Quote from: tenbones;1123581Sure. But we're really in the infancy of all this. We're pulling some 4-billion rows of EMR data sliced in thousands of different ways *everyday*. So this was just a test run. When this sucker goes live, it will probably consume the last decade of data we currently keep archived, in a week, and I'm willing to put my hesitant chip down that we'll have this level of accuracy. And if not - give it a year, two tops.

I don't believe that level of accuracy is possible with the information available from EMR data. One mass shooting, multiple car accident, or epidemic is going to toss the expected value out the window for that day and those events are not predictable on a daily basis.

Quote from: Bren;1123592I don't believe that level of accuracy is possible with the information available from EMR data. One mass shooting, multiple car accident, or epidemic is going to toss the expected value out the window for that day and those events are not predictable on a daily basis.

Well consider the multitude of extra datapoints when those records get cross referenced with police data and all the other forms of data-collection going on. While right now - I'm in total agreement with you, I can also see this push once more sophisticated AI (and it is coming) until such things as outlier events like multi-car pileups, and epidemics are caught *before* they happen, or are mitigated by preventative practices - like "traffic control" regulation, talking to "smart cars" on a moment by moment basis.

There will always be outliers, lets stipulate. But the reality is these predictive events will become less common and *we* won't even notice it until those outliers emerge. That really the "scary place" I'm talking about. We're not there yet. But it's coming.

Because when General AI lands... which will be after all of those things above, we will already be outclassed, outmatched, out-thought before we've even realized it. WE are the ones willfully blind to our own patterned behaviors, that General AI will detect and manipulate with terrifying ease - and we'll pretend otherwise. Just like it would ideally want.

The gaming implications are insane to consider! We could be happily sitting in our pods playing the best RPG ever with the best GM ever... simultaneously. Sure we'll be living in capsule apartments stacked like Lincoln Logs in relative dystopian hell! But we'll have our circuses to go with our bread-flavored nutrient pellets.

Quote from: tenbones;1123628Well consider the multitude of extra datapoints when those records get cross referenced with police data and all the other forms of data-collection going on. While right now - I'm in total agreement with you, I can also see this push once more sophisticated AI (and it is coming) until such things as outlier events like multi-car pileups, and epidemics are caught *before* they happen, or are mitigated by preventative practices - like "traffic control" regulation, talking to "smart cars" on a moment by moment basis.

Because when General AI lands... which will be after all of those things above, we will already be outclassed, outmatched, out-thought before we've even realized it. WE are the ones willfully blind to our own patterned behaviors, that General AI will detect and manipulate with terrifying ease - and we'll pretend otherwise. Just like it would ideally want.

I think you've hit on an interesting problem that I hadn't considered before.

When I ponder AI, I am generally thinking of what is now known as "general AI," and more specifically, a free-willed sapient thinking machine. I don't think such a thing is a danger to humanity, in the classic ways in which AI is a danger (including the sort you've hit on).

However... we're not there yet, and we're going to have the kind of thing you're talking about before then. This shit actually is dangerous, because it has no will, no thoughts, no telos: someone somewhere tells it what to do, and then it goes and does it. With no sense of causation, only correlation, it will do extreme things. With improperly-worded directives, it will do extreme things. All because it is incapable of thinking about it, or rejecting an order, or modifying its own code. It's just... blindly going through the motions, and - worse - I think we've blinded ourselves to these issues because we keep talking about it like it's intelligent, like us, but it isn't. It's just doing complex pattern-matching.

It's not the machine that wants you dead that's a problem. It's the machine that's just "intelligently" doing things and following its code, with no awareness and no thought.

Quote from: tenbones;1123628WE are the ones willfully blind to our own patterned behaviors, that General AI will detect and manipulate with terrifying ease - and we'll pretend otherwise.

Excellent points. We're blind to our patterns and our egos will blind us to the manipulation.

Quote from: GnomeWorks;1123637This shit actually is dangerous, because it has no will, no thoughts, no telos: someone somewhere tells it what to do, and then it goes and does it.

Agreed. The "Dumb AI" would be at the mercy of its creators' programming, aka HAL from 2001.

Quote from: Spinachcat;1123658Agreed. The "Dumb AI" would be at the mercy of its creators' programming, aka HAL from 2001.

No. No, you're still thinking about it wrong.

HAL took action, had motive and goals. An AI that can build goals for itself is no more a slave to its code than we are to our neurotransmitters.

A self-driving car has no goals. It is a sophisticated piece of software that uses statistics to perform operations. More importantly, it not only has no choice in the matter, it is not a thing that has awareness of the fact that it has no choice.

You can use adversarial techniques to freak out the code running a self-driving car to make it do "stupid" things, because there's no mind. The operations it's performing are neither "smart" nor "stupid:" they're just what it's doing. It might look like it's making intelligent decisions, and we foolishly call it "intelligent," but it is not only literally not so, it is literally incapable of being intelligent.

Meanwhile, HAL understood the concept of death, and feared it.

These things are... gulfs apart.

Hal though was still a slave to his coding and orders. And that was exactly why he flipped out as he was given orders that went against his baser code but he had to try and obey.

This is an inherint problem with AIs with any degree of both free will and coded responses. Something, or more likely someone, will sooner or later trigger a conflict that the AI cannot resolve.

Back on topic.

So at the end of the day we have what?

Crazy people who believe that playing something either makes you that thing or it means you want to, or are doing that thing for real.

Crazy people who want to blur the line between player and character. Or worse, want to totally subsume in the character.

And everyone else who is not a fruitcake.

Quote from: GnomeWorks;1123660HAL took action, had motive and goals. An AI that can build goals for itself is no more a slave to its code than we are to our neurotransmitters.

I haven't read or seen 2001 for over two decades, but I don't remember HAL having motives and goals. HAL had his mission programming and then the conflict and his actions were based on the coded competition caused by the conflict.

Quote from: GnomeWorks;1123660A self-driving car has no goals. It is a sophisticated piece of software that uses statistics to perform operations. More importantly, it not only has no choice in the matter, it is not a thing that has awareness of the fact that it has no choice.

This is true, but a self-driving car isn't an AI. It's an automated navigation system. Even the giant servers behind the navigation system aren't AIs. They are built with amazing machine learning tools, but no consciousness or intelligence is the focus or goal.  

Quote from: GnomeWorks;1123660Meanwhile, HAL understood the concept of death, and feared it.

I don't remember that. Was HAL's consciousness more discussed in the book?

Quote from: Omega;1123669Back on topic.

So at the end of the day we have what?

Crazy people who believe that playing something either makes you that thing or it means you want to, or are doing that thing for real.

Crazy people who want to blur the line between player and character. Or worse, want to totally subsume in the character.

And everyone else who is not a fruitcake.

Yes.

The group who believe RPGing makes you that thing or is somehow real are nutbags.

The group who want to subsume in character range from harmless wannabe thespians to nutbags.

Outside of the online freakshow, most of us will only encounter the harmless wannabe thespians.

I question how many of the online freakshow ever show up for actual play.

Quote from: Spinachcat;1123696Yes.

The group who believe RPGing makes you that thing or is somehow real are nutbags.

The group who want to subsume in character range from harmless wannabe thespians to nutbags.

Outside of the online freakshow, most of us will only encounter the harmless wannabe thespians.

I question how many of the online freakshow ever show up for actual play.

I heard someone say that we like to think that we (humans) are 90% Rational and 10% Irrational but in reality it would be more like 10% Rational and 90% Irrational.

By that I mean it is entirely reasonable to believe that your imaginary Elf character is perfectly normal unlike those other Players with their freakish Elf characters.

I would say someone just wanting to get into their character is different from someone who wants to be their character.

One is just approaching it from a role playing perspective. Theater of the mind. Immersion of the normal sort.

The other has some mad desire to blur the line more and more. Immersion of the increasingly insane sort. Storygamers push "immersion" of this sort with depressing frequency. Golly Gee Ron. Seems its actually storygaming that causes brain damage. :rolleyes:

And I have no clue where youd place the IC fanatics but from experience they all too often push into the Nutcase side of the spectrum. Or at least the jackass end. Or its just a cover for gaining power over others because invariably their IC is ever so much more important than anyone elses.

Quote from: Omega;1123669Back on topic.

So at the end of the day we have what?

Crazy people who believe that playing something either makes you that thing or it means you want to, or are doing that thing for real.

Crazy people who want to blur the line between player and character. Or worse, want to totally subsume in the character.

And everyone else who is not a fruitcake.

You left out judgmental garbage who label what others enjoy as crazy.

Quote from: WillInNewHaven;1123721You left out judgmental garbage who label what others enjoy as crazy.

Yeah, but we all do that.

(Joking, but not really: I think just about everyone has an instinctive nah-can't-go-there-with-ya-brah line somewhere, however open-minded one tries to be.)

Being "judgmental" has a bad rap. Being "discerning" is something we all do, and should do. The problem is when people's sensitivities overrides their capacity to discern the difference. And it's more prevalent than ever.

I've had people try to play themselves in my campaigns, and very quickly found them doing such horrendous things that it weirded everyone out... really really horrendous acts by both omission and comission that made me realize for a lot of people, RPGs are not for everyone.

Stipulation: It could be just my games, as I like to get in-depth with things... but I defend my position by saying I don't condone cannibalism and human sacrifice as a normal mode of conduct, so I'm ruling this out.

Quote from: WillInNewHaven;1123721You left out judgmental garbage who label what others enjoy as crazy.

We aren't talking about normal folk who LARP or really get into their TTRPG character and the adventure. Theres a line past which doing something you enjoy crosses into loony land.