General Gaming News and Discussion

joz123 said:

Damn guess they gotta find another job that gives them a ton of money to show up in pajamas and slippers

Click to expand...

All I hear is "I'm shit and I like being fed shit."

Every AI voice acting tech we've heard so far is garbage, and even if it gets to average it'll still just be average, never good. It will be used to make worse games so future Bobby Koticks can capture more margin for themselves. Many of the best games are the best games because they're carried by extremely strong vocal performances.

Captain Suave said:

Will the quality continue to improve? Yes.

Click to expand...

Actual vocal quality, maybe. Acting ability? Unlikely. Emotional quality? No.

All these tools are is averaging engines.

Pyros is correct. While I think that we'll see a few studios pull off something amazing in the next decade with these AI tools, for the most part, GenAI is the new goyslop. Companies like Microsoft and EA are gonna try to force-feed that slop into the average gaming consoomer because what choice do they have? Not buy new AAA video games to fill their empty incel lives?

Captain Suave said:

Of course they will, if for no other reason that someone will train one on higher quality reference material.

Click to expand...

That's not how the math works lol.

If you refine the training set to a subset of higher quality content, you make it worse at generalization.

And an AI is never going to tell the director "Well, I think this line would read better like this..."

This fantasy that you can remove the human element and still have content that connects on a human level is both a dark one and a dumb one.

*hands you an oat bag of goyslop*

popsicledeath said:

algorithmic goyslop is good actually, because people don't deserve better

Click to expand...

ftfy

Hateyou said:

I agree it doesn’t need to connect on a human level. Every person that brings up AI to me is blown the fuck away by and absolutely loves it. They don’t care how it works, they don’t care it has minor inconsistencies, they don’t care it has blatant lies. They just fucking love it.

Click to expand...

And that alone should tell you that it's actually slop.

We are entering a world of Peak Garbage. Please do not cheer it on, or else be pronounced another Funko-Pop-Buying consoomer.

Captain Suave said:

I'm not sure what in the history of human technology makes you think that these tools in their infancy represent the maximum potential, but dollars to donuts you're going to be very badly mistaken.

Click to expand...

Linking this again.


TL;DR - 70 years of AI research has demonstrated that trying to make AI 'smarter' via clever tricks and cognitive hacks produced effectively 0 results. People are under the mistaken assumption that this new batch of AI is new tech, but Deep Learning is actually a return to a very old approach that just wasn't feasible at the time. "The Bitter Lesson" basically says that the only way to make anything resembling 'smart AI' is to just throw a ton of data and compute at the problem--aka there are no shortcuts to smart. The reason the latest round of models work as well as they do is because they collected basically all of the coherent data they could for the domains they cared about, then shoved it through a tremendous amount of compute to distill it.

To build models twice as smart, they'll likely need ~20x the data and ~20x the compute. The compute is 'easy' so long as we continue building chip fabs and avoiding a war over Taiwan. But there's very unlikely to exist 20x the high-quality data you'd need, unless humans keep producing new, high-quality data.

Do you see the trap here? As the internet gets more filled with mediocre goyslop, it becomes harder to build future smarter models using this method, with a high probability of leading to a doomloop of mediocrity as the models starve on their own shit. And 70 years of research has shown that none of the shortcut methods work.