Bluesky

Hard disagree.

Most technology goes through phases like:

• Novelty of limited utility (think pre-1980 home computers or a pre-1900 motorcar)
• Growing maturity that shows the product has mainstream viability, but does not conquer the market permanently (think original IBM PC or Ford Model T)
• Final stabilized product.

You can get onboard late in the "Growing maturity" phase and still make a great business, while avoiding some of the pie-in-the-sky thinking of the "novelty" phase.

AI has not really left the "growing maturity" phase. You could even argue it's still in the "novelty" phase, as the "let's add AI to everything" mindset is a lot like flogging three-wheeled cars with boat steering wheels, or home computers with paper tape punches to store recipies on.

There's still a huge hype bubble to prick. People think AI does things better than it actually does, because it avoids some specific, familiar, human foibles. Notably, the whole "confident and grammatically clean" thing that can disguise huge factual inaccuracies. Hence all the summarizing/desummarizing tools. As we see through that, will the AI business see a retrenchment, cut back to much more narrow markets where it might have more value?