TAM - part the third

Well... that didn't take long. Yet more reporting, with data, from NYT.

There is, in the econ biz, an acrnonym TANSTAAFL which stands for There Ain't No Such Thing As A Free Lunch. Origin not known, although some/many/most sources go back to 19^th century taverns - "free lunch" if you spent enough.

The AI, as currently implemented, craze has been operating, if not "free", at least way, way below cost.

This latest report from the NYT explores evolved GPT models, driven by the same request. The models are 3.5, 4.o, and 5.5, the number offered for each is the number of "tokens" to process the request. So, here we go:

3.5 448 at $0.00

4.o 771 at $.01

5.5 6,463 at $.19

The report goes on to explain how each is, to all intensive porpoises, just a black box. There have been some reports that not all systems/models/developers actually know what's going on inside these Black Boxes.

Justin Summerville, an executive at the company, said many businesses "spent the past year yelling at their employees to use A.I., use A.I., use A.I."

But that's come at a cost, he added. "And now the finance organizations are waking up and saying, 'Whoa — this is a big number,'" Mr. Summerville said.

Dontcha just hate the bean counters? Always ruining the party. What was that the Wicked Witch of the West said? "I'm melting!" Just as fast as the AI TAM.

Here's another take, similar but not identical, to address the issue.

Faced with escalating variable costs, enterprises are capping their cloud consumption. The financially viable alternative is shifting to owned server architectures. By deploying open-source AI models on proprietary hardware, organizations bypass restrictive and costly subscriptions entirely.

With a grain of salt, naturally. The report is from a stock analyst with a bone to pick someplace. But, if it happens that there does emerge a significant shift from AI-operator operated models to in-house "open source AI" on owned hardware? More melting for the likes of OpenAI, et al.

TAM - part the second

So, Ok, how's it going for the AI TAM? According to recent reporting (today in my dead trees version) in the NYT, not so well. As stated before, any product falls (mostly, generally) into one of two buckets: intermediate good which is incorporated into a final user product, and, as you might surmise, a final user product. With AI, the latter, so far, is mostly adding steroids to web search, aka Gemini mostly. As to the former, this report deals largely there. And there isn't good. At least from making Bongo Bucks profits, that is.

Is there any evidence, or even logic, that makes a case for search on steroids as a Market? IOW, is it reasonable to expect that Google, et al (to the extent that there is an et al) will reap even more advert revenue traceable to AI function? My guess is just, no. At some point Google, et al will have to fess up.

On the next hand, there've been reports of various AI being used to generate technical papers, of questionable veracity. That's a Brave New World.

[T[here's a [sic] been a growing wave of AI-generated and AI-assisted academic papers hitting journals and preprint services like arXiv. It's a trend that is both hinting at new forms of research and discovery, and threatening to overwhelm academic literature with a tsunami of pseudo-intellectual AI slop.

An earlier report said this

But that level of AI use does not come cheap. Just months ago, many businesses encouraged employees to use AI for nearly everything, a trend known as "tokenmaxxing." Now, as the bills arrive, some are introducing limits and asking whether all that usage is actually producing enough value.

That's an example of the intermediate goods situation. The NYT report is even worser.

As A.I. costs skyrocketed, Andy Markus, the company's chief data and A.I. officer, looked for cheaper alternatives. He landed on "open" models, which can be downloaded and modified without payment or approval.

By May, open models accounted for 20 percent of AT&T's A.I. use. That has since risen to 40 percent and may jump to 60 percent in the coming months, Mr. Markus said in an interview.

"We believe it could go much, much higher," he said, adding that AT&T was saving up to 80 percent on A.I. costs compared with earlier this year.

Where do we go from here? One avenue -

At AT&T, the telecom company is now taking open-weights models and customizing them to make new tools that can transcribe calls and facilitate customer service, Mr. Markus, the company's chief data and A.I. officer, said.

Once again, the intermediate goods venue, which depends on firing humans to generate "savings" Bongo Bucks. What happens after you've fired the Last Man? Good question. Where's that part of TAM? It's Nowhere Man. "Making all his nowhere plans for nobody".

So, beyond the recent and continuing euphoria with The Latest New Toy, what's the long term TAM for AI? I don't think anybody knows, cares, or thought about it. I still think it ain't anything worthy of Trillions of Bongo Bucks. The bubble may well break before the first Trillion is booked. One can only hope; the fall won't be as far or hurt as much.

Life, Liberty, and the Pursuit of Grift

This summary is not available. Please click here to view the post.

Gimme Shelter

So, if any in the audience thinks that AI, as currently implemented, may be regulated by Kim Don-Felon and his Merry Band of Fascists (so says this Tesla investor), think again. Real hard. Having the Fascists in control is the goal, as described here. That's the NYT review of what happened, and it surely caused every manjack of the Kim Don-Felon and his Merry Band of Fascists (so says this Tesla investor) to get an instant hardon.

Just think of how simple and easy it will be for Kim Don-Felon and his Merry Band of Fascists (so says this Tesla investor) to totally control The United States of Alabama. Putin and Hitler would be proud. For as long as it lasts, since Batshit J. Fucking Moron^©Tillerson will drive the economy over a cliff. And he'll depart to Moscow with his Bongo Bucks safely in Switzerland.

This time, the A.I. collective didn't seize a military network, hack a hospital or shut down an electrical grid. This time, humans regained control.

Next time, we might not be so lucky.

Next time, we won't be so lucky. Kim Don-Felon and his Merry Band of Fascists (so says this Tesla investor) will be the humans in control. Only for themselves.

Cruelty is the point.

That's Entertainment

A few times in the course of these missives, I've mentioned that measures of "economic impact" of various consumer-optional entertainment are basically bogus. Here's a long piece from the NYT on the economics of the US Open tennis program. It does make the case, but only indirectly, that value of such activity matters only if measured spending is from consumers outside of the locale in question. One might be picky and assert that the appropriate venue is the USofA in toto. Why? Because of what we in the econ space call substitution effect. Or, as one skeptic put it (near the bottom of the piece, of course)

Some are skeptical of these kinds of projections. Andrew Zimbalist, a professor emeritus of economics at Smith College and an expert on the economics of sports, has not seen the study. But he cautioned that there are ways to make data fit a narrative.

"It's quite possible that there is some positive impact," he said, adding that money spent by New Yorkers at the tournament is money that most likely would have been spent elsewhere in the city — on Broadway, at a Yankees game, at a restaurant — and now won't be. "In that sense it's a net negative," he said.

A spokescritter for a restaurant org went on to say that benefit outside of the immediate area around the center doesn't amount to much. Duh.

Exactly how much Damn Gimmint Bongo Bucks were used to create the current facility, I haven't found. But it was built, apparently for free, on city land. There is an audit from 2019, which shows that USTA isn't exactly being taken to the cleaners by NYC on an annual basis. The report says the US Open is worth $1,200,000,000 to NYC each year. The City Gummint doesn't get but a few pennies. Sweet heart, I luv you so!

One can site casinos and data centers as boons to the local economy. Which never seem to work out that way; Atlantic City, I'm talking to you. And how many uneducated shitkickers are going to work in those centers in fly-over country? Yeah, right.

TAM

For those not knee deep in bidnezz, that's total addressable market. If you're not some scammer or grifter (hard to find such, given these days of Grifter In Chief ), then the first thing you do when you want to move some new (or importantly updated) widget into Mr. Market is figure out how many end users (aka, buyers) you expect exist. In the pharma world these days, drug companies are actually looking for a TAM that's tiny thanks to Uncle Sugar, who passes on some substantial benefit to them if FDA approves a drug for very limited number (200,000 ... well sorta, kinda) of patients. Although gaming continues as ever. This is one segment of the economy where the TAM is sorta, kinda explicit. And being 'small' is good.

For AI, as currently implemented, it's a bit different. And that difference has been recited here a bit. And it matters, given that Big Tech is spending some non-trivial Billions of Bongo Bucks on hardware (data centres) alone. Who are the patients to be served? Do they have, taken as a whole, enough Bongo Bucks to recoup all those trillions of Bongo Bucks, plus a bit of profit? Here's a recent assessment. As predicted in these missives, the arithmetic looks awful. And, some of the tech bros are worried.

At one point this year, Meta internally projected that it could spend as much as $10 billion annually on Anthropic's A.I. models, said two of the people, who declined to be identified discussing private information. That would have formed a major chunk of Anthropic's yearly revenue, which the start-up estimated in July would pass $65 billion.

Now, that cash flow is wonderful for Anthropic and all of the AI vendors out there. But Meta, and Google, and Microsoft, and the like are buying what we econ types call 'producers' goods'; IOW, stuff that the buyer turns around and uses to make some finished good which Americans on Main Street, Anytown, USofA actually pay for. Well, Meta et al, ain't selling any new thing to Americans on Main Street. Remember, so far as Meta goes, the user is the actual product, sold to advertisers. Microsoft, is more or less, the exception (no one uses Bing).

As of April 2024, Bing holds the position of the second-largest search engine worldwide, with a market share of 3.64%, behind Google's 90.91%.

So, where is the flood of Bongo Bucks to make it all worthwhile? It's a verity that Real Producers always bitch and moan about the Bean Counters ruining the party. I suspect the Real Producers will lickety split concoct some tale to blame the Bean Counters when it all goes Pop.

Mr. Zuckerberg has described Hatch as Meta's next breakthrough, a personal "agent" that works "24/7 on your behalf to help achieve your goals and improve your life, your health, your relationships, your finances."

(By the way, what was the last New Product that Meta made that was a Bigly Win? Just asking for a friend.)

OK. A semblance of recognizing that somewhere, somehow a real live Human (American or otherwise) will use an AI derived service/product. But will such a Human pay a user's fee for the privilege? And, if so, how much? Or does The Zuck assume that Hatch will somehow, some way entice his real customers, advertisers, to pay more and more often? In either case will end users or advert makers generate enough new revenue to make it all make sense? How would you bet on Kalshi?

It's been mentioned here and there that, given existing accounting rules, all this data centre spending is hidden in corporate capital budgets. IOW, very little of it hits their current financials. There will be a reckoning and there will be blood.

Thought For The Day - 28 August 2026

The hits keep coming. More like attack drones than top 10 songs, though. Latest grift

Companies Can Tell Investors Less Under Proposed S.E.C. Rules

Insiders can make an even bigly boodle. Won't take long for Grifter In Chief to trade ever more eagerly. And, of course, his best billionaire buddies. The USofA is looking more and more like any Eastern Europe -stan.