Not really going anywhere with this one. I just read Ava’s post about the LLM circus at her work. I get Ava’s blog on RSS but in this case she blew up on Hacker News, for better or worse. (I use adblock to vaporize the Hacker News comments section because it always leaves me in a bad mood.)

Hey Ava—I have the same LLM circus at my work! It’s ridiculous. Everyone is being asked to show how they are using LLMs to make things more efficient and yet all of the use cases we have are either more inefficient or solving problems that don’t exist.

One of the shittiest use cases I encountered was an executive who asked the LLM team to build him an agent that would automatically generate thank you letters. Who wants an AI-generated thank you letter? Just send me nothing, at that point!! Or send me an obvious form letter instead of pretending to actually care.

I wish I knew where all this LLM stuff would end up. I’m not one of these extreme anti-AI types who thinks they are never useful or that they are always bad for the environment (that depends on what the LLM is replacing). But the good use cases are few and far between. And we’re already paying a steep price in terms of the social upheaval and workplace tensions caused by companies pushing this AI stuff way too hard.