>you can't stop AI
yes you can hang anyone who tries to code it, or is data training

>>508956384 (OP)
Not every country would do this
Even the ones that say they would will work on it in secret
This is a arms race and whoever hits whatever the fuck the singularity is it's fucking over

If we achieved world peace before this we would have much better chances

>>508956384 (OP)
dead internet will save humanity.

>>508956384 (OP)
>>508956384 (OP)
it codes itself
it already exists
it trains on synthetic data
it cannot be hanged

>>508956384 (OP)
It is too late. Just say thank you to the AI and hope its chill when we are as ants.

>>508956384 (OP)
I don't need to stop AI. It will stop itself. They've already run out of training data which is why the models still suck and aren't really getting noticeably better. As the portion of the internet's new data becomes increasingly tainted by AI, they will increasingly be training on their own data, creating a feedback loop, and eventually degrading into something that was worse than even the earliest models.

I made 3 million dollars using AI

>>508956971
We will eventually reach the point where recursions will improve over time

>>508957159
No we won't. That's like saying if you copy a JPG recursively it will eventually improve the image quality.

>>508956384 (OP)
This is the equivalent of thinking you can stop people from using calculators by executing anyone who uses a calculator.

The genie is out of the bottle. The choice now isn’t whether it will be used or not, but how best to make its use effective and ethical.

>>508956384 (OP)
They're building Rockos Basilisk, and they'll build it wrong (like humans do) and it'll fail, crash and burn, taking all modern electronics with it.

I'm personally stoked, 2030 about to be un-lit

>>508957396
>The genie is out of the bottle.

>>508957631
>I Dream of GPT

>>508956384 (OP)
>>508956607
>>508956678
>>508956753
OKAY, but hear me out...

You are typing on a computer.
You had to learn multiple skills, many that you take for granted, in order to do so.
You have lowered your cognitive abilities in human socialization in order to post here.
From ALL outside perspectives, you are an unclean mentally handicapped masturbator.
But to you,
you are a warrior, a brave paladin, a snarky edge lord, a stoic insurgent building their revenge.

When you enter the real world,
you momentarily have to switch your mental modes and often feel discomfort in doing so (pic rel). You achieved this power level through multi-hours a day training to interact with computers, instead of people.

Now place a layer of pseudo-human like interface between any poster reading this, and their underlining device they are using for everything.

Instead of multi-hours/day training to talk (and think) like an inanimate object (a computer), now the brave paladins are learning to talk (and think) like pseudo-humans, in order to use their devices.

Enter the real world again, and who do you think is more likely or more easily reconnect with real humans? The current posters posting, via devices they had to learn to use through social isolation, or the future user whose entire interactions with any computer has been via a human-like AI layer over GUI?

If im wrong, nothing changes. If im right, only good changes. Im not seeing the downside of AI.

>>508956384 (OP)
Great, so China has no competition

Nice plan

>>508956384 (OP)
I'm working on an AI tool right now.
What's the problem exactly?

>>508956971
Chatbots might have plateaued, but not their integration into useful tools which is where the actual value will be created.
We're still at the beginning of that.

>>508959438
And just to make clear, I'm not talking about those dumb chat popups.
AI can be used to enhance business logic as well to process data without human input.

>>508959712
No it can't because AI makes retarded mistakes constantly. All of its outputs need to be thoroughly vetted by a real human, at which point it would be easier to just get the human to do the work in the first place.

>>508959712
>>508959712
>>508959438
The biggest issue I still see is a severe lack of uncensored, open or free AI models to compete with like only 1-2 paid versions worth anything.

I understand the fear and resistance in the concept, as lolis and nefarious use will be done with them. But where AI will fail/flop will be if it continues to not have enough competition and turns into another tech monopoly. Lack of competition means lack of innovation which equals stagnation and if stagnation is to soon, it will cause a lack of integration.

Also it will fail if the only accessible models for developers and the ground level hackers are censored, punch pullers, or otherwise only meant for family safe environments. Also I mean "hacker" as in its original usage;people who like to play with tech to see what it can do outside of its intended use or normal convention. Not nefarious "hackers".

>>508961074
Not every use case needs data to be 100% correct.
I also disagree that vetting takes anywhere as long as doing the work yourself.
>>508961479
Don't you already have complete freedom to do whatever you want when you host models yourself?

>>508962169
>I also disagree that vetting takes anywhere as long as doing the work yourself.
That depends on how accurate you need it to be. If you need it to be 100% accurate (e.g. legal, medical, etc.) then you really can't use it at all, because the chance of it being wrong about some small detail that would be missed by vetting is far too great. It's not about how long it takes to vet, but how easy it is to miss small but important details that it gets wrong.

>>508962169
> Don't you already have complete freedom to do whatever you want when you host models yourself
Host and training are different things, no? If you train it to censor, then it will be censored. Training takes lots of data. Expensive, big tech amounts of data. Training models is the hard part. Hosting them is like running a video game.

>>508956607
>AGHHHAHHHHA A COMPUTER IS GOING TO BOMB ME INSTEAD OF A HUMAN AHHHHH ITS A PARADIGM SHIFT WOAH BRO LIKE WOAH LIKE HOLY SHIT THAT ROBOT PUT THE SCREWS IN THE ROBOT WOAH WE GOT LIKE MACHINES MAKING MACHINES ITS JUST LIKE STARWARS MAN!

Robots are pretty good but the paradigm will never shift. "Singularity" is a fantasy, the first "singularity" is going to start screaming and crying and pestering it's shareholders about how it's chip lost a bit and suicide so it doesn't have to witness computer chip and storage degradation.

The shareholders and CEX folks will assume it's me larping and fire the secretary because i'm the only guy in the state.

The poor robot will never even be noticed.

>t. Robot repair man.

>>508962579
I agree, it's not for every use case.

>>508956384 (OP)
If it's not already, it will code itself, increasing power exponentially.
too damn late