Print

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

[Professor Lemur]

I looked down in the shower today. Now my night will be sleepless.

[GiglameshDespair]

Genitals aren't that terrifying, man. Unless there were teeth or something.

[Professor Lemur]

It was like my genitalia turned into a hairy bezoar.

[SirQuiamus]

It's called puberty, man.

[Professor Lemur]

You don't know man. You weren't there man.

[Graknorke]

https://www.google.co.uk/search?q=disposable+razor&tbm=shop

[Spehss _]

Man, this conversation changed quick, man.

[Professor Lemur]

I know man. Y'see my main man, I gotta occasionally shitpost to help things stay all fine and dandy. Man.

[GiglameshDespair]

Well, mission successful, I guess? Man.

[Eric Blank]

Professor Gorilla.

[Spehss _]

He'll get his PhD by writing a thesis about the economy eventually producing a gorillion dollas, at which point everyone will holla to get a dolla.

[i2amroy]

And there aren't necessarily any do-overs.

Also consider that intelligent AI would not necessarily be governable by the humans that made them.

Why wouldn't it be? We have 40 years or so where we will be doing bugfixing on AI's that are almost as smart as humans but not quite. I'd certainly hope we would be able to iron out problems like "develops an uncontrollable urge to destroy the human race" in the 40 or so years of beta testing.

I think you might be making a common mistake, which is assuming we will go current computers -> uncontrollable AI in a single leap. We won't, we'll slowly get smarter and smarter AI's as we develop the current ones, until eventually we get one that is smarter than we are (and even that probably won't kick things off immediately, since you not only need an AI smarter than we are, but also need an AI smarter than we are working on AI development). We're not talking about something some kid is going to randomly make in their basement some day, nor are we talking about some global defense program that stupidly put cutting edge machine learning in without any safeguards. Most likely the first smarter than us AI will be created in a lab designed for AI research (much like the ones we have today) that has been working for decades towards creating one.

[Professor Lemur]

I guess you can call it Lemurilla's Thrilla Paper in Teaching How to Holla for a Dolla Dolla Bill, Y'all. 

[Lagslayer]

And there aren't necessarily any do-overs.

Also consider that intelligent AI would not necessarily be governable by the humans that made them.

Why wouldn't it be? We have 40 years or so where we will be doing bugfixing on AI's that are almost as smart as humans but not quite. I'd certainly hope we would be able to iron out problems like "develops an uncontrollable urge to destroy the human race" in the 40 or so years of beta testing.

I think you might be making a common mistake, which is assuming we will go current computers -> uncontrollable AI in a single leap. We won't, we'll slowly get smarter and smarter AI's as we develop the current ones, until eventually we get one that is smarter than we are (and even that probably won't kick things off immediately, since you not only need an AI smarter than we are, but also need an AI smarter than we are working on AI development). We're not talking about something some kid is going to randomly make in their basement some day, nor are we talking about some global defense program that stupidly put cutting edge machine learning in without any safeguards. Most likely the first smarter than us AI will be created in a lab designed for AI research (much like the ones we have today) that has been working for decades towards creating one.

I'm not saying it's going to all happen at once, but that's also a problem. We can barely maintain the computers we have today, and our ability to do so is not even close to keeping pace with our need. This combined with the gradual onset and development of the AI means we may not even recognize there's a problem until long after it's been established. Small steps just don't grab that much attention. It's the problem of "just a little bit more won't hurt", until it catches up with you and it causes serious problems. It requires constant attention and I'm not confident it's getting it.

[Yoink]

Jeez, I skip ten pages (on my phone) and you guys are still going on about this?!


Anyway: had a nightmare that I slept in and missed my flight, woke up in a panic. Argh.