> Generative AI is a novelty that makes us crazy productive at certain tasks. But it doesn't yet seem to fundamentally change what we build or why. We just do it faster and sloppier with AI. It's a tactical tool to help you win, whereas interest rates define the rules of the game.

We're building some stuff that actively uses it—not (just) using it to write code, but integrating it into business processes.

This is both:

1) A far, far more valuable use of it than as a replacement for e.g. macros in your editor, assuming it worked as one might hope it would.

2) In practice so incredibly brittle, tightly-coupled, expensive, and slow to develop (not to mention some of the most boring work I've ever done in my 25 year career) relative to other options that the business could have embraced at any time in the last 15 years (but didn't because it took the hype of "AI" to gain activation energy for the project) all with no evident path toward any of that meaningfully improving, that I'm looking for an exit to another project that's ideally non-AI-related for when this one turns into a nightmare before eventually imploding and staining everyone involved's reputation, if not getting them fired. I reckon the nightmare phase is about six months out, for this one, and the implosion 12-18 out.

I expect similar stories are playing out all over the place.

The end of ZIRP coincided with some reorganizing of the tax code that wasn't favorable to developers, as well. Both hit after a year or two of windfall profits and massive hiring due to covid.

AI is a very convenient way to tell that story as being about an ascendant new technology, rather than a post covid decline for the tech sector.

The increased interest in using generative AI to replace high-paid workers may well have been caused by the end of ZIRP.