Tobias Burns

News investing writer, CNBC

[Miranda Nazzaro @mirandanazzaro]

Read: https://t.co/qDjdDx0qGa

[Tobias Burns @tobyburns1]

The contemporary tableau economique from Godley, Lavoie 2006. Better late than never.

[Tobias Burns @tobyburns1]

Amazon is getting tariff refunds; they got $600 million in Q2. They "identified a limited set of circumstances" where they passed the cost onto customers and will be "proactively" contacting those affected.

[Deirdre Bosa @dee_bosa]

lots of folks want Kimi K3 to be a distillation story....

distillation may be part of it but it doesn’t explain how Chinese AI labs keep designing, training, and shipping models this good, this quickly.

"they copied" is not an AI strategy... time to build an American open source one

[Jason Paladino @jason_paladino]

Scanner audio reveals that a woman in Baltimore makes the 911 call, but Lindsey Graham is home alone in DC with a locked door that has to be forcibly opened.

[Tobias Burns @tobyburns1]

https://t.co/cmza5PaU62

[Reuters @Reuters]

EXCLUSIVE: Beijing is looking at curbing overseas access to China's top AI models, sources say https://t.co/IVVXnf2qI7 https://t.co/IVVXnf2qI7

[Ed Zitron @edzitron]

Here's my full interview with CNBC, covering my bear case against generative AI, OpenAI's questionable finances, AI's lack of ROI, and how all of this is a symptom of the tech industry running out of hypergrowth ideas.

It's great to see the mainstream media discussing this.

[Lee Edwards @terronk]

Not 2026 enough. Should have just been AIbirds and gaslight everyone about the L always being an i.

[Erez Maggor @ErezMaggor]

📰What was Bidenomics?
My new paper at @pasupdates uses developmental-state literature to reassess the IRA,CHIPSAct, and infrastructure law.I argue these policies were more than derisking: they marked a break with neoliberalism and revealed a no-longer hidden developmental state

[Tobias Burns @tobyburns1]

https://t.co/iANfB5NKrY

[Tobias Burns @tobyburns1]

https://t.co/pSZcgdW2LR

[Tobias Burns @tobyburns1]

Warsh/Greenspan productivity calls
https://t.co/5CthZrk7f1

[Milk Road @milkroaddaily]

Kevin Warsh just made an argument that, if he's right, sends interest rates significantly lower than the market expects (save this).

His claim is that we're at the front end of a deflationary wave driven by AI making the cost of production fall across nearly every industry.

His deeper concern is that the Fed is still running on economic models from 1978, with no institutional framework for recognizing what happens when a technology-driven productivity boom permanently changes the relationship between growth and prices.

He pointed to Alan Greenspan's famous bet in the 1990s...

Greenspan held rates steady and let the economy run hot because he believed the internet productivity boom was real and would show up in the data.

He was right. Output per hour grew 2.7% annually while inflation dropped to 1.9%, one of the most beneficial macro environments in modern history.

Warsh thinks that bet is on the table again, and that a Fed anchored to old models risks tightening into a productivity boom and strangling growth that would have otherwise been non-inflationary.

The data backing the thesis goes like this:

The cost of running frontier AI has collapsed from $30 per million input tokens in early 2023 to pennies today, and by the end of 2026, today's frontier capabilities will likely be available near-free.

When the price of intelligence approaches zero, the cost structure of every industry that uses intelligence starts falling with it.

But there's a flip side worth taking seriously.

AI-related price pressures have already added roughly 0.3 percentage points to core PCE through hardware prices, software hikes disguised as AI upgrades, and electricity costs from data center power consumption.

Microsoft raised M365 by 30%. Adobe raised Creative Cloud Photography by 50%. Intuit raised QuickBooks by 45%. All justified with AI features, all showing up as price increases in the inflation data.

Almost everyone agrees AI will eventually be deflationary if it delivers on the productivity promise.

The main fight is over timing. Is the payoff 2 years away or 10? And should the Fed act on that expectation now or wait for the data to confirm it first?

Nearly 60% of economists surveyed by the University of Chicago's Center for Markets said Warsh's AI thesis would have minimal impact on inflation or rates over the next two years.

If he's right, we're entering an environment where growth stops being a rate hike trigger, deflation comes through the supply side, and the companies investing most aggressively in AI now pull further ahead.

If he's wrong and the old models still apply, rates stay higher for longer and the productivity payoff gets pushed further out than the market has priced in.

The market is betting he's at least partially right.

If you want to track how this plays out, what it means for rates, AI infrastructure investment, and the risk assets (like crypto) positioned to benefit from a Fed that finally sees that the productivity boom is real...

Join Milk Road PRO, link bio: @MilkRoad

[Tobias Burns @tobyburns1]

https://t.co/7VbcjdkNXf

[Tobias Burns @tobyburns1]

https://t.co/h9pbwmPL7g

[Tobias Burns @tobyburns1]

https://t.co/lrgZCcSPhO

[Tobias Burns @tobyburns1]

https://t.co/RfmeFL8Mqf

[Tobias Burns @tobyburns1]

https://t.co/XHcucqvrlv

[Tobias Burns @tobyburns1]

https://t.co/RK8W9tD7iO

[Deirdre Bosa @dee_bosa]

"One of the biggest misconceptions"

Cerebras CFO @BobKomin pushes back on the small-models narrative.

"We serve all models, and there is no limit to the size of the models that we can serve. Today, we're serving trillion parameter models. We're serving trillion parameter models that are internal for OpenAI today. We are currently running OpenAI 5.4 and 5.5 with them."