/sci/ - Science & Math

[Anonymous No.9970487]

Do you think artificial general intelligence will be created within your lifetime? And if so, what do you think the implications of it will be?

[Anonymous No.9970612]

Yes. Im sure we will have prototypes within 10 years. AI weak will be the first example. With low capacity. Then it will go on and become stronger and better and better until AI strong. Strong AI is maybe decades away.

[Anonymous No.9970616]

Yes, we have natural General unintelligence now! Can’t happen soon enough!

[Anonymous No.9970628]

>>9970487
It is spelled General AI.

[Anonymous No.9970643]

>>9970628
No it isn't https://en.wikipedia.org/wiki/Artificial_general_intelligence

[Anonymous No.9970692]

>>9970487
It'll be bad for scientists.
One of the few jobs that customers don't care if its done by a human or not.
Also doesn't require high dexterity.

[Anonymous No.9970716]

>>9970692
But if AI is ever capable of doing human-level scientific work, it will surely do it at dramatically superhuman speeds because the processors are already operating at the gigahertz level whereas our neurons can only run at 200hz, tops; so the AI would be completing thousands of years of human-level scientific work for each actual year worked.

[Anonymous No.9970724]

>>9970716
if it happens and we can control it then it will make everything that isnt physical labor obsolete. And if you assume that robotics doesnt stop on its current progress trajectory then AI linked to robotics will make all physical labor obsolete too.

Then its either share the goods and services produced by the AI/robots and have some kind of utopian, post-work economy, or only the "owners" of the robots/AI get the goods and services they produce and everyone else dies.

[Anonymous No.9970808]

>>9970724
>Intelligence automatically comes packaged with an ego
This meme needs to stop.

[Anonymous No.9970907]

>>9970716
>2045
>put the ultra-AI hypercomputer to work on entry theoretical physics
>instantly prints 14 books worth of incomprehensible postulates and theories that need a team of joint team of elites to even make sense of the first five words
>spend the rest 50 years making sense of it all

[Anonymous No.9972176]

My lifetime? No. Not saying something like that is impossible. But We have a long way to go before We have something that could be considered an AGI. I don't get people that say We'll have it in a couple of decades, I don't see it.

[Anonymous No.9972209]

>>9970907
We'd just automate engineering as well.
Then just tell it to notify us if it ever figures out ftl.

[Anonymous No.9972258]

>>9972176
Agreed. Still, ANI like we have now is still an incredibly powerful tool by itself. Even without AGI the world could potentially be very different 20-30 years from now.

The only way I could see AGI happening in our lifetimes is if another Cold War breaks out and governments start pumping massive resources into R&D again.

[Anonymous No.9972265]

>>9970628
>Breaking news: Biggest leap in technology in the 21st century, GAI has been created
lol

[Anonymous No.9972272]

>>9972265
They changed the acronym so that it wouldn't sound like "gay"
I fucking hate /sci/

[Anonymous No.9972275]

>>9970487
Yes, within 10 years tops

[Anonymous No.9972292]

You'll lose your STEM job

[Anonymous No.9972299]

>>9972275
Ha. We are decades out from AGI.

[Anonymous No.9972327]

Deep Learning and neural networks capable of filling out CAPTCHA'S, yes. True Artificial General Intelligence? Not by a long shot.

You can only hope by the time those come along that humans have genetically engineered themselves to be a little better, because you ain't gunna have a job. True Artificial General Intelligence will come just before Artificial Super Intelligence.

[Anonymous No.9972425]

>>9972275
More like 2.

[Anonymous No.9972449]

https://arxiv.org/pdf/1705.08807.pdf