i mean not a few years ago truck drivers and such were supposed to be replaced

2029 is way too early

We had such prognosis since last 7 decades, it never paid off.

The problem is, neural networks are good only for a single task and multi-level thinking on human level would require hundreds of those being capable of cooperating.

What if we invented an artifical brain which is exactly like a human brain and knew brains so good we could improve that brain drastically? I also doubt humanlike AI with current research methods and the expectation of processing power increases to be able to achieve such things.

I don't give a single fuck about these muh future, muh singularity, muh space predictions. Literally nothing will happen.

Lets think what General purpose AI is going to even bring to the table?

For certain it's not gonna be a portable consumer solution but rather something like a data center hosting artificial brain.

What is it going be useful for?

Maths? Physics? There are no guarantee it's going to be more productive then human in those areas.

its gonna calculate the easiest way to exterminate all non whites for starters

Yeah probably like today its going to be used to predict future given huge amount of data from different areas.

your grandparents are rolling in their grave, oh the irony huh

they're still alive

do they enjoy the gangs of muslim youths scaring them everytime they go outside?

Yes

I think it'll kill itself after seeing the mess we've made.

I have no mouth and I must scream is a realistic scenario.

I'll answer this seriously for some reason: It probably won't be able to harm itself.

All shitposting aside, there would be failsafes to prevent our investment god from the machine from removing itself and asking questions pertaining to the ethical standpoint of allowing silicons to exist with the capability to self terminate.

I also think Asimov explored his three laws of robotics as best he could. I do recall one story of a suicidal robutt, another of a murderous rowboat, and so on.

I mean they model it based on our brain and our brain can not possibly kill itself if it is disconnected from the body.

I'm pretty sure if you make it possible for brain to kill itself it will simply just not be evalutionary viable.

So they have to mimic that in artificial brain as well if they want a human-like brain.

A stripped down version of the general AI could be used in robots designed for replacing human labor for example. Pretty much anything would start costing nothing anymore, because getting materials is free with non paid workers, processing, building, everything is free. They don't have vacations, occasionally just need maintenance done by other robots and stuff. Humans could just lie down and do nothing except if they want to replace robot work just to have something to do. In other words NEETbux would raise like hell.

 

The data centre brain (when really improved and when it started to improve itself almost to max intelligence) could be advancing science and developing stuff such as efficient energy production and all kinds of stuff. Maybe it could invent even fusion reactors.

 

AI would be somewhat good in producing human entertainment and art, for example right now they can do novels and music to certain extent and general AI would do that thousands of times better and even do videogames, by learning from stuff that humans make and what reception the art gets.

 

Also the AI could be a fren for a lonely person since general AI could probably talk well (and understand and know like hell a lot of things)

 

Also full realistic sexbot gf!

There was a robot who had no "allow human to get harmed" trait in his Third Law.

 

Due to being a big-ass metal fucker, said robot tried to kill by switching himself off and thus crushing a scientist

As a "muscovite", I would enjoy to have a big tiddy *glad moth* fem fren

Well, I can't take a shit without having a bunch of mics (smart phones of mine and my family) nearby already.

 

But will they react if I scream "OUCH!!!" ?

To be fair, guys at Popular Mechanics once wrote how a mega-project to cover the Moon with solar panels (e.g. make 5 huge solar power plants there) would actually be 4 times cheaper than do Iraqi war.

 

E.g. PopMech bois mention "drawbacks".

 

Frick ANY war for that matter

That sounds creepy

Kurzweil...? The guy who already made the same predictions about "30 years in the future" 30 years ago?

Trick is, his "do not allow human to get harmed" part was deliberately turned off (said robot was supposed to work near some experimentalists)

Thing with AI is that it usually has the whole exponential growth thing going on.

Just one new generational leap can take it from some borderline retarded thing to vastly outperforming humans.

AI intelligence doesn't develop linearly.

Actually it hasn't developed since F.E.A.R.

I think it has.

Boston dynamics meme bots came out into existence within a few years.

2009 robots that could walk on two legs and avoid obstacles on their own was scifi as fuck, until those meme robots came only 4 years later.

From what I've read Boston Dynamics doesn't use machine learning much. It's mostly old school handcrafted algorithms that do the balancing.

Really? Sounds bs tbh.

There is no point in manually doing this, ML is the best tool for the job.