Interesting Development on Artificial Intelligence - Facebook

[GoldMember]

Elon Musk may have been right, maybe Mark Zuckerberg's understanding of AI was 'limited'. Probably not so much, now.

https://www.inc.com/thomas-koulopoul...sk-bill-g.html

[PB]

Hmmmm .... maybe, but I for one am looking forward to the arrival of our AI overlords. Think of the free time!

[PNWbrit]

the arrival of our AI overlords.

They won't come here.. Server closet couldn't handle it.

[PB]

So ..... we reeaaallyyyy need them because we're clumsy and easily distracted, like I was saying.

[Timberridge]

To be honest, I'm more worried about the apes right now. And corvids.

[PB]

Wait a minute .....WHAT IF THE RUSSIANS GET A HOLD OF THIS AND ..... AND .....

Never mind

Russians already be doing' this FO'EVAH!

[MontuckyFried]

Meh. As long as they do this before they take over, then I'm happy:

[old goat]

I don't know whether or not AI poses a threat to humanity or not. It's certainly possible; it wouldn't be the first human invention to pose a threat to humanity--take nuclear weapons and the internal combustion engine for starters. What I do know is that "experts", like Musk and Zuckerberg, turn out to be exceedingly poor predictors of the future. Of course if Musk and Zuckerberg make opposite predictions, one of them will likely turn out to be right. The question is which one?

[PB]

Not necessarily - pretty sure there's more than just two possible outcomes.

[The Suit]

No, Facebook's Chatbots Will Not Take Over the World

[neufox47]

What were the chat bots saying to each other?

Sent from my SM-G930V using TGR Forums mobile app

[The Suit]

What were the chat bots saying to each other?

Sent from my SM-G930V using TGR Forums mobile app

Bob: i can i i everything else . . . . . . . . . . . . . .
Alice: balls have zero to me to me to me to me to me to me to me to me to

[old goat]

Not necessarily - pretty sure there's more than just two possible outcomes.

There are many possible outcomes--but they boil down to 2 basic ones. A) Computers will become independent of our control and take over the world, and B) they won't. I guess there is a C) computers will become independent of us and decide to chill on the beach. The point is that your guess is as good as mine is as good as Elon Musk's and Mark Zuckerberg's. ( See Tfhinking Fast and Slow, by Daniel Kahneman--an economics Nobelist--on the inaccuracy of expert predictions of the future.)

Personally I think we're not smart enough to create machines smarter than us.

[neufox47]

Bob: i can i i everything else . . . . . . . . . . . . . .
Alice: balls have zero to me to me to me to me to me to me to me to me to

So you are saying we can't even crack their code!?

[Elkhound Odin]

Bob: i can i i everything else . . . . . . . . . . . . . .
Alice: balls have zero to me to me to me to me to me to me to me to me to

Computers verbalizing their sex? oh god oh god oh god

[Tippster]

Elon Musk may have been right, maybe Mark Zuckerberg's understanding of AI was 'limited'. Probably not so much, now.

https://www.inc.com/thomas-koulopoul...sk-bill-g.html

Article was corrected.

[JimmyCarter]

There are many possible outcomes--but they boil down to 2 basic ones. A) Computers will become independent of our control and take over the world, and B) they won't. I guess there is a C) computers will become independent of us and decide to chill on the beach. The point is that your guess is as good as mine is as good as Elon Musk's and Mark Zuckerberg's. ( See Tfhinking Fast and Slow, by Daniel Kahneman--an economics Nobelist--on the inaccuracy of expert predictions of the future.)

Personally I think we're not smart enough to create machines smarter than us. And Kahneman would say that my guess is as good as Musk's or Zuckerberg's.

Kahneman's point is that experts can be biased and trapped by fallacies like anyone else, but that's not equivalent to saying every guess is equal.

[PNWbrit]

What were the chat bots saying to each other?

Ambition makes you look pretty ugly
Kicking, squealing Gucci little piggy
You don't remember
You don't remember
Why don't you remember my name?

[GoldMember]

Article was corrected.

Thanks, Tipp:

"Correction: The original version of this column was based on inaccurate reporting and hype that has been debunked. The column inaccurately portrayed the outcome of a Facebook chatbot research project and the company’s reason for ending it. That bots use their own language is expected, not unanticipated by experts. The project was ended not out of alarm but because the researchers hadn’t designed them to communicate in a way comprehensible to humans, not bot shorthand. In addition, the earlier version also incorrectly said that the reports on this research broke after Facebook CEO Mark Zuckerberg and Tesla CEO Elon Musk engaged in a public debate over AI; the debate came after the news reports."

[old goat]

Kahneman's point is that experts can be biased and trapped by fallacies like anyone else, but that's not equivalent to saying every guess is equal.

His research showed that experts were wrong more than 50% of the time--flipping a coin, which you or I could do, would do better. Best example is the well-known fact that index funds do better over the long run than actively managed funds. The reason that experts do so badly in predicting the future, despite knowing more about a subject than the average person, is that they cannot anticipate all the things that might happen in the world that would affect their prediction. How many experts predicted the outcome of the last election? Because who predicted Russia hacking the election or Comey's last minute bomb or that Hillary would abandon the northern midwest? (Obviously there were other factors that were predictable and were ignored.)

To put it another way--experts do well when outcomes can be calculated by the rational part of their brains based on a predictable set of facts. When there is not a precitable set of facts experts make their predictions using the emotional part of their brains, the part of the brain that we all use the great majority of the time, and experts are just as vulnerable as the rest of us to their biases, as you say. At that point the part of their brain that makes them experts is turned off. One of the reasons that most experts missed the Trump victory is that he is repugnant to the kind of people who do most of the political predicting, so they missed the signs that pointed to a Trump victory.

BTW--if machines get too uppity can't we just unplug them? Isn't that what happened to Hal?

[Skistack]

Ambition makes you look pretty ugly
Kicking, squealing Gucci little piggy
You don't remember
You don't remember
Why don't you remember my name?

Sounds like a Morrisey song.

[Game Over]

Many more notable people are concerned about the dangers than not. And let's get real Zuckerberg is a moron compared to the likes of Musk, Hawkins, Gates and countless others.

[PB]

Why so pessimistic?!?! They could all turn out to be really swell fellows.

More likely they'll be just as crazy messed up as we are, cuz ...... <image of God and all>

I would like, however, to get my request in right now that they avoid at all costs recreating DD and BF.

[old goat]

I can't think of anything more horrifying than being a conscious intelligence with no physical agency--like being quadraplegic. Maybe we should ban AI not for our sake but for the sake of the machines.

[PB]

Your wet-ware bias is showing. If they're as smart and powerful as everyone fears, they'll just manipulate everything at the sub-atomic level using reverse gravitonic midichlorotic femto-bots, allowing themselves exist in a constant state of orgasm, thereby removing themselves from the mundane.