Post

Hot take: I don't like calling LLMs "AI" or image/video/audio generation as "generative AI" Because it mystifies the underling technology as smart in a way that it isn't.

An "AI that sorts your emails for you" might *sound* cool and futuristic to these marketing execs but all I need it to do is mark everything coming from an "@kde.org" address as "work" which is a thing I can do with a single if statement and a regex

I also would prefer "I added AI to my game" to continue to mean "I wrote a finite state machine that makes it so enemies in the game world move around in a convincing way" instead of "I wrote the story with chatgpt"

I also do like the idea of experimenting with generative technology (which is probably a better name for it than "generative AI"). Imagine an RPG where all the characters talk to you with dynamically generated, realistic, human-sounding language instead of pre-written scripts.

An LLM could be useful for that. What else could I use? A markov chain?

But I'm worried people would misconstrue "experimenting with generative technology" with the unethical and legally-dubious use of "AI"-generated artwork, writing, or software.

Or the unethical and bandwidth-consuming scraping of information done by "AI" companies to feed into their borg or sell to advertisers.