How far are we from AI recreating EQ99?

[loramin]

P.S. check out Steve Yaegee's Gas Town. It's basically what you want (throw gas on the fire of current tech) ... and it's barely a thought experiment, let alone something that can produce working software.

[BradZax]

Originally Posted by loramin [You must be logged in to view images. Log in or Register.]

No, you're still not understanding. Let's say you pay someone to level your toon. Each new level takes more time, so it costs more money. Up to 20 you can pay them for a level a day, but after that they take multiple days to level.

You can get a new job that pays more, and then you can afford to pay for faster leveling. This is current AI.

But now you want a live toon instead, max level, with all AAs. It doesn't matter if you pay your player more, they still can't get you a level 100 with all AAs. It doesn't matter how much money your player gets, he can only play so many hours a day.

That's getting an AI to make something like EQ. Throw the very best models at it, give them their own personal nuclear reactor even ... they still won't be able to do it. New tech is required, not just better-powered existing tech.

Right now you can generate an image on google gemni, it uses recursive learning and LLM models.

You can also generate an image on one of their diffusion models, these do not use recursive learning or LLM models.

The recursive memory generation costs 10$ a second of wattage to generate.

The problem isn't "this is as much wattage as technology allows us to pipe into it"

It's that, "currently it costs 10$ to generate that much wattage."

So, they don't allow users to use the full version of their LLMs or they would go bankrupt in processing costs.

Every question, causes the LLM to turn on compactors that need to be cooled.

It could turn on more compactors now, but they don't have the power to cool them.

If overnight we built enough pwerplants to generate enough power so that it costs 10 cents to generate 1 second of video, you would see massive improvements in LLM results that you thought were impossible, overnight.

The new version of chat GPT and shit, thats trying to do MORE with less. Optimizing their limited model to function as efficiently as it can with current power limitations.

But the thing they are investing in, is capable doing more, with MORE!

[DeathsSilkyMist]

Originally Posted by BradZax [You must be logged in to view images. Log in or Register.]

Right now you can generate an image on google gemni, it uses recursive learning and LLM models.

You can also generate an image on one of their diffusion models, these do not use recursive learning or LLM models.

The recursive memory generation costs 10$ a second of wattage to generate.

The problem isn't "this is as much wattage as technology allows us to pipe into it"

It's that, "currently it costs 10$ to generate that much wattage."

So, they don't allow users to use the full version of their LLMs or they would go bankrupt in processing costs.

Every question, causes the LLM to turn on compactors that need to be cooled.

It could turn on more compactors now, but they don't have the power to cool them.

If overnight we built enough pwerplants to generate enough power so that it costs 10 cents to generate 1 second of video, you would see massive improvements in LLM results that you thought were impossible, overnight.

The new version of chat GPT and shit, thats trying to do MORE with less. Optimizing their limited model to function as efficiently as it can with current power limitations.

But the thing they are investing in, is capable doing more, with MORE!

Let's run with your idea for a moment. If we built 100x the datacenters overnight, genie 3 still wouldn't be able to make Everquest. The tech just doesn't work in a way that is condusive to persistent online worlds with thousands of players. Maybe genie 3 with 100x datacenters would allow you to play a basic singleplayer game for a few hours.

[loramin]

Originally Posted by BradZax [You must be logged in to view images. Log in or Register.]

Right now you can generate an image on google gemni, it uses recursive learning and LLM models.

You can also generate an image on one of their diffusion models, these do not use recursive learning or LLM models.

The recursive memory generation costs 10$ a second of wattage to generate.

The problem isn't "this is as much wattage as technology allows us to pipe into it"

It's that, "currently it costs 10$ to generate that much wattage."

So, they don't allow users to use the full version of their LLMs or they would go bankrupt in processing costs.

Every question, causes the LLM to turn on compactors that need to be cooled.

It could turn on more compactors now, but they don't have the power to cool them.

If overnight we built enough pwerplants to generate enough power so that it costs 10 cents to generate 1 second of video, you would see massive improvements in LLM results that you thought were impossible, overnight.

The new version of chat GPT and shit, thats trying to do MORE with less. Optimizing their limited model to function as efficiently as it can with current power limitations.

But the thing they are investing in, is capable doing more, with MORE!

Did you read what I wrote about EQ? You think that because you can pay some kid to get you from level 3 to 4 that money = leveling ... and it does ... to a point. Even if level 40 takes a lot longer than 4 you can (probably, depending on the kid) pay them more to make them spend the hours needed to get you to 40.

But if you try to make that kid farm you a live toon with max AAs, the kid will go off to college before he finishes. You can throw ten times as much money per hour at him and it won't change a thing.

It's an analogy, but that's the gist of it for AI too. If an AI can (extending the analogy) get you to 60 with today's tech, then yes, throwing more power at it might get you to 60 with 1 AA. Even more power (a lot: remember, diminishing returns) might get you 2 AAs.

But if you want all AAs (ie a working EQ), even with all the power in the world, current LLMs could never do it. But don't try to argue with me: go learn more about how this amazing tech works yourself!

[Goregasmic]

It is all fun and games in theory but right now it costs like 10B$ to build a powerplant over 10 years and pretty much all of these companies have zero path to profitability. Funding will dry up before they get online. AI doesn't have much social acceptability right now, not sure how much funding they'd get from public money either. There's already huge clashes over data center water usage for what gain exactly?

[BradZax]

Originally Posted by Goregasmic [You must be logged in to view images. Log in or Register.]

It is all fun and games in theory but right now it costs like 10B$ to build a powerplant over 10 years and pretty much all of these companies have zero path to profitability. Funding will dry up before they get online. AI doesn't have much social acceptability right now, not sure how much funding they'd get from public money either. There's already huge clashes over data center water usage for what gain exactly?

Well in foundation the anti ai people killed all the pro ai people and all the ai.

The anti ai crowd is getting pretty far left behind IMO but the reactionary karens that want to protect children from deep fakes are still clinging to their paranoia.

I genuinely think we're at the 70s and some of us like computers and some of us hate them.

[OriginalContentGuy]

This is like Nexii's Erotic Revelations thread except with Brad. 'Nuff said.

[OriginalContentGuy]

Respect to OP though for sharing

[BradZax]

Originally Posted by loramin [You must be logged in to view images. Log in or Register.]

But don't try to argue with me: go learn more about how this amazing tech works yourself!

Just curious.

Why do you resort to like this "im smarter than you, your argument comes form a lack of knowledge" approach?

[Goregasmic]

Originally Posted by BradZax [You must be logged in to view images. Log in or Register.]

Well in foundation the anti ai people killed all the pro ai people and all the ai.

The anti ai crowd is getting pretty far left behind IMO but the reactionary karens that want to protect children from deep fakes are still clinging to their paranoia.

I genuinely think we're at the 70s and some of us like computers and some of us hate them.

I think the average person looks at this and all they see is shitty image generation, hallucinations, broken google overview, more sophisticated scams, increased workers/citizen monitoring, deep fakes, disinformation and we'd have to trade our local power/water for this? And we aren't even talking about the more nefarious uses like defense departments using it to dodge accountability but that's probably beyond what the average person notices I guess.