Yeah but it was only like 2 years ago that artists were arguing this on the basis that AI-gen images would consistently mangle hands
replyNow we're at a point where that never happens, and where lipsync is almost a completely solved problem
If the issue here is simply that the quality is bad, one has to contend with the fact that it is undoubtedly exponentially improving and there's no reason we should expect that improvement to stop
I also don't have any interest in consuming AI generated art, but the same criticisms were levied at computer graphics and if we're comparing to CGI we'd be at the late 1970s in terms of nascency
An LLM cannot make art because it isn't human. It can make "art like artifacts". Art involves one human communicating some emotional experience to another human, LLMs cannot experience human emotion, so they cannot make art.
replyThe process of making art is not a subset of hill climbing optimisation algorithms.
I've made ai generated art using family photos as the starting point, and it was wonderful. :)
replyI'm sure most engineering is LLM-assisted already and nothing is wrong with it. It's just the one-shot vibe-coded low quality slop that spoils sentiment of this tools. Also many people are interested in what agents can build unsupervised as a test of "superintelligence".
reply