules (MJ) of energy out of a reactor using just 2.1 MJ of input, demonstrating the first-ever net gain in nuclear fusion research? It was last December, and Neil deGrasse Tyson went on a bunch of podcasts to talk about lasers, which you half-listened to as you stressfully navigated TSA lines over the holidays?
Well, they did it again, but better:
Researchers at the federal Lawrence Livermore National Laboratory in California, who achieved ignition for the first time last year, repeated the breakthrough in an experiment on July 30 that produced a higher energy output than in December, according to three people with knowledge of the preliminary results…Achieving net energy gain has been seen for decades as a crucial step in proving that commercial fusion power stations are possible.
It’s always hard to gauge how seriously to take news like this, but screw it, it’s officially Science Month.
(3) AI: don’t forget about me!
Generative AI is still a gamechanger, even if it ends up taking a backseat to LK-99 or accelerating fusion progress as the crowning achievement of the 2020s.
In hindsight, the frenzy over the release of ChatGPT a few months ago seems appropriate, but narrow. It became the fastest product ever to get 100 million users, but pretty much every noteworthy demonstration of its awesomeness was in the domains of coding or writing. This is still worth celebrating, since coding and writing are things that literally take up tens of billions of person-hours each year. That is a massive, massive addressable market for improving productivity.
The techno-optimist argument, though, is that these are just the first applications of large language models’ usefulness. There are many other tasks, representing many orders of magnitude more person-hours, that AI will help streamline, freeing up humans to do the things that matter more, like solving new problems (and leisure, too).
Convergence
So here’s the dream scenario:
Oppenheimer is so good that it changes the public consciousness.
The LK-99 achievement is a physical science breakthrough on the order of the Manhattan Project.
Bonus points: Fusion starts going down the path of Moore’s Law, from 1.5X net energy creation today to 100X – enough to be commercially viable – within a couple decades.
AI is the intellectual companion the optimists hope it can be, helping us skyrocket productivity in both the digital and physical realms.
We kickstart a science (specifically, physics) renaissance that changes the trajectory of homo sapiens. It even becomes socially acceptable to discuss e/acc at a dinner party.
The probability of this happening, exactly as written, is very deeply below 1%. A far more likely outcome is that this wave of attention on superconductors is a dead end; fusion news trickles along throughout our lifetime on obscure websites and basically nowhere else; Oppenheimer occupies a cultural significance somewhere between Interstellar and Memento; and AI encounters some sort of glass ceiling on improving economy-wide productivity.
Even more likely is we land somewhere in the middle, and the descriptions in both scenarios look incredibly naive five years from now. Such is the condition of the world.
There is reason to celebrate: progress is in the news, and our culture has an opening. Dead ends or not, it’s a high upside moment that we should be excited about.
It is worth mentioning that the Census is over $1 billion. I don’t mean to be reductive, but like...I can’t really fathom how counting people gets more funding than nuclear energy research.
I’ll eventually correct it, but “X handle” sounds like an NBA Jams power-up, and idk, I’m just not quite there yet.
…which led to the dumb and confusing commentariat a Paul Graham tweet typically leads to, from which I will protect you via omission.