TAM - part the seventh

Yet another report from the NYT, which you might be able to see.

But basically, the idea was something like: These systems are merely good at seeming clever and are not going to be even capable enough to do a ton of economically valuable work.

AKA, TAM ≡ 0. Ha Ha Ha

This is the original from which that quote was built. It's not shown in quotes, and the original is behind a paywall, so it should be a precis of the original. But being behind a paywall... OTOH, the wiki has a page devoted to stochastic parrots. The term means what you think it does.

Again, from the NYT report

I thought: This is going to be a useful tool, and it's going to be powerful for a lot of things, partly because I believed then it would continue to get better. But I didn't think that people who were worried about more catastrophic scenarios were right.

He's changed his mind.

Two conflicting notions: 1) AI/LLM/whatever will never produce anywhere near enough end-user revenue to offset the trillions of Bongo Bucks already spent and 2) Ayatollah 17% Don^©'s newly annoited AI czar has surely been given the remit to take AI for Batshit J. Fucking Moron^©Tillerson's control of the United States of Alabama. Just as I, and ever more others, have been warning for a while. If you disagree: why wouldn't he? He does take whatever he wants, under threat of annihilation. This is the crown jewel of Dictatorship. He's going to pass it up?

Clayton unqualified for DNI, and unqualified for AI. Well... appointing a Gumby is kinda what he do. He's just a Gumby his own self.

I Told You So - 2 October 2026

Well... may be. Turns out, the AI hacking community does know SQL

In another case, the agents used what is called a S.Q.L. injection attack — where unexpected information is injected into a database in an effort to break through the network's defenses — to burrow into the systems of a company that operated a call center. They could then listen in on any of its customer calls.

Naturally, there's no guarantee that these AI engines either know how to build 5NF databases or care to. Since they're, so far as anyone knows, driven by kiddlie coders, they're most likely trodding the same RBAR world as their COBOL slinging GrandPa. Oh well. One might speculate that the "open weight" models have sense enough to store such weights in a RDBMS. Time will tell.