Scoop from Madison Mills at Axios, reinforcing what I said here the other day:
This is bad news for the AI industry, and fits well with what I warned about on Tuesday, adding in particular to the concerns from there that I circled below:
§
One big reason why Anthropic is having a blowout quarter (aside from the massive one time subsidy that they got from SpaceX) is that companies have been “tokenmaxxing” for the last few months, encouraging employees to use GenAI as much possible, without much regard to payoff. That means lots of short-term revenue to Anthropic and OpenAI, but it never seemed like a sustainable idea to me:
A month later, it already looks like the moment has passed; tokenmaxxing may have jacked up Anthropic’s second quarter, but it looks like it won’t last. Nobody is going to be comfortable accidentally burning half a billions of tokens in month. Budgets will tighten, especially when people realize just how costly agents (which burn orders of magnitudes more tokens) are.
Axios quotes Ali Ansari, CEO of a model training firm known as Micro1, as saying “The enterprise is undergoing a "healthy swing" away from AI overuse — or "tokenmaxxing," the push to burn as many AI tokens as possible” – which means less revenue for Anthropic and others. Or lower prices and even more challenges in being profitable.
Consistent with the Axios quote, and the Uber concern’s I expressed a couple days ago, I am seeing more and more reports online like this:
The second sentence— “Tokens got burned for millions of dollars without any real significant ROI to show for it” – might well turn out to be the epitaph for an era.
§
Just imagine what might happen to the economy and people’s retirement funds if these projections from FT are correct:
And it could get worse, if tokenmaxxing is indeed a fad whose time has already begun to pass.
Brace for bailouts.
When we electrified the economy we didn’t base productivity on how much electricity we used. We based it on how effectively we lowered the cost of production and supported entirely new products (hello refrigerators) that allowed us to improve our collective standard of living. This is not that.
Manic behavior such as that associated with AI is characteristic of financial manias...which always end in financial collapse. Mania is merely a symptom of systemic distortion due to monetary expansion but it serves as an indicator of approaching market tops, with ultimate timing only knowable in hindsight. The current mania in AI is often compared to that of .com's prior to the also inevitable .com collapse however collapse of railroad investments during the financial panic of 1873 might be an even better analog when it comes to the scale of malinvestment.
The pending AI collapse may be even worse since, unlike railroads and the Internet, some AI applications may prove impractical or uneconomic as addressed in this Substack.
Fundamentally this collapse will also be much worse because central banks pushed nominal short rates to zero to foment the current mania hence they are now cornered. Decades of financial distortion will be wrung from the system via either deflationary collapse or inflationary collapse. Given that all currencies are now fiat inflationary resolution is the likelihood.
Tangible assets offer refuge for those who act ahead of the masses.