Post #2501984
2026-05-13 16:27 UTC
Replies (7)
-
@TheRealPomax@mastodon.social 2026-05-13 16:34
@sue@glasgow.social "it's a really hard problem. If only there was some kind of... something, I don't know, that would have let us compare numbers to other numbers, but unfortunately that does not exist so no one could have predicted this"
-
@shimst3r@famichiki.jp 2026-05-13 16:34
@sue@glasgow.social hello, I have a certificate in FinOps for AI, hire me market! :fingerguns:
-
@ainmosni@social.ainmosni.eu 2026-05-13 16:46
@sue@glasgow.social "It's a really hard problem" Is it? Is it really? Seriously, they REALLY don't want to employ people, even if the end result is worse and more expensive. I'll solve the "really hard problem", although anybody reading it here would find it obvious. Stop using the problematic expensive slopmachine, and just hire competent people.
-
@401matthall@mastodon.xyz 2026-05-13 16:46
@sue@glasgow.social ... They haven't even _raised_ the rates to cover their _costs_ yet. Other people's compute is _expensive_.
-
@shanecelis@mastodon.gamedev.place 2026-05-13 16:59
@sue@glasgow.social "If your AI agents aren't trading crypto on the side, you're leaving money on the table to be left behind!" /s
-
@djl@mastodon.mit.edu 2026-05-13 17:57
@sue@glasgow.social I love all the articles that say "It only took Claude 14 minutes to do X and 20 to do Y", where there is absolutely no concept of the vast amount of money 14 or 20 minutes of AI datacenter compute would actually cost if you actually had to pay for it. LLM-based AI is stupid on a plethora of levels. I need to copyright this: "The LLM idea is the stupidest idea in the history of computer science."
-
@drgroftehauge@sigmoid.social 2026-05-13 18:04
@sue@glasgow.social Companies both measuring employee performance on how many AI tokens they burn through (more is better) and whining about blowing their budget on AI tokens...