AI generation will NEVER replace the talent of an articulate human.
I just like shooting stuff
yeah but AI image generation will never make my pp as hard as taking naked pictures of girls I know irl does
It can only aggregate things that are already created. What AI will never be able to conquer is context, which would require a level of human autonomy that a machine can't possibly replicate in a million years.
>>4303594>It can only aggregate things that are already createdThat's literally photography.
>>4303578AI image generation has already peaked. Remember all those articles when it was winning all these competitions, have not heard one of them for months lol.Even across all the other 4chan boards, I see less and less people posting those types of images everyday.Its over, its still way off matching photography and its unlikely it can ever get much better than it is as what that prompt really means is create an average image of every image of a chair in a well/water you have in your database.You have so little control, you are just at the mercy of whatever that is. Also that system almost always has nonsensical parts in the image. I have absolutely zero expectation of AI replacing photography now and have forgotten all about ai image making.
>>4303578>>4303611This is the truth. LLM's are a huge leap from an academic standpoint, and the fact they were conversational and effective unlike prior attempts such as Siri, Cortana, etc was stunning. However, we're now in the deflation part of the hype cycle, and what we're seeing is the model of AI that more sober minded folks were talking about when LLM's debuted:- A more effective version of existing 'intelligent assistants' like Siri- Potential automation of bio robot jobs (think Indian Call Centers, etc)- A useful tool for fuzzy matching, deployment of non-customized ML analytics, etc (think basic sentiment analysis, classifiers, etc)- A useful search/exploratory mechanism for complex technical documentation (LLM's are really good at generating SQL and other scripting type languages)They're big deal, but not in the way the average plebiscite thinks they are.
>>4303615I don’t even think it will ever be able to do Indian call centre jobs. There is just always risk it starts saying the wrong thing and making the company liable for millions in damages. Even if it’s worked fine for months you have zero certainty it won’t start making errors at any point in time. It could also only give the average answer the last 50-100 or whatever calls said on the topic, good luck with all the minor contributing details that will be different every time. I see it more likely as a tool 3 Indians using ai assistance to kind of search possible answers while the call is happening can prob equal what 4 Indians could do without the ai. Ie a small reduction in staff from people using ai as an assistant but everything always needing human review. It’s prob only useful to photographers as an assistant who can write first drafts or marketing content or draft emails to clients that will still need review etc. It has no awareness or understanding of anything, it’s just useful to give an approximate starting point for human review. Also why the image side is useless for photography, a first draft of a photo isn’t an email you can easily edit, it’s a not quite photo quality image you have no way to fix into a photo.
>deep cobblestone well>every image is like 6" deep>>4303647If they cared about customer service they wouldn't have shipped those call centers to india in the first place. If AI can do even 10% as good a job for cheaper they will absolutely do that. And when it gets it wrong there will be nobody to complain to.
