Derek Thompson

folks in and near the labs sometimes say “we’re not prepared for a world where AI *isn’t* a bubble” and I think that’s true inasmuch as predictions like this imply that frontier lab revenue could exceed total global Apple or even Amazon sales by the end of next year

to say that scenario would likely be an inflationary event seems … almost quaint

Today's pod: @DKThomp on Abundance discourse, AI's econ impact, whether Claude has consciousness, & we get a lil ahead of ourselves on AES/Ossoff 28.

Then polling miss analysis, Senate rankings, and what we learned from primary season with @admcrlsn

podcasts.apple.com/us/podcast/der…

there are dystopias everywhere for those with eyes to see

New pod: THE GOLDEN AGE OF BIO & THE DARK AGE OF BIOWEAPONS

I talked to @ashishkjha about the Russian plague mystery, the age of bio miracles, the pre-AI threat of bioweapons research, the novel threat of AI to global bio-security, and what we can do about it.

We begin with this remarkable observation by Jha:

Think about the history of modern science in the last 100 years. In the late 1800s and early 1900s, chemistry took great leaps forward, nitrogen fixing. But soon after, in the 1910s, the same scientists who fixed nitrogen also turned their science into a weapon, as Germany deployed fatal chemical weapons on its adversaries in WWI.

... then in the 1920s and 1930s, physics had a golden age, as scientists cracked open the secrets of the quantum world. One decade later, they used that science to build another weapon, the atomic bomb, in another world war.

... now, in the 21st century, it’s biology in the spotlight. From CRISPR to mRNA drugs to cancer therapies and GLP-1s, we are in an age of wonders in biology and the life sciences. But the modern history of sciences teaches us that horrors often follow wonders. Seen in this framework, if the biological breakthroughs of the last few decades became agents of war, that would not be an abnormal break from history. It would be the continuation of history. And of course, that’s the scary part.

youtube.com/watch?v=8G-cMu…

Having 2 kids under 3 is an unspeakable joy in so many ways and also “statistically speaking, someone in the house has some kind of cold every other day” sounds just about right.

Having a sniffle, sore throat, or sinus pressure goes from a condition you once called Being Sick to a condition better known as Wednesday

One way to put it, maybe, is that the frontier labs are "frenemies" of the rest of the AI ecosystem.

In some ways, what’s good for the frontier labs is good for the entire AI industry: better AI → more useful capabilities → more usage → more revenue and investment → even better AI → new capabilities → even more revenue ...

But in other ways, what’s challenging for the frontier labs is good for the rest of the industry: cheaper models → lower costs → more applications → more AI use (Jevons, etc) → more inference → more chip/compute/cloud demand → compresses frontier-lab margins.

Another way to put this: There are several ways for AI to go right, but not all of those ways are best for the labs *as their businesses are currently constructed*

1. What the labs probably prefer: Intelligence becomes more valuable, a capability breakthrough allows for short-term scarcity profits, which funds the next frontier breakthrough, flywheel continues.

2. What others might prefer: Useful intelligence becomes too cheap to meter, disproportionately benefiting everybody building on top of the models

(... wrote this quickly, open to feedback!)

👀

Second

My toddler to me, everytime my wife leaves for a business trip:

good summary by @daveweigel

Really good AI data!

Several facts regarding the AI bubble question that are all true at the same time, with most points illustrated by graphs

1. AI revenue growth is incredible, and possibly historic.
2. Big public tech companies have relatively low P/E ratios right now, which defies lazy predictions about how they're wildly mis-priced and doomed for a big correction
3. But model competition could compress AI token margins ...
4. ... at the same time that the hyperscalers are going deeper and deeper into debt to build out AI ...
5. ... and the level of future free cash flow anticipated to justify this build-out is absolutely enormous, like genuinely without precedent in corporate history ...
6. ... and any correction in frontier lab valuations will reverberate through public companies, which own huge shares in the labs

A must read.

This is TOP tier

I’d love economists to take a crack at this theory that the US is experiencing a mini-Dutch Disease effect from AI.

My big-picture thesis about AI economics right now is that while the discussion about AI safety and existential risk is important, AI’s construction is changing the world faster than its capabilities.

This.
from @DKThomp

New newsletter: THE 26 MOST IMPORTANT FACTS ABOUT AI AND THE ECONOMY

After weeks of feeling a bit lost in a sea of decontextualized AI factoids, I decided that what I really wanted to read was a structure, a story. So I wrote one.

Today’s newsletter is a fact-by-fact, chart-by-chart story about what AI is actually doing to the economy, with answers to questions like:

1. I keep hearing about trillions in "AI capex," but where exactly is all that money going? (Chart 1 has got you)

2. If AI is powering an otherwise moribund economy, how is it also starving other sectors of capital and resources? (Chart 2 has got you)

3. Do Americans find AI helpful, or do they think it's going to destroy society? (Definitely yes, and ... uh, also yes)

4. How many of the best-performing stocks in 2026 are tied to the AI buildout? (arguably, 19 out of 25)

derekthompson.org/p/the-state-of…

question for econ/finance:

1. what's the strongest case that AI is behind the increase in treasury rates?

2. what's the strongest case that AI is behind the rise in inflation?

From @hamiltonnolan—"Every government will, in time, be faced with the question of what to do when the power of rich individuals begins to rival the power of the state itself." Sounds incendiary, but read carefully, it's hard to disagree. My chart...

The New York TImes: What’s one issue in politics that you’ve changed your mind about?

Zohran Mamdani: The role of the private market in housing construction.