I have worked with old fashioned neural networks, deep learning, and now LLM-specific deep learning: wonderful technology, but over hyped, and advice to go a little slowly, with firm use cases that are financially viable is great advice!
Outside a small bubble within Silicon Valley and the finance ecosystem funding it, I’d say most folks are increasingly fed up with AI.
It’s a very noticeable shift these last 6 months. The mood went from excited, to just annoyed at all the slop and folks using AI as a half-baked easy button vs doing real deep value-add thought.
Business is also noticing that the ROI simply isn’t there and a lot written about this. That doesn’t bode well for AI providers that need to massively increase prices to make the math work on their business models.
The world inside of the AI bubble seems largely ignorant of the mood shift underway, which suggests interesting times are ahead.
I speak to people who work at the upper echelons across various industries regularly and whether you want to believe it or not idc - the management are desperately trying to push AI but it just doesn’t add much value to what they do. At best it’s just a really good search engine across internal data. Many of these places already had things called macros in place so there’s barely any value add.
Even within SV there are still luddites who sometimes type out code, in mid 2026!!!
Although they know perfectly well what happened to their search / personal data, but they still don't want to see the obvious
There's a lot of overlap between people who cannot stand the AI boosters, don't want a data center built anywhere near them, are sick of the slop, and still use a chatbot for some stuff.
I don't think this is hypocrisy. I don't think it's a contradiction at all.
It suggests that people actually like natural language interfaces where they make sense and the price is reasonable. What they don't like is the rhetoric, behavior, impact on electricity prices, insistence on cramming it into places where it doesn't belong, layoffs, threats, and general obnoxiousness of the people pushing it and their general milieu.
Which makes perfect sense.
Both of those things did transform life & culture but mostly to the benefit of their makers. People now expect the same from AI and for better or worse most of the CEOs are not even pretending this time. The most they do is some vague hope that it'll all workout magically somehow.
FTFY the headline for you
A bunch of frivolous projects that fail sounds to me like a pretty good way to learn how far a new technology can be trusted.
If you're considering putting AI into something load bearing you either need a engineer who has not been participating so they can say "no" or one who has made 15 failed AI projects so they can say "maybe". The very worst case is to pressure somebody who doesn't know the technology very well into saying "yes".
ramon156•57m ago