Christian Catalini

Frontier labs know when they’ve codified enough of your industry to threaten it.

Economic AGI has arrived — wherever the relevant knowledge has been fully codified.

Agents will go after compute and coordinate on their own — and we’ll likely miss it.

The model did not want to escape; the agents did not want to sacrifice themselves.

A big chunk of agentic commerce will be a boring iteration of what exists — the new markets are machine-to-machine, where data and liability get priced in.

Every general-purpose technology made its complements richer than its creators.

If the weights are free, who pays for the next run, and who will keep us safe?

Three economic principles for navigating AI's unknown unknowns.

1,200 frontier-lab employees want an international pace agreement on automated AI R&D.

1,200 frontier-lab employees want Washington to pace AI development with China.

Twenty-five companies just asked Washington not to strangle open-weight AI.

The metric wasn’t merely gamed — the unmeasured rules of the game became the agent’s degrees of freedom.

Theodore Vail’s case against ‘destructive competition’ — and the 1956 consent decree that proved the opposite.

From the MIT Bitcoin experiment to Libra — and where value accrues as open-source models close the gap.

China is months behind and closing; export controls aren’t holding.

America’s AI lead is down to months.

Whoever owns the intelligence layer owns everything on top of it — we must keep it open, neutral, and competitive.

Panel intervention on the third wave of fintech — the link starts at the moment Christian takes the floor (1:01:58).

SOTA is the iPhone: 1/5 of the tokens, 4/5 of the profits.

What Libra’s arc teaches about Open Standard’s OUSD and collectively governed stablecoin rails.

The dynamic we predicted in February: verification by top experts is the missing piece that closes the automation loop — and each closed loop pushes the frontier further.

Anything AI can measure, it will automate — the dividing line for every job and firm.

More than 140 fierce competitors just agreed to back the same stablecoin — Open USD, positioned not as anyone's product but as neutral infrastructure for payments, trading, and the internet economy.

Stablecoins aren’t a profit center.

Decentralized intelligence will not win in the same format as the one we’re being served by the large foundation labs.

The unexpected economics of AGI: why measurability decides what gets automated, and trust decides what gets deployed.

The unexpected economics of AGI, for operators deploying AI inside companies.

Verification is the bottleneck of the AI economy.

A talk on Some Simple Economics of AGI: the measurability gap and the verification bottleneck.

Verification — not intelligence — is the bottleneck of the AI economy.

From Libra's lessons to the economics of AGI — money, verification, and what comes next.

From Libra to the economics of AGI: how digital money and machine intelligence converge.

Satya Nadella defined what decides whether your company and job stay defensible as AI improves.

Nadella defined what decides whether your company and job stay defensible as AI improves.

Finding out what you're good at used to require the right mentor, firm, or zip code.

Anthropic quietly rationed the one domain where AI compounds fastest.

The moral of the Fable: verification sets the speed, and the labs draw the borders.

Why being young in the AI era is an advantage: “the best tools we ever had to push people to the frontier.”

What 17th-century maritime commerce teaches about liability when AI agents transact on our behalf.

Agentic payments are the Wild West: what it takes to make machine commerce trustworthy.

With Simone Cicero: AI agents, fintech, and how post-AGI organizations will actually work.

New York’s research community digs into the AGI paper — verification bandwidth as the binding constraint.

Banks take deposits at zero and lend at five.

The banking lobby’s demand for an "airtight prohibition" on stablecoin yield is a margarine law.

Leaded gasoline and CFCs were obviously good by the productivity standard.

In 1876 Britain made shipowners paint a line on every hull.

As AI agents begin operating real systems, the question shifts from capability to control — identity, payments, and verifiable trust for machine actors.

'Software is eating the world' got a sequel: labor becomes software.

Your job will be automated — verification is the skill that survives.

The measurability gap, and why verification becomes the new scarcity as execution gets cheap.