Chebucto Regional Softball Club

@futurebird > I don't know if this makes anyone else upset, but it's very annoying to me.

It's not just you.

LLMs making equivalent fsckups of basic info in other sciences (especially astronomy and chemistry, where I have fairly extensive background) is something I found intensely obnoxious from the very start.

Literally the very start. One of my first 2 or 3 prompts to ChatGPT back in 2022 or so was an astronomy question that it completely butchered in a manner that was close to jaw-dropping.

I won't go into extra detail here (I post egregious examples here on Fediverse every so often), but this phenomenon has made me dislike this tech from the beginning. Even if it has some utility in other areas (and I believe it does, albeit massively overhyped), it's not something I feel I can be comfortable using for anything meaningful.

@futurebird My employer is dealing with this now.

A bit over two years ago: my data engineering team determines that a rather niche 3rd party database product, already in use by another team, is a logical choice for my team as well, since the two teams will likely need to share data in a way that the database is a good fit for.

About 1 year ago: get email that the 3rd party database company is bought by an "AI"-focused company.

About 1 year ago minus 10 seconds: *groans internally*, knowing what will probably happen soon.

About 2 months ago: new company stuffs "AI" features into the database service and doubles the price for the "AI"-enabled service. There is no "old service, hold the AI, please" option. Since my company is gigantic, it is able to negotiate a one-time 12-month extension of the current DB service, but it's not going to be able to do that indefinitely.

4 hours ago: Discover all of this in a meeting, and get annoyed at having to contemplate completely throwing out a bit over 2 years of work to get something similar to what we were using just fine back then.