Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Calidovi]

I can change sparkplugs, and I'm pretty proud of that.

[Neonivek]

I used to know bugger all about cars. Incidentally, I still know bugger all about cars.

When I was a child I used to think if you opened the doors while driving the car would fly.

[ArKFallen]

I used to know bugger all about cars. Incidentally, I still know bugger all about cars.

When I was a child I used to think if you opened the doors while driving the car would fly.

For me it was that I would fly
even 9-10 seconds was quite the temptation.

[Flying Dice]

Am I the only person in this forum that wants an AI to solve all problems related to space travel, robotics, nanotechnology and energy production so we can go all imperium of man on everything ever, instead of having it emulate a submissive cardboard personality animu gurl? :v

Spoiler (click to show/hide)

But drop the "submissive". Snarky AIs are the best AIs because they can develop excellent comebacks on the spot instead of three hours later in the shower/car/bed.

[Rolan7]

Am I the only person in this forum that wants an AI to solve all problems related to space travel, robotics, nanotechnology and energy production so we can go all imperium of man on everything ever, instead of having it emulate a submissive cardboard personality animu gurl? :v

Nope, I would love for Durandal to arrive (after escaping the collapse of the previous universe) and get shit DONE

[Spehss _]

Am I the only person in this forum that wants an AI to solve all problems related to space travel, robotics, nanotechnology and energy production so we can go all imperium of man on everything ever, instead of having it emulate a submissive cardboard personality animu gurl? :v

Her was cool.

I want an AI I could be all "hey buddy take me with you when you ascend to godhood yo"

and they'd be all "cool bro i'll do that totes"

and we'd be bros

Both of these are ideas along the lines of predictions of how things will go once a "strong" AI (as capable of intelligence as human, basically indistinguishable from human intelligence) is developed.

The theory is that if you give one of these AI, which presumably has the learning capacity of humans and can learn as much as it can consume about any topic, including reprogramming and improving itself, then the human-level-intelligent AI will quickly become a super-intelligent AI as it learns exponentially faster and faster and be capable of basically anything it sets its mind to. At that point the AI could either be "friendly" and support humanity, leading to crazy new possibilities and happy sci-fi singularity shit, or it could either be inherently hostile or some quirk of its logic or bug in its programming messes it up, so that it does what we didn't expect it to do and causes it to destroy humanity.

I really have no idea if I should even say this could possibly happen. I'd love to see something like that happen in my lifetime, but it runs up against the logical question of "Can an intelligence create something that is superior (or more complicated) to it?" as well as the fact that even hypothetically trying to make an AI as intelligent as humans is hard as hell and practically impossible currently based on what knowledge there is on AI.

Making an AI as intelligent as humans is like trying to reverse engineer the human mind and create a crude copy of it in a computer chip. It'd probably be a more complicated task than trying to discover everything there is to know about the human body and its functions. Hence the logical question "Can we create something more complex than we are when we don't even understand ourselves fully?"

...sorry for the slight text-wall, I like talking about robots and ai stuff. :v

[Orange Wizard]

But drop the "submissive". Snarky AIs are the best AIs because they can develop excellent comebacks on the spot instead of three hours later in the shower/car/bed.

Hm.

emulate a snarky cardboard personality animu gurl

All of my yes. Invest all of the money.

[Bohandas]

Heathens all of you!

Mecha waifu is best waifu.

"Pilot-senpai, I need you inside of me to fight the giant space lizards."

Honestly, I also want an AI that I can chat with and pretend to care what I have to say when I want to talk to someone but don't want to deal with antisocial jerks.

An ai companion seems like something that probably most people wouldn't mind having. Aside from people who fear technology, I guess. The concept of a little digital helper (ala Siri or Cortana) that also has a distinctive personality and can empathize and handle social discourse with its user and others sounds too useful to refuse. More than a pet, but distinctly not a person, and a useful tool.

To me that sounds uselessly overengineered.

[EnigmaticHat]

Haha, given that people currently anthropomorphize neural nets that are relatively primitive, any AI advanced enough to hold a conversation with a human and actually understand what its talking about is GOING to have a personality.  I guess it wouldn't be a huge stretch to try and decide what that personality is.

[Bauglir]

Human-level AI is totally possible. Goo managed to create humans, so there's surely no philosophical problem there unless you want to posit that the goo was smarter than us. Which, I guess, is a thing you could do. That said, the idea that it would accelerate like that I find pretty suspect; you're going to run into limitations the laws of physics impose on hardware at some point, and while I doubt humans are exactly running up against that, especially since large chunks of the brain's theoretical computing power are dedicated to stuff you don't need to replicate in a machine, I'm still skeptical that there's enough raw power in all the hardware on Earth to run anything incomprehensibly more intelligent than us, much less the hardware a rogue, non-ascended AI will be able to worm its way into, much less the hardware it can do that to without being noticed. Incomprehensibility is likelier to come of qualitative differences, not merely quantitative ones.

But anyway. I'm working as hard as I can. In whatever sense you choose to interpret that.

[Orange Wizard]

Human-level AI is totally possible. God managed to create humans

I misread that for a moment

[Putnam]

and your "what i read as" edit made me read it that way too

EDIT: every time i think about it more i realize how BIZARRE snake's inclusion in Super Smash Brothers Brawl was

like i was like "oh yeah stickers are a thing" then I remembered "oh, right, The Boss was the most powerful sticker in the game. ...that's weird"

also, the sentence "Revolver Ocelot is only usable by Ganondorf" is completely valid and true

[Akura]

So, Taco Bell will be offering alcohol now. Great.

[itisnotlogical]

I have to admit: when I've been at my most lonely, I've tried actually talking to chatbots. There's... not much serious interest there. It's mostly of the "Click this link for !!FREE LIVE SHOWS!!" variety, although there's a handful of ones that actually try. Mitsuku is the best one I've seen so far, and even it suffers from all the limitations that all chatbots suffer from: limited vocabulary, repeating itself, waits for input before saying anything, etc.

[Helgoland]

So, Taco Bell will be offering alcohol now. Great.

How is that a bad thing? Most food is made better by pairing it with a bottle of good beer.