Nothing said by or about the AI industry can be trusted at this point because there are too many speculative bets swirling.
Until this AI speculation bubble deflates, nothing said about LLMs, chatbots, data centers, or anything related can really be trusted because people have placed big bets on it with insider info they don’t want the people on the other end of the bets to know about.
The big stories about AI so powerful it will find bugs and ruin systems, I believe it’s a decoy, a sleight of hand where you’re looking there and missing the really huge AI disaster waiting to happen: People using AI slop error filled output from LLMs in very serious situations, and that bad information pollution poisoning a lot of very serious projects and systems, while the humans in the loop set on vigilance tasks have, like all humans do, automation bias, the tendency to just blindly trust an automated system, and on top of that alarm fatigue, and at this point isn’t proofreading anything at all anymore, if they ever were actually bothering to fact check every bit of output by the chatbots. Would you use a calculator that you had to fact check every single calculation because 10% was wrong? Or does everyone accept this extremely high error rate and just stop bothering no matter how serious the consequences? And like they say, they’re putting this everywhere “whether you like it or not” so these errors are being haplessly deployed in your credit scores and in your medical records, it’s in our food supply, and it’s in all our phones and every surveillance system deployed everywhere.
There’s a bug in the system, that’s for sure; it’s the error-prone AI tech tools being forced into everything. The tool is the bug.
