>>4
This. I am genuinely perplexed as to why everyone has been so fucking captivated by poorly written paragraphs with predictable and clearly recognisable sentence patterns and shitty generic images for the last 3 years. Large-scale application is a different conversation entirely, but in terms of the average end user It will never make sense to me.

>>4
i tend to agree, but my single biggest problem is the corperate datamining it does. with self hosted that's not a problem, all that's left is the toy. people are definitely using it to make the world more shit, but that has been true of all new tools. email used to spam ads, gunpowder used to kill, etc. it's possible to use it benificially, but the clear and predominant usecases are all shit, it's easy to go off the rails if you don't approach it skeptically, and the personalities of the corpo models are literally my least favorite type of people.

I don't use AI, but I do incorporate images generated by users of AI (if I think they look sufficiently good) into my own images ('art' is an annoying concept when my intent is to entertain myself.)
>>5
When those AI materials are used properly they are not recognized as such which is why providing an example of "good AI" is impossible. I've seen people really like something until they find out it is AI, meaning that the context matters more than the result.
>>4
I find AI being everywhere annoying, but I also find the "AI = bad" crowd equally annoying; counterculture alt-normies moved from IRL subcultures to Internet subcultures (we must admit that there is a trend towards these types of attitudes, as accurate or "well-intentioned" as they may be.)

>>4
>>5
LLM usage is only as prevalent as it is because corporations are shilling it as the solution to labor and endlessly investing into each other so as to keep that illusion alive. Once the true ineffectiveness of "AI" forces the trillions of dollars invested into this toy to crumble under its own weight, this retarded economic bubble will pop.