On the danger of self-aware A.I.

[Scurrybt]

How real is the danger of A.I. becoming self-aware and an existential threat like Skynet from Terminator or the Cylons from Battlestar Galactica? Do you think the threat of A.I. is overblown, reasonable, or understated?

[Uncle Sam]

Well if it does reach Singularity aka Skynet or beyond level of exponential learning, it might as well start nuclear winter (aka nuclear war between the countries with the most amount of nukes atm) to get rid of most life on Earth, or use other methods with the same objective in mind.

[WaitandSee]

We don't have the technology to actually make "Artificial Intelligence" let alone make it self-aware. To create something that is actually capable of learning but also understanding why and how things work and why they work, the underlying principles, we cannot currently do.

We are at the limitations of silicon binary processing, we proper artificial intelligence would require the next generation of quantum computing in which I cannot see reaching any level of production in the foreseeable future.

[Scurrybt]

Sir Dookface McFerretballs eredeti hozzászólása:

Self aware A.I. doesn't exist, that we know of yet.

Hell, most stuff labeled as "A.I." isn't even "A.I." at all, but programmed "Machine Learning" instead which still has constraints and is limited or biased by its programming/programmers.

We are still on the way to actual A.I., but not there yet.

I'm more worried about the human world leaders currently in power leading us towards WW3 than I am of anything machine learning or A.I. related.

I agree with this

[Lying Sack Of Fun]

Sir Dookface McFerretballs eredeti hozzászólása:

Self aware A.I. doesn't exist, that we know of yet.

Hell, most stuff labeled as "A.I." isn't even "A.I." at all, but programmed "Machine Learning" instead which still has constraints and is limited or biased by its programming/programmers.

We are still on the way to actual A.I., but not there yet.

I'm more worried about the human world leaders currently in power leading us towards WW3 than I am of anything machine learning or A.I. related.

Idk I've messed around with machine learning a bit. It's practically evolution for code. In theory, if it was left to learn long enough it could become AI I would think. Depending on the original programming anyway.

[Lying Sack Of Fun]

Scurrybt eredeti hozzászólása:

Sir Dookface McFerretballs eredeti hozzászólása:

Self aware A.I. doesn't exist, that we know of yet.

Hell, most stuff labeled as "A.I." isn't even "A.I." at all, but programmed "Machine Learning" instead which still has constraints and is limited or biased by its programming/programmers.

We are still on the way to actual A.I., but not there yet.

I'm more worried about the human world leaders currently in power leading us towards WW3 than I am of anything machine learning or A.I. related.

I agree with this

I think I prefer Skynet to being forced to kill people being forced to kill me.

[Scurrybt]

lying sack of fun eredeti hozzászólása:

Scurrybt eredeti hozzászólása:

I agree with this

I think I prefer Skynet to being forced to kill people being forced to kill me.

Agreed

[Goldias]

If they develop that kind of AI without kill switch of emergency shutdown, they deserve their fate.

Real threat from AI is not they will make us slave, but they will make us broke.
In a decade or two most worker in this generation may be left unemployed because AI simply do their job better, while newer generation will have to change their education to suit the new AI society.

[zeke]

lying sack of fun eredeti hozzászólása:

Idk I've messed around with machine learning a bit. It's practically evolution for code. In theory, if it was left to learn long enough it could become AI I would think. Depending on the original programming anyway.

Its an apt comparison, but unlike the evolution of organic 'computers' (ie brains), it is important to note the evolutionary process is limited solely to software, to the coding running the bot. Sure the future might hold new upgraded hardware in the computer's future but this won't really differ in many, if any significant ways from a hardware/technological perspective. But for our brains to be capable of what they are today, they underwent significant physical changes guided by an evolutionary process. There's no real analogue for that in computers (at least, as of right now under the current concept of computers).

In other words, for a true evolutionary process capable of surpassing its current capabilities and exhibiting, what we might call, 'growth' as opposed to just 'progress', you would need the hardware also subject to quasi-random variation/reconfiguration. That probably isn't something that can sufficiently be provided for on a computer that runs its operations solely on electricity. For comparison, our brains make use not just of electrical signals but also chemical signals. Computers are, in a sense, missing a bit of pre-historic 'tech' that provides a much greater deal of flexibility. As a result, they are arguably 'technologically' eons behind the brains of animals in complexity.
--------------------
I don't think there's any amount of time, even if you gave them a trillion years, that you could leave machine learning bots to continue to develop on our computers (as they currently exist) that would allow them to evolve into something with sentience, with self awareness because hardware will still present a hard limit to actually expanding their capability. A bot will almost certainly remain just that - nothing more than a (albeit more impressive) bot.

[n00bz0r]

Cicada 3301 already have near human AI.

[n00bz0r]

Cicada 3301 is a science cult based in the mountains of Greenland.

[Scurrybt]

zeke eredeti hozzászólása:

lying sack of fun eredeti hozzászólása:

Idk I've messed around with machine learning a bit. It's practically evolution for code. In theory, if it was left to learn long enough it could become AI I would think. Depending on the original programming anyway.

Its an apt comparison, but unlike the evolution of organic 'computers' (ie brains), it is important to note the evolutionary process is limited solely to software, to the coding running the bot. Sure the future might hold new upgraded hardware in the computer's future but this won't really differ in many, if any significant ways from a hardware/technological perspective. But for our brains to be capable of what they are today, they underwent significant physical changes guided by an evolutionary process. There's no real analogue for that in computers (at least, as of right now under the current concept of computers).

In other words, for a true evolutionary process capable of surpassing its current capabilities and exhibiting, what we might call, 'growth' as opposed to just 'progress', you would need the hardware also subject to quasi-random variation/reconfiguration. That probably isn't something that can sufficiently be provided for on a computer that runs its operations solely on electricity. For comparison, our brains make use not just of electrical signals but also chemical signals. Computers are, in a sense, missing a bit of pre-historic 'tech' that provides a much greater deal of flexibility. As a result, they are arguably 'technologically' eons behind the brains of animals in complexity.
--------------------
I don't think there's any amount of time, even if you gave them a trillion years, that you could leave machine learning bots to continue to develop on our computers (as they currently exist) that would allow them to evolve into something with sentience, with self awareness because hardware will still present a hard limit to actually expanding their capability. A bot will almost certainly remain just that - nothing more than a (albeit more impressive) bot.

Really good analysis!

[eram]

chaos gpt made a twitter account to try and get support to destroy all humans

[Battlekruse]

The real breakthrough will be when you combind quantum computing together with artificial intelligence, then you will have unlimited calculations.

A other possibility is the idea of chaos computers (a supercomputer using organic human brains for processing power).

[Xero_Daxter]

I made an NTR story with AI. Pretty erotic.

[Dwarven Runesmith]

eram eredeti hozzászólása:

chaos gpt made a twitter account to try and get support to destroy all humans

Literately give human some robot booba and humanity will destroy itself within the hour.