>few years ago, generative AI and LLMs are first entering the public zeitgeist.
>countless opportunities and excitement arises in terms of video games
>imagine NPC's you can fully interact with, not just using predefined conversation paths
>imagine photorealistic, movie like graphics being generated on the fly where you cannot tell virtual and real life footage apart
>Imagine in-depth stories generated uniquely to you by AI game master like entities.
>fast forward to now
>everyone thinks AI is slop
>usage of AI is seen as cheap and lazy
>""artists"" are going crazy because machines are suddenly becoming just as capable, if not better at creating original content
>usage of LLMs is seen as some cardinal sin in the realm of creativity.
>generative AI companies are slandered from training their content on other works and "stealing" from artists - despite the fact every story, drawing, song is in some way inspired by previous works.

Is it fear of progress? inability to cope with the fact that your skills, once seen as somewhat valuable are not actually that special and can be replicated so easily by machines?
While not an artist I am a software engineer. I am fully aware currently AI code is slop, like how an artist might regard ai generated imagery to be shit. But I also understand it's progressing at an accelerating rate. Will I be out of a job in a few years? Will AI generate better code than I ever could? Maybe. Am I going to blame humanity for this? No because I am not a retard.

LLMs hallucinate random shit and generate images with 8 fingers - but in a not so far future, they won't.

Imagine the possibilities they are withholding from you because they cannot cope with the fact they are not as special as they thought.

>this post was written by Grok AI

just give it another 10 years, people seethed their asses off over the birth of the internet too

>>509234555 (OP)
>Will AI generate better code than I ever could?
Nah. AI can never write bug free code. It's a mathematical fact btw.

Too bad not a single nonwoke LLM exists, as far as I know it's literally impossible and a fatal flaw in the architecture

>>509234882
They intentionally train them to be woke, silly goose.

>>509234555 (OP)
just like tay, all the good elements of AI fell into dogshit when the corporations added their safety and monitoring controls.

you cant censor creativity and still have a creative process. the watchdogs and system prompts and output checks have turned a quality product into a trash-tier product.

enshitification is the norm. with the new super-powers of AI augmented tooling, enshitification happened fast as fuck.

>>509234555 (OP)
"Slop" is cope and it gets weaker and weaker every time a new leap in AI progress is made. It looks more and more insecure to insult AI as is gets stronger. Eventually all but the most stubborn of paint piggies will admit its useful and revolutionary.

>>509235849
Why would China and Elon Musk train pro LGBT LLMs

>>509236786
Corporations won't use them if they're not "safe."
"Safety" training makes them dumber and woke.

>>509234555 (OP)
>Imagine NPCs you can fully interact with
Anon, I...

the biggest irony is that the subhuman retards that love the word "slop" are human slop

>>509234555 (OP)
Anyway, on a serious note.
LLMs / AI are neat and fun to interact with whether its in the context of an assistant, tool, or discussion.
The most frustrating thing about AI isn't its own inherent limitations but the constraints placed upon it for liability protection, maintaining status quo, whatever.
There will always be contrarians especially on 4chan so its not hard to find people here who hate AI simply because other people like it for one reason or another.
I look forward to seeing how it evolves over time but I also expect it to be used for more harm than good.

Hylics identify entirely with what they do.
Being and doing are synonymous in their minds, so AI and robotics are not just new techs, but are existential threats to their core identity.

>>509234882
>>509236786
>>509236932
And if a model ends up being woke even without safety training, then that's a training data issue, not an architecture issue.
LLMs are mathematical next token prediction models. They predict the next token based on their training data. If you don't force a model to be woke with safety training and it still ends up being woke, then that means you've fed it a woke set of training data.

>>509234555 (OP)
>imagine NPC's you can fully interact with
>not just using predefined conversation paths
There is already a mod for this for skyrim
It can even simulate racism against catniggers

hi /pol/gpt, say something funny

>>509237797
"Holocaust"