The technology is promising, it’s just not remotely ready for what they’re trying to use it for, and may never be in its current iteration (transformer-based LLMs). Like, yes, an AI will probably eventually be able to read many articles from search and integrate that information together in a useful way, but right now it’s almost as likely to just start making shit up halfway through and tell you to eat glue lmao.

The problem is that AI is the new corporate buzzword like web was back during the dot com bubble. The web did end up being massively successful, but it just wasn’t ready for like 90% of what investors wanted from it back then.