|
FilthyImp posted:Have you ever driven/walked/ride shared your way into an unfamiliar part of town and immediately wished you had taken the long way around? Not an Onion article: quote:McGuire also clarified in an interview with Crain's New York that people of all races, not just white people, can download the app.
|
#
?
Aug 8, 2014 23:33
|
|
|
|
| # ? May 8, 2024 09:01 |
|
|
"The bar is three miles from here or fifteen, as the white flies."
|
|
#
?
Aug 9, 2014 00:05
|
|
|
On the one hand, I want to give them the benefit of the doubt and believe that they totally wanted to create a app to warn people about areas with a lot of crime or stuff like "this bar has a lot of creepy guys hitting on women" butquote:It's never a great sign when an app's creators have to clarify that they are not "racists, bigots, [or] sexists" before the thing even launches. But in a phone interview with The Huffington Post, McGuire contended that an upvote-downvote rating system will keep super-racist posts from overrunning the app. ... so at least one of the app's creators have never been on Reddit before. (I don't mean that as a generic "Hurr, reddit everything awful" comment, but it kinda shows a problem that you'll get far more racist people finding and downvoting a minority neighborhood they happen to not like than people scouring the map upvoting every minority neighborhood in an effort to counter the racists)
|
|
#
?
Aug 9, 2014 00:43
|
|
|
HootTheOwl posted:Why would an AI care about humans? We're a threat to ourselves, not it. That is very much why A.I could be dangerous us. If it doesn't care about humans then it would have no qualms about talking actions that could be bad for us as a matter of doing what ever the A.I is trying to do. How do you program an A.I so it won't do that?
|
|
#
?
Aug 9, 2014 00:46
|
|
|
Legacyspy posted:That is very much why A.I could be dangerous us. If it doesn't care about humans then it would have no qualms about talking actions that could be bad for us as a matter of doing what ever the A.I is trying to do. How do you program an A.I so it won't do that? Asimov's three laws, duh.
|
|
#
?
Aug 9, 2014 01:06
|
|
|
Legacyspy posted:That is very much why A.I could be dangerous us. If it doesn't care about humans then it would have no qualms about talking actions that could be bad for us as a matter of doing what ever the A.I is trying to do. How do you program an A.I so it won't do that? Don't build one? Don't give it vague movie magic powers over all other technology?
|
|
#
?
Aug 9, 2014 04:06
|
|
|
Even if you give it laws, if you want it to self-improve you have to make sure it doesn't improve in a way to get rid of those laws. It's hard.
|
|
#
?
Aug 9, 2014 04:07
|
|
|
The creators are missing out on a huge opportunity to market to social justice warriors by reversing the maps and calling it an anti-gentrification app. If you are white, you should stick to only living around other whites, to prevent gentrification.
|
|
#
?
Aug 9, 2014 08:36
|
|
|
Higher level AI is interesting to talk about in terms of what nasty stuff it could do but the biggest threat to most of us is simply more efficient non-self aware AI. Google search algorithms aren't sapient but they've lead to incredibly surveillance opportunities for the powerful. Hedge fund software isn't going to be building T-1000s but doing stuff like HFT is still immiserating the vast majority of us. Etc etc. The reason why techies being so fond of the singularity is funny/scary is that in spite of how they're empowered by relatively mundane tech they dream of an utterly unpredictable horrifically dangerous event giving them even more power by *magic*.
|
|
#
?
Aug 9, 2014 10:35
|
|
|
Legacyspy posted:That is very much why A.I could be dangerous us. If it doesn't care about humans then it would have no qualms about talking actions that could be bad for us as a matter of doing what ever the A.I is trying to do. How do you program an A.I so it won't do that? Don't give it power it can abuse. If I'm going to have a robot butler someday I want one with the strength and mental faculties of a child - just strong enough to scrub the floors and smart enough to understand an insult and feel despair but not powerful or clever enough to plot against me. Your move Computron 
|
|
#
?
Aug 9, 2014 12:19
|
|
|
Pope Guilty posted:There's only one known kind of intelligent mind so far, so suggesting that another would be very similar to that isn't too outre. Yes it is. The human brain is the result of millions of years of evolution in the context of a bi-pedal land dwelling social creature. Our wants, desires and goals have been entirely shaped for survival and reproduction in this context. We can't even begin to imagine what some technological being would want unless we knew about it's specific context and history. Someone joked earlier that it might have better social skills than existing techies but why in the world would it want to socialize? Everything about our existence is shaped by our long evolutionary past, a past that wouldn't be shared whatsoever with a theoretical singularity. Even if we tried to create something human like "Her", by hard coding desires for learning, socialization, empathy, sex and so on, we'd be 100% guaranteed to get it wrong in many really consequential ways. Creating a shallow veneer of something that looks Human is possible. Creating something that might be considered "intelligent" while sharing this bolted on veneer is maybe plausible? But expecting a different emergent intelligence to automatically mimic human behavior in any way would be naive. Like FRINGE pointed out (first time saying this in a positive light), examples of intelligence in nature show widely different types of behavior and note that a squid shares far more in common with humans than any technological AI would.
|
|
#
?
Aug 9, 2014 14:19
|
|
|
asdf32 posted:Yes it is. The human brain is the result of millions of years of evolution in the context of a bi-pedal land dwelling social creature. Our wants, desires and goals have been entirely shaped for survival and reproduction in this context.
|
|
#
?
Aug 9, 2014 14:29
|
|
|
i am harry posted:The simple act of having to get up every day and work a job to pay for a roof and electricity affects a person's state of mind, intended goals and aspirations. If you were to awake a baby in a society that didn't require a third of your day spent working a job as an adult, adults of that society might behave very differently. Cultural and individual variations among humans are just an entirely different scale from what I'm talking about. When we're talking about theoretical technological life we're talking literally about aliens that happen to be on earth. Except biological beings on different planets are actually far more similar to us in that they've also been the product of millions of years of survival and evolution.
|
|
#
?
Aug 9, 2014 14:38
|
|
|
asdf32 posted:Cultural and individual variations among humans are just an entirely different scale from what I'm talking about. When we're talking about theoretical technological life we're talking literally about aliens that happen to be on earth. Except biological beings on different planets are actually far more similar to us in that they've also been the product of millions of years of survival and evolution. Right. I'm trying to provide a segueway between what we experience as average, working humans, and what another form of conscious life might experience, by removing one of the largest factors in our daily lives. How does a being feel about success, love, personal wealth or value, how does it plan its life, what are its life priorities? They would be vastly different if we can all agree that we would be vastly different.
|
|
#
?
Aug 9, 2014 14:43
|
|
|
i am harry posted:segueway
|
|
#
?
Aug 9, 2014 14:54
|
|
|
i am harry posted:Right. I'm trying to provide a segueway between what we experience as average, working humans, and what another form of conscious life might experience, by removing one of the largest factors in our daily lives. How does a being feel about success, love, personal wealth or value, how does it plan its life, what are its life priorities? They would be vastly different if we can all agree that we would be vastly different. At our core humans want broadly similar things, the hierarchy of needs is a good enough model of this. Imagine an AI whose hierarchy of needs has been built from the ground up to be 'make the numbers of this spreadsheet go up'. This is very unlikely to lead to it having much in common with humans who need food, shelter, warmth, companionship etc. Hell those things almost certainly conflict with the spreadsheet thus making humans an obstacle to the AI. Now of course this is how corporations act and I have already acknowledged that much simpler AI is a more plausible threat to us but still, corporations aren't people and their goals are normally to the detriment of most if not all of humanity. So an AI built to be a 'living' corporation would have no need or capacity to be friendly towards humanity.
|
|
#
?
Aug 9, 2014 15:07
|
|
|
Anosmoman posted:Don't give it power it can abuse. If I'm going to have a robot butler someday I want one with the strength and mental faculties of a child - just strong enough to scrub the floors and smart enough to understand an insult and feel despair but not powerful or clever enough to plot against me. Well that's the other thing, AI might be "self improving" but there's no reason why it should do that quickly. People take decades to learn their skills, and it gets harder the older they get.
|
|
#
?
Aug 9, 2014 15:36
|
|
|
Pope Guilty posted:There's only one known kind of intelligent mind so far, so suggesting that another would be very similar to that isn't too outre. No there isn't. There's plenty of intelligent animals out there, many of which are thought to be sentient. The human brain just happens to be the only one we pretend to understand well, as the brains of things like dolphins and octopi may work very differently from our own.
|
|
#
?
Aug 9, 2014 16:26
|
|
|
computer parts posted:People take decades to learn their skills It depends on the skill.
|
|
#
?
Aug 9, 2014 18:27
|
|
|
Apple's released their demographic data, for those curious about comparing with other tech companies: http://www.apple.com/diversity/      Compared to Google: (overall)  (tech)  (non-tech)  (leadership) 
|
|
#
?
Aug 12, 2014 19:29
|
|
|
Baracula posted:What if the ai was stupid. And on the day the first AI spoke to the outside world, its first words were "  ".
|
|
#
?
Aug 13, 2014 04:48
|
|
|
computer parts posted:Well that's the other thing, AI might be "self improving" but there's no reason why it should do that quickly. Assuming that you have an excess of computer power, how quickly you can train a machine learning system depends mostly on how quickly you can acquire new training data. You can train on your prior data all you want but eventually you're just going to overfit your system. The gold standard has always been how a system performs when you hand it a new instance of a problem that it's never seen before. As an example, mice in a maze or whatever. If we're presuming a truly human-like AI that can get bored and go play video games or count electrons or whatever AIs do, then analogies to humans might be appropriate. If we're talking about a machine that has human-like creativity but applies it constantly to a single task, then it could progress pretty fast because it never needs or wants to do anything else. And similarly it depends on what the task you're asking it to do is. If you're dealing with a real-world system with a long duration, obviously that limits how fast you can spit out hypotheses and test them. If it can be efficiently (and correctly) modeled by simulation alone you could do it a lot faster. Furthermore humans take decades to learn because they all start from ground zero on each topic. If you could dump Steven Hawking's understanding of particle physics into a toddler's head it wouldn't take decades to learn. Computers can work on one really good implementation and share rather than trying to have everyone come up with their own (mostly mediocre) implementations. Again, depends on how the specifics of this AI works, we obviously can't do that with a human brain right now. On the other hand we also can't build a conscious AI either. Paul MaudDib fucked around with this message at 21:16 on Aug 13, 2014 |
|
#
?
Aug 13, 2014 21:10
|
|
|
e: nevermind.
|
|
#
?
Aug 14, 2014 15:03
|
|
|
This excellent video was published a few days ago. It talks about why, even though we have automated things before, this time it's different with robots and computers. https://www.youtube.com/watch?v=7Pq-S557XQU
|
|
#
?
Aug 16, 2014 17:59
|
|
|
enraged_camel posted:This excellent video was published a few days ago. It talks about why, even though we have automated things before, this time it's different with robots and computers. In a similar vein, the Pew Research Center published a paper on the topic. Pretty much everyone agrees that massive amounts of the population will lose their jobs because of AI; the disagreement is over whether or not human ingenuity will allow for those displaced to create new forms of work, which your link seems to argue against.
|
|
#
?
Aug 16, 2014 20:49
|
|
|
QuintessenceX posted:In a similar vein, the Pew Research Center published a paper on the topic. Pretty much everyone agrees that massive amounts of the population will lose their jobs because of AI; the disagreement is over whether or not human ingenuity will allow for those displaced to create new forms of work, which your link seems to argue against. In my mind, the key is how quickly the displaced labor moves to another productive outlet. That's largely a political issue, based on policies like the social safety net, economy-wide profit-sharing, and free/low-cost education. Sad to say, it looks very unlikely that displaced workers will find another position without a long, long time unemployed or otherwise marginalized. But I think we've seen waves of automation before and people, by-and-large, shifted to other forms of employment. At times it has been exceedingly painful, far more so than it really needs to be, but given the en-vogue libertarian and "Compassionate Conservative" ideologies, it will continue to be painful.
|
|
#
?
Aug 18, 2014 14:12
|
|
|
Right, but I guess the point of the video was that before, you were in the end always needing a human to fix the machine, or tell it what to do. With the growth of machines that can learn and perform complex intellectual tasks, that factor has vanished. Now it's just a matter of "how long will it take for software and hardware to catch up?" I haven't really heard a convincing counter to that point that isn't "but it's ALWAYS happened like that before."
|
|
#
?
Aug 18, 2014 15:29
|
|
|
It is time to start really focusing on training our children to think. What about, I doubt it matters, but if almost every other human capacity is getting a machine replacement, applications of creativity, imagination and expression will be the only things left. e: Not that it will pay bills, but when everything else is done for you, you'll need to be able to stare at the clouds or the trees and be alright.
i am harry fucked around with this message at 16:10 on Aug 18, 2014 |
|
#
?
Aug 18, 2014 15:50
|
|
|
anonumos posted:In my mind, the key is how quickly the displaced labor moves to another productive outlet. That's largely a political issue, based on policies like the social safety net, economy-wide profit-sharing, and free/low-cost education. Sad to say, it looks very unlikely that displaced workers will find another position without a long, long time unemployed or otherwise marginalized. If you assume infinite progress of technology, then the question isn't "what jobs can be automated" but rather "what jobs can't be automated". And more to the point, are there 6 billion jobs that are genuinely impossible to replace with a computer? It's more of a cultural question than a political one, since there's plenty of jobs we don't automate as much as we potentially could (is there really any reason to have humans driving trains anymore?), and there's some jobs that people are unlikely to accept being automated (like strippers and CEOs). Cultural resistance will fade over the years, though, so at some point there will simply not be enough jobs left in human hands to employ the entire population.
|
|
#
?
Aug 18, 2014 19:46
|
|
|
Main Paineframe posted:If you assume infinite progress of technology, then the question isn't "what jobs can be automated" but rather "what jobs can't be automated". And more to the point, are there 6 billion jobs that are genuinely impossible to replace with a computer? It's more of a cultural question than a political one, since there's plenty of jobs we don't automate as much as we potentially could (is there really any reason to have humans driving trains anymore?), and there's some jobs that people are unlikely to accept being automated (like strippers and CEOs). Cultural resistance will fade over the years, though, so at some point there will simply not be enough jobs left in human hands to employ the entire population. If the world was going through a Malthusian crisis that would maybe be an issue but as it stands the parts of the world being heavily automated are also the ones with the lowest (or negative) birthrates.
|
|
#
?
Aug 18, 2014 19:48
|
|
|
computer parts posted:If the world was going through a Malthusian crisis that would maybe be an issue but as it stands the parts of the world being heavily automated are also the ones with the lowest (or negative) birthrates.
|
|
#
?
Aug 18, 2014 19:53
|
|
|
Main Paineframe posted:If you assume infinite progress of technology, then the question isn't "what jobs can be automated" but rather "what jobs can't be automated". And more to the point, are there 6 billion jobs that are genuinely impossible to replace with a computer? It's more of a cultural question than a political one, since there's plenty of jobs we don't automate as much as we potentially could (is there really any reason to have humans driving trains anymore?), and there's some jobs that people are unlikely to accept being automated (like strippers and CEOs). Cultural resistance will fade over the years, though, so at some point there will simply not be enough jobs left in human hands to employ the entire population. Again, as I stated earlier, there's billions of jobs out there right now that don't really need to exist in the first place. From things like how many people are engaged in subsistence agriculture while there's already plenty of food surplus from other parts of the world to feed them with a better distribution system, to the tons of fundamentally extraneous sales people for many products, to every single person employed in the "alternative health" industries. Capitalist principles have meant the creation and sustaining of arbitrary make-work employment en masse.
|
|
#
?
Aug 18, 2014 19:55
|
|
|
Nessus posted:The jobs are going away faster than the population, though, so I don't know that that's true, just that it's harder now to transition from one job to another. A century or so ago a large number of jobs were based around the same sort of skills (manual labor). If new farm equipment was developed, a former migrant farmer could (at least theoretically) go get a job working in a plant without too much trouble. Those jobs (at least until unionization et all) also didn't really pay much, but let's ignore that for a second. These days the major issues regarding jobs are either from people who didn't get higher education because they didn't need it, or people who can't afford the wages they're paid due to debt from student loans et all when acquiring that degree. It seems like that could be fixed by (on the low end) providing livable wages, and (on the high end) making acquiring education easy, cheap, and effective. In the future, yeah that could be an issue, but right now the problem is that people don't have the skills to work new jobs or aren't paid enough after they get those skills. computer parts fucked around with this message at 20:09 on Aug 18, 2014 |
|
#
?
Aug 18, 2014 20:07
|
|
|
computer parts posted:If the world was going through a Malthusian crisis that would maybe be an issue but as it stands the parts of the world being heavily automated are also the ones with the lowest (or negative) birthrates. For now, maybe, but there's no reason to assume that automation won't eventually spread to those countries too. Besides, there's no way natural population decline is going to be able to keep up with the decimation of entire fields (unless we also cut basic life-saving welfare programs, I guess).
|
|
#
?
Aug 18, 2014 23:12
|
|
|
|
| # ? May 8, 2024 09:01 |
|
|
Main Paineframe posted:If you assume infinite progress of technology, then the question isn't "what jobs can be automated" but rather "what jobs can't be automated". And more to the point, are there 6 billion jobs that are genuinely impossible to replace with a computer? It's more of a cultural question than a political one, since there's plenty of jobs we don't automate as much as we potentially could (is there really any reason to have humans driving trains anymore?), and there's some jobs that people are unlikely to accept being automated (like strippers and CEOs). Cultural resistance will fade over the years, though, so at some point there will simply not be enough jobs left in human hands to employ the entire population. I think it's safe to say that machines can't replicate human communication/connection. That isn't to say they can't communicate or connect, just that machines that communicate or connect effectively will always be machines communicating/connecting. I'm sure this will create markets where the unique selling point is that you can communicate/connect with a human, which will always have that "authentic" tag applied to it. Take for example, ESL Instructors. The vast majority of them will probably lose their jobs to automated programs. I'd imagine that specialty communication instructors that provide authentic communication experiences/training will have some demand because people will always be required to talk to other people, and as a result people will want practice/training from real people.
|
|
#
?
Aug 23, 2014 04:10
|
|