Message from @KawaiiSamurai

Discord ID: 542934508610256906


2019-02-07 05:00:42 UTC  

This is known.

2019-02-07 05:00:45 UTC  

not only that, but the models for constructing AI and testing it has really come from those proto-AIs interating with human players in those games

2019-02-07 05:01:02 UTC  

@Fitzydog maybe, it's just a fun conspiracy theory

2019-02-07 05:01:19 UTC  

Yeah, kind of like the Joe Rogan 'cocoon theory' lol

2019-02-07 05:01:36 UTC  

But without video games, I don't see what industry would have spurred the proliferation of parallel processing

2019-02-07 05:01:43 UTC  

Look if we honestly did accidentally figure out AI, we made it too retarded to be viable. We can't even figure out why WE are intelligent.

2019-02-07 05:01:53 UTC  

It wouldn't have happened nearly as quickly through scientific pursuits

2019-02-07 05:02:06 UTC  

It was probably all inevitable

2019-02-07 05:02:15 UTC  

Everything is.

2019-02-07 05:02:26 UTC  

but that's just one example

2019-02-07 05:02:32 UTC  

What will happen, will happen.

2019-02-07 05:02:46 UTC  

Well, anyways, the one on one discussion was good, but I g2g

2019-02-07 05:04:02 UTC  

there's also a lot of interesting new things being developed in the next iteration of programming languages, which have marketed individual uses, but combined have a lot of utility towards machine learning models outside of the current generation of neural networks

2019-02-07 05:04:27 UTC  

so it's going to be very interesting to see where that goes

2019-02-07 05:04:34 UTC  

Fitzy is hungry for more trap cock

2019-02-07 05:04:50 UTC  

Machine learning is sort of shit though.

2019-02-07 05:05:15 UTC  

It's impressive at first I think.

2019-02-07 05:05:17 UTC  

Yes! Which is why it's all so exciting, because there's so much potential work to be done

2019-02-07 05:05:44 UTC  

But then you realize... Oh. It's just luck of the draw over and over again.

2019-02-07 05:07:37 UTC  

No a lot of predictive models are very successful

2019-02-07 05:07:54 UTC  

they are integrated in everything you do online

2019-02-07 05:08:37 UTC  

I'm not talking about predictive models.

2019-02-07 05:09:31 UTC  

well they use machine learning to make predictive models and offer suggestions

2019-02-07 05:10:39 UTC  

it's the #1 killer app that it is currently being used for, from commercial stuff to STEM stuff

2019-02-07 05:10:46 UTC  

By the way, machine learning that gives us shit like Alexa, can really infuriate me. Because you realize, by Alexa's responses, thatshe either A) adaptedto judge you as retarded, or B) so many retards use Alexa that she has learned to be retarded from them.

2019-02-07 05:11:29 UTC  

yeah the problem is that when the suggestions become too influential, the tail begins to wag the dog

2019-02-07 05:11:32 UTC  

Always irritates me when Alexa cannot do a simple fucking search for me for my literal words without citing Wikipedia on something totally unrelated.

2019-02-07 05:11:34 UTC  

but welcome to politics

2019-02-07 05:11:50 UTC  

most people are robots anyway

2019-02-07 05:12:28 UTC  

I hve to think, "How do retards ask questions?" And then THAT QUESTION goes into the feedback loop too, haunting me in the future.

2019-02-07 05:13:40 UTC  

well the fundamental problem right now is that these models are all relative, and therefore suffer from a lack of creativity

2019-02-07 05:13:49 UTC  

Look. As long as we can have some sort of Asimovian laws working, I say, DO NOT let AI learn from us. They will learn our worst possible traits.

2019-02-07 05:14:30 UTC  

Because, smart people, don't communicate nearly as much as dumb people.

2019-02-07 05:14:39 UTC  

That simple.

2019-02-07 05:14:50 UTC  

I don't think it will be any different than what we deal with now

2019-02-07 05:15:05 UTC  

you might as well also say that dumb people shouldn't vote, because they will corrupt the system

2019-02-07 05:15:29 UTC  

which btw is not to say that is a wrong thing to believe

2019-02-07 05:15:35 UTC  

They shouldn't learn from each other, but it's too late to stop that.

2019-02-07 05:15:44 UTC  

We haven't made AI yet.

2019-02-07 05:16:52 UTC  

I do think that uninformed people shouldn't vote, but that would require a basically perfect unbiased system of vetting, that could never be trusted to actually work.

2019-02-07 05:16:59 UTC  

Well my definition of intelligence applies to more constructs than just bio and tech, so for e.x I think a government or corporation can develop an intelligence if it is complex enough