Around 2023, the early days of AI and Google shoving it in everyone’s faces in their search, I was trying to look up how to get nail glue off of a phone screen because my coworker at the time, during her shift, was doing her nails at the desk and got nail glue on the work phone.

The first result, Google’s AI summary, suggested to “microwave the phone for one to two minutes.” I clicked the source and it linked to a forum comment by a guy who was clearly joking about how to clean the phone, but the AI presented it as legitimate advice.

Just got an AI generated PR from a coworker that implemented an entire microservice to replace the user id with username in one error message.

The kicker was there was already the username in the api method’s context. It was like 1500 lines, a new cache, a bunch of weird asynchronous thread code… and zero screenshot or proof it worked.

EDIT: Thanks to Iran’s refusal to allow reporters or international investigators to look at the debris of one school which was bombed recently killing about 170 children, we may never know with any certainty who was responsible, but it is highly likely that the USA bombed the site because the building was included in legacy documents and AI would likely not exclude such older intel.

Changhan Kim, CEO of the South Korean gaming company KRAFTON, decided he needed a way out of a costly acquisition deal, he asked ChatGPT.

It didn’t work.

Obviously this was only a $250 Million dollar mistake (and ruined what little reputation his company had) but if you want an example which results in potential loss of human life: UnitedHealthcare has spent the last few years automating claim denial.

The US Department of State has also been using AI to shut down federal programs which include the words “gender”, “binary”, and “inclusive” for being woke dei. As a result grants were denied for museum air conditioning and publishing mathematics papers.

Early adopters of AI included Airlines which resulted in lawsuits over refundable tickets, and drive-thrus resulting in the purchase of 9000x Water (free).

TBH, in my opinion, all LLM output is worthless slop, equally stupid, always the dumbest. The only way we quantify one being dumber than the other is by the idiocy of the user justifying their insane actions with it because it’s a machine which inherently causes psychosis in the dumb and feeble.

The headliner about flock cameras getting 71 percent of the stolen tag alerts wrong. You will have better luck flipping a coin on every sus car you see