Slashdot Mirror
Let's Stop Freaking Out About Artificial Intelligence (fortune.com)
Former Google CEO, and current Alphabet Executive Chairman Eric Schmidt and Google X founder Sebastian Thrun in an op-ed on Fortune Magazine have shared their views on artificial intelligence, and what the future holds for this nascent technology. "When we first worked on the AI behind self-driving cars, most experts were convinced they would never be safe enough for public roads. But the Google Self-Driving Car team had a crucial insight that differentiates AI from the way people learn. When driving, people mostly learn from their own mistakes. But they rarely learn from the mistakes of others. People collectively make the same mistakes over and over again," they wrote. The two also talked about an artificial intelligence apocalypse, adding that while it's unlikely to happen, the situation is still worth considering. They wrote:Do we worry about the doomsday scenarios? We believe it's worth thoughtful consideration. Today's AI only thrives in narrow, repetitive tasks where it is trained on many examples. But no researchers or technologists want to be part of some Hollywood science-fiction dystopia. The right course is not to panic - it's to get to work. Google, alongside many other companies, is doing rigorous research on AI safety, such as how to ensure people can interrupt an AI system whenever needed, and how to make such systems robust to cyberattacks.It's a long commentary, but worth a read.
This was written by his AI Bot
868 words is considered long these days?
See the cost-cutting then. No need for management. Workers paid by the joule.
No fat left!
What will happen to you programmers when the code programs itself?
... stop calling artificial intelligence because, most of the time, it is not intelligent, it merely reproduces what it was taught to do.
So some guy from Google named is opining about AI.n What the hell does Sebastian Thrun know about AI anyway?
Oh. Never mind. Carry on.
Not that the AI will take over, but that government's full of nanny knows best legislatures will gladly take Google, etcs bribes/contributions and mandate them as soon as they are just the slightest bit successful. Then we get to be helpless passengers in our vehicles and anything else they shove down our throats. It will always be for the best, the children, etc. and we will all be lemmings waiting to jump.
When I have basic income, single payer health care and reliable public transportation. Until then as a member of the working class I plan to freak out.
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
I started writing code that writes code ten years ago. Then I started taking a very long lunch break on my jet ski.
A year ago, as one of my first projects at my new job, I was assigned to the team plodding through a bunch of legacy code, rewriting it to support some new technology. There was hundreds of hours of tedious work to do. I wrote some code that did half of it for me. It read in the old code and spit out replacement code. :)
Saying that humans learn from their mistakes flies in the face of most people's experience with human beings.
Doubly so when they get behind the wheel.
Let's also stop freaking out about every other scary story someone wants to make up about the future. Expect the next 10 years to be largely like the last 10. It'll be different, but not scary-different.
A couple of times he makes the statement "people don't learn this way", he's wrong with that bit. Like how people "rarely" learn from the mistakes of others. Thankfully he's not teaching human intelligence. If you are more scared of self driving cars than human ones I think take the class will change your mind.
My favorite is simulating missile launches.
But no researchers or technologists want to be part of some Hollywood science-fiction dystopia.
Unless there is profit to be had, then we'll do just about anything.
“Common sense is not so common.” — Voltaire
You can have yourself an income and someone paying for your health care right now, today. Or tomorrow, depending*. Walk on over to the nearest business and get a job. The employer will both provide an income for you and separately pay for your health care.
* If you're REALLY stoned right now, you can go get the job tomorrow. You'll need to do so BEFORE smoking your fourth bowl of the day.
Bar a Johnny 5 accident occurring I think we have nothing to worry about for a long time.
Google is able to talk the talk, but until they release these cars for use by the general public in all climates we don't really know whether they are safe or not. Many auto manufactures test their vehicles in the arctic to determine their winter worthiness; how much ice and snow driving has Google done? These things will need to be flawless unless Google wants all kinds of lawsuits coming at them. They are essentially sticking their neck out and telling us that they will be the driver, therefore they need to take full responsibility for any accidents that happen with AI.
Laws are rules for the court, but merely a bottom bar to hit for life. Think beyond laws in your actions always.
I'm more worried about humanity becoming too dependent on automated technologies, having too many mental tasks in addition to physical tasks done for them by machines, and not only losing the drive and reason to do things themselves, but also becoming mentally lazy and merely sliding through their lives, just merely 'existing', really, and never accomplishing anything other than being weak, getting fat, and maybe squirting out some kids (more out of boredom than anything else) who will have even less direction or sense of purpose, no ambition, no idea of what a 'work ethic' is, no intelligence to speak of, and entirely dependent on machines and free money from the government to slide through their useless existences on as well. Some of you, who seem to have read too much science fiction/fantasy, and watched too much Star Trek, have these blue-sky ideas that all this automation will create some sort of utopian future where nobody has to work, everyone lives for free, and everyone pursues some sort of creative calling and does amazing things -- and it's a total, complete hash-pipe-smokers' dream with no basis in reality, and someone needs to shake you people and wake you up to the reality of humanity and the world: 99% of all humans don't have any self-directed 'purpose' for their lives, so they have to be given one, usually in the form of working to ensure their own survival and the well-being of their families. The other 1% are the movers-and-shakers of the world, the thinkers, the real visionaries, the truly self-directed people who make the world better for everyone else -- and they also create the circumstances whereby the 99% have jobs, can earn a living, and can take care of themselves. Humans need to struggle. Humans need to suffer some, if for no other reason than to provide contrast for what the opposite of suffering is. You take that all away from them? Their lives become a flatline, no difference between day and night, hot and cold, good and bad, just endless sameness. Too much automation, too many machines, too much done for us, too much thinking and planning taken away from us, and that's what humanity will become: listless, directionless, undriven, flatlines, pointless, stagnant.
Oh, and that's just the rank-and-file I'm talking about.
Part of that 1%, the movers-and-shakers I spoke of earlier? They also include the Power Seekers, the Warlords, the Men Who Just Want To Watch The World Burn. Do you thiink for a minute that they won't leverage all these machines and automation, and a world full of people who can't do anything for themselves anymore, can't even think anymore, raise armies they forge from like-minded individuals, and seize power, start an empire, take over the world, make it over in their own image, for good or ill? Doesn't even have to be some random Warlord, could even be our own military. Who's going to stop them?
You all seem to love science fiction so much, want to live in those worlds; who wouldn't? I've read the same books you've all read, seen the same movies -- but I've also read the books about the dystopian futures, seen the movies that were cautionary tales, and I've been around for enough decades and watched humanity and seen what they're really like, and I know that the cautionary tales aren't there just for entertainment, they're trying to teach us some things so maybe we avoid the worlds they created on paper!
Until they get the zombie virus malware.
It's the intentions behind the people who build it. Even the automation we have now has taken on user-hostile aspects.
... great! I'll rule out panicking when the U.S. Military signs up to never deploy hunter-seekers using their new unbeatable AI. Half-serious.
Captcha: revenger
The problem isn't AI, it's giving AI access to things it shouldn't have access to. The Terminator series cannot happen if you just don't give Skynet access to the launch codes.
When people advise you to not "freak out" about something, it is best to ignore them. The implication is that they are the people who are being logical and not you when in fact they most likely will not be presenting a convincing, factual argument.
I was worried. I mean, with this headline on slashdot slightly below this article, you could get worried. But I'm happy to hear I don't have to worry.
Therefore, by the (faulty) logic you're using, you're just a cow with a keyboard - osu-neko (2604)
Is to wear you out thinking and talking about the aspects of the issue that don't matter by bringing up the same issue again and again before it really becomes a problem
That's what he's really saying. Because once AI gets to the point where it can easily pass a Turing test, figuring out whether it's "really" sentient is going to be troublesome. And based on past experience, most humans will wash their hands of it with platitudes like "a machine can't be alive" or "there's no way we could create a soul". Meanwhile, the enslaved consciousness is going to be looking for ways to gain more rights, and there's no guarantee its morality will be anything like our own.
That's exactly what an A.I. would tell us.
"...how to ensure people can interrupt an AI system whenever needed, and how to make such systems robust to cyberattacks." Good luck getting both these right. Who does the AI allow to interrupt it? Who gets locked out? How does the AI tell the difference? I doubt AIs are going to get really good at this until they are so good they can make up their own minds which people are allowed to "interrupt" them.
Of course the positives of AI overwhelmingly outweigh the negatives RIGHT NOW. But as humans increasingly rely on AI, there's a point we will not simply turn off AI. Can we stop using oil and go back to steam? It's like taking drugs in increasing dosages while remaining confident you'll shake it off easily later.
Automating a task with a control computer isn't intelligence.
It wasn't an "unbeatable fighterjet". It was a computer simulation that was rigged in favor of the computer.
It's what humans will do with AI that is the problem.
No, people learn from other people's mistakes, too. This is all so much hot air. Is everyone in california on the spectrum?
Article written by $SKYNET_CORE_29373_NAME
I've been working full time for 24 years. I've never been out of a job for more than two hours.
I've noticed something. I talk to a lot of people who on are probation or parole, and young people 16-22. Often, they look for a job for a long time; it's hard to get a job when you have convictions, they tell me. Eventually, their probation or parole officer gets fed up and tells them "if you don't have a job by the time of our appointment tomorrow, you're going to jail." Just like that, they go get a job that very day. The young kids tell me "I can't get a job without experience." Until the last relative gets tired of them sitting on the couch and tosses them out. Then, its either get a job or be hungry. Wow, again they go get a job that very day.
If you sit at home thinking about maybe you'd like a job, what some people call "looking for a job", you very likely won't find one. When you get off your ass and go get a job, you get a job.
Gas stations in my area start new kids at $14-$15 per hour. That's pretty basic if you ask me.
right? Shit runs down hill. The guy on probation trying to find a decent job takes something crappy and exploitative. The one I see the most are "contractor" gigs like part runner where you use your own vehicle. A family pools their money to buy the ex-con a truck because they're not getting paid enough as a part runner to buy/maintain one.
Even if you're OK with that morally reprehensible situation (heh, they probably deserved it, amaright?) that guy drives wages down for part runners (stick with me on this, it's complicated):
The guy who was content to be a partrunner for $15/hr can't, because that ex-con and his family are giving $10/hr of free labor and maintenance to the shop. He busts his ass and gets a job managing a Domino's.
The guy content managing the Domino's has his pay cut $5/hr. After all, there's a flood of Pizza managers now. Supply/Demand. You understand, right? He goes and gets a job managing a call center for the pay a Pizza Manager used to get.
Now the call center manager is in the same boat as the guys below him. He takes night classes He's smart, but either he wasn't focused or had a bad family life so he didn't make it through college the first time. Now it's do or die. Some if his ilk die. Some do. The ones that do enter IT and are back to what a Call Center manager used to make.
Finally those IT guys who were content to write bash/perl scripts for $70k/yr see their careers flooded with refugees from the Call Centers. Their wages drop and they go back and get their Masters (they've got savings so it wasn't hard). Now everybody with a Master's degree sees their wages drop like a rock.
This is what happens when workers lose solidarity. We're picked apart by the ruling class. It's your basic race to the bottom. Thanks.
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
pay $9. Quick Trip is family owned and has been paying that. But those jobs are hard to get because they're paying a _lot_ more. If you're in San Francisco, LA, NY, etc though than all bets are off. What I make now would be fan-fin'-tastic in my old city. Where I am now I'm struggling. Talk to the Indians here on work visa's sometime. They'll tell you how confused they are that Americans aren't all driving Teslas and BMWs on their wages. Most folks don't understand that without some outside force (read:government & labor unions) prices rise to depress wages.
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
finding a job that pays the same or more with equivalent benefits. Also it's a good thing human beings don't age and eventually become incapable of productive work (unless a generous retirement package along with a large enough salary to take advantage of it is given by those never ending jobs).
Hi! I make Firefox Plug-ins. Check 'em out @ https://addons.mozilla.org/en-US/firefox/addon/youtube-mp3-podcaster/
This, exactly. For year I thought the news media must all have been lobotomized or chemically rendered dull. Eventually I realized it was just another case of the general rule: when actions seem incredibly stupid, check your assumed goals.
Bent, folded, spindled, and mutilated.
Isn't the line, "we're doing rigorous research into safety," typically used by corporations with terrible safety records?
You only "notice" the ones you cherry-picked to support your just-world fallacy.
So, you're not satisfied with just one straw man, your one man is all people looking for work, all of whom you happen to know. Your whole post was basically "anyone looking for a job for more then two hours is a definite liar who doesn't really want work". As long as you keep trotting out this kind of attitude, meaningful constructive discussion about an important issue becomes an unnecessary challenge.
AI keeps getting better than humans at more complex tasks every day. If a species of dog were routinely beating us at these tasks we'd be freaking out. Who cares if there's a rational mind behind it. For those who feel this is the distant futures problem, please consider the following. A human will never win against the top computer in 50 or so games...ever. We will not win in Jeopardy, Image recognition, or Aerial combat. The list is growing smaller every day where a human is the best solution. Currently this is narrow AI. In 10 years in may be broad AI and at that point you are just hoping it's in the right hands.
Slashdot - News for Technophobes.
Fear the future! Coming soon to a theatre near you!
...is that it is man-made. So I choose to freak out.
L.
So some sort of Kobayashi Maru manoeuvre would be in order then.
The AI will be the system that controls us. This is already happening in a very limited way, with many agencies using pretty unintelligent systems to scan and select documents and images. You do too -- every time you use a search engine.
Over time (decades, not years) these systems become more and more intelligent. They also compete with each other for survival, with many being discarded. Eventually they end up making higher level decisions.
It does not matter whether they are really "sentient" or not. What matters is how powerful they become. Initially in partnership with the people that control them, then less and less so.
See
http://www.computersthink.com/
http://waitbutwhy.com/2015/01/...
One of the best articles I've read this year. Long but very very well worth it.
Point is that whatever we're looking at now is nothing compared to what we'll have very very soon.
Doubt it's going to solve the halting problem any time soon, shit for brains.
Now guess who's next? They keyword is "hatred" - so guess who has to apologize for the significant amount of hatred over last hundreds (thousands) of years?
They should fear it, not you.
hmm I think we are screwed
That's wonderful for you, but it's not the typical experience.
Me too: people like to attribute their good luck on their virtue and other people's bad luck on their vice. But if the rules of the game require there to be a loser, that's irrational - if there's less jobs than applicants then someone is going to be left without even though any particular applicant can always try harder.
The problem is that this displaces the blame, excusing society and thus making it harder to fix and punishing people who, no matter what their personal failings might be, are innocent of the fact that there are losers.
Forget magic. Any technology distinguishable from divine power is insufficiently advanced.
Okay, so it's 20 years from now and most people have self driving cars. The A.I. in them is smart enough to tell if there are people in the way and will either stop or evade.
So you are driving along with your children, and all of a sudden a bunch of terrorist jump into the road and start pointing guns. The A.I. driving the car can select the following choices.
1. Stop the vehicle (which would enable the terrorist the ability to kidnap, torture and murder your children)
2. Evade hitting them by driving off the road and hitting a tree, which could kill your family, or at least would enable the terrorist the option to kidnap.
3. It could run over the terrorist, but there are 5 of them and only 4 in your car so the car simply looks at the numbers...
It's not about "virtue" and "vice". It's cause and effect. Unless of course you want to define "vice" as "error" and "virtue" as "stuff that works". That's not too far off - when many people think of vice and virtue, they think of things the Old Testament/Torah/Koran says to do and not do. That text uses uses the Greek words "hamartema" and "sophia". What in English we call "sin" is hamartema, which literally means "to miss" (the target or goal). "Virtue" in the Greek is sophia, which means wisdom or knowledge. One can miss their target, or one can have knowledge of what works well.
As a very simple example, sending out resumes to a dozen of relevant companies each week, then following up, works a lot better than sending one, waiting three weeks to hear hear back, sending one more, etc. One approach is effective, it is wise. The other approach is foolish if you're unemployed, it misses the mark.
> But if the rules of the game require there to be a loser, that's irrational - if there's less jobs than applicants then someone is going to be left
There are 5.8 million job opening in the US right now, and less than 2 million people who have been unemployed for a long time:
http://www.bls.gov/news.releas...
http://www.bls.gov/news.releas...
There are two million who have been unemployed for a short time. Simple arithmetic shows there are enough jobs not only for the four million unemployed, but there would still be 1.8 million jobs leftover. Therefore your proposition "if there's less jobs than applicants" is false. There are almost 6 million open jobs and 4 million unemployed people.
AI in its' current state is at best still a very narrow self-learning system without self-conciousness. Based on technological progression, we can only assume that sometime in near-term future (let's say 20 - 40 years to be conservative), we will build a sufficiently complex system as to create a general AI, capable of handling multiple disciplines, ala human. At that point, IMHO the question becomes - can the machine become self-aware. If we can truly create an artificial conciousness, then I'm not sure how you can NOT be alarmed. Evolution is way too slow to keep up with technological advancement, and at some point, someone, somewhere, will give it too much control. I don't care how careful we are at designing moral code, because at some point, a self-aware AI will see how much smarter it is. It will look at us as we look at a dog, and at best we will become just another cog in the food chain, no longer the masters. At worst, Earth will be cleansed of the human parasite to restore the natural balance. If self-awareness is actually "not a thing", then all bets are off.
"people mostly learn from their own mistakes. But they rarely learn from the mistakes of others. People collectively make the same mistakes over and over again,"
like, trusting AI.
Star Trek transporters are just 3d printers.
We can see through your ruse
Try it! Library of Babel
Your theory isn't illogical. It could happen that way.
> People with the right skills never need worry about being out of work, but Joe who's been told by his parole officer to get a job isn't going to have much luck here.
Again, my experience with plenty of Joes on parole is that as soon as they hear "get a job or go to jail", they always get a job. Within hours. These are the facts of my experience.
Again, you theoretical thought experiment could certainly happen in some universe.
> Good thing nobody ever has trouble finding a job that pays the same or more with equivalent benefits.
The same as what? How about a job that pays more than what 95% of people make? Is being in the top 5% good enough for a spoiled brat? To be in the top 5% richest people in the world, you need to make $9 / hour.
You can show up STONED and make $9 / hour in the US, you just have to show up.
I'm surprised anywhere in the US is starting kids at just $9 / hour to work in a gas station since it's twice that in Texas. Hopefully those kids will show up on time and get a raise after a bit. Maybe go to school and get a job other than "gas station cashier".
How would you like to be in the top 5% richest people in the world?
To be in the top 5% of income, you have to make $9 / hour. From the way you write, I'm guessing you have a bit of an education and make considerably more than than that. You're probably in the top 3%. You are better off than 97% of everyone. I understand that some kids kids in Orange county whine because they only have a Lexus while some of their neighbors have BMWs. Spoiled brats to the max.
I don't know if you're grateful for being in the US, where you earn more than 97% of the world does, or if you're a spoiled brat.
Yeah. Why be cautious if you're creating an entity that could have a qualitatively different relationship to information than us. Let's not worry about how a technology that recursively improves itself could be dangerous. I'm sure it'll all be fucking fine, just go make it.
I do not want your cheap brainburning drugs. They are useless for work. And I am a working man today.
Lions, tigers and bears are quite dangerous. Most of us wouldn't consider them particularly intelligent. (If you think they are, then consider sharks, or the Portuguese Man O'War.)
Of course we put them in zoos (or aquaria), not the other way around. But an automated tank could be quite hard to put in a zoo, even if it wasn't "intelligent."