When I was studying computer science, in a class of about 40 people we would have on average 2 or 3 girls. I'm shocked it's so high. I guess things have gotten better since when I was in college.
It's not about blaming men or women. If anyone is to blame, it is parents for helping to enforce gender stereotypes and not doing enough to make sure that their daughters have the tools necessary to compete in the tech sector. By the time these girls grow up to be women, bu they don't have these tools, it's too late. They don't even want to be computer scientists because they were never exposed to the sorts of things that would both inspire and train them to be attracted to and prepared for that field.
So if you deceive a company into hiring you by tricking their HR department, won't they just figure out you are incompetent at some point?
This doesn't seem like so much a problem of racial discrimination as it is with the fact that some people are not good interviewers. Are you suggesting that minorities are inherently bad at interviewing or something?
From a strategic point of view this is a clusterfuck. Why did Nokia put real money in the bag if they were planning to arrest the person that came to pick it up? If the police had succeeded then it wouldn't matter if the money was real. If the blackmailer gets away, then maybe, if you are lucky, he might keep his promise if he thinks you acted in good faith. But now I am reading a story on slashdot about how they tried to catch this guy and botched the plan, so now the blackmailer knows that Nokia was not acting in good faith. Now the blackmailer has no reason not to leak the keys, unless he plans to try to extort more money.
Lets say I decide to modify my ikea furniture myself. The modifications I made are dangerous. If my furniture causes me to injure myself, should I be allowed to sue ikea for selling me furniture that I was able to then turn into something dangerous? No obviously not. Does it matter if some (stupid) people perceived ikea's furniture as being safely modifiable? It seems like people's perception shouldn't really matter unless that perception is reasonable.
I would suggest that assuming that every modification presented in ikeahackers is safe because ikea allowed them to use their trademark (i.e. didn't sue the shit out of them), is also not a reasonable perception, and shouldn't be legitimized.
I find it ironic that someone who argues that everyone (even small children, and mentally incompetent and psychotic people) should have the right to own firearms, because freedom is infinitely more important than safety, would be defending trademark laws on the grounds of safety.
Considerate people may feel this way, but they shouldn't. Short term trips are a necessary part of the system; it may not be good for the one driver who takes you (on that day), but it is better for every other driver behind him in the queue, so it is better for everyone.
Actually the better solution is to not even use traditional taxis for short hauls. That way no taxis need to lose the lottery by picking you up. You can use Uber for short hauls, and pay a fair market price. You can also use Uber for long hauls to also get a fair market price for long hauls as well (i.e. where you won't get ripped off yourself), then the taxis go out of business as all the taxi drivers decide that their best option is to become an uber driver.
We're not quite there yet, actually. The problem is that we're in a transition period where traditional taxi companies need to coexist with services like Uber and the traditional companies need to serve people who aren't using smart phones. So there needs to be at least some temporary solution to keep things fair. E.g. an airport could have a short haul queue and a long haul queue. Or they could have the passengers line up at a kiosk where they are matched up with the best driver.
The temporary solution is airport kiosks that provide access to Uber for a premium. Maybe Uber could also provide a phone service for people without smartphones, where you talk to a live operator who help you set up an UberX ride (i.e. one that works more like a traditional taxi in that the driver is an uber employee driving an uber branded car, but with uber-style variable pricing)
In fact airports are places where a system like Uber might even be able to take advantage of the fact that lots of people might be going to the same place and have groups of people share a ride for better efficiency.
Going to the airport as a taxi driver is like participating in a lottery with a positive expected value.
What is the benefit of having an airport trip be a lottery? Why not just charge people based on the actual cost of their ride?
Having a system where cabs must compensate for their losses on short hauls with disproportionate profits on long hauls, not only creates volatility (i.e. a driver can have really good days and really bad days), it means that long haul passengers are subsidizing the cost of short haul passengers. It also means that considerate people will feel guilty about going on short hauls as they are depriving a driver of his livelihood, and price conscious passengers are turned off from going on long hauls as they know they are being ripped off.
Taxi companies have fought hard to get the flag drop and first mile rates they already have. Good luck getting them to raise them further to implement your fix to the system.
The whole point of a free market is that you don't have to fight to raise prices. You can simply refuse to take jobs that are not profitable to you. If enough drivers refuse to take short hauls, then either you will have people willing to pay higher fares (due to lower supply), or maybe it just doesn't make economic sense to have taxis do short hauls from the airport at all.
At one point it may have been inefficient to exhaustively analyze the fair price for a cab fair or time consuming to bargain with each cab, and the benefit of having simple pricing schemes, outweighed the disadvantages of volatility and a distorted market. Now we have computers that can make price calculations fairly and quickly. Why not take advantage of them?
But taxi drivers, and Uber drivers are greed-driven.
The question is why taxi drivers are forced to take short hauls.
the city forces regulation on its taxi companies so that visitors know that taxis in Vegas are a reliable always-on-demand service.
When you must force someone to do something for money, it means they are not being paid enough to do it.
Letting the market "correct itself" so that short hauls are more expensive, the tragedy of the commons [wikipedia.org] happens and suddenly there are fewer and fewer long hauls, since there are fewer and fewer short-hauls, as more people rent cars or avoid Vegas entirely opting for Atlantic City, Foxwoods, or Macau.
I am well aware of the concept of the tragedy of the commons, and this not an example of it.
Why would allowing the price of short hauls to be dependent on it's actual cost cause fewer long hauls?
An unregulated free market doesn't always solve everything.
I never said it solves everything. I said it solves this particular problem.
Without laws forcing them to do short hauls, they simply won't do them. Uber drivers won't be any different in this regard.
I don't think you know how Uber works. If no drivers are willing to take a particular fare, then the price goes up until an Uber driver is willing to take the fare.
Furthermore, why is it fair to expect a really cheap short haul ride when a driver is required to wait in line at an airport for an hour as in your example?
Why shouldn't prices be more reflective of their true cost in resources like gasoline and people's time and energy?
If all these taxis only want to do long haul trips, then it simply means the long haul trips are overpriced and the short haul trips are underpriced. Uber solves this problem much better than government regulations forcing taxis to take fares they don't want.
Requiring cabbies to accept all fares at a fixed rate made a lot of sense years ago because it made visiting other cities far easier without being ripped off or stranded.
Why do you think requiring cabbies to accept all fares at a fixed rate is fair? If no cabbies want to accept certain fares because they are less profitable than other fares, or they all want to charge certain fares more money, it's not because they are greedy, it's because the government is bad at setting market prices.
Having a running meter is the most fair thing anyone could think of last century. Now we have better tools. No one is getting ripped off with Uber. Neither the customers nor the drivers. It's all just supply and demand based pricing. In fact as a tourist, I feel much more comfortable using Uber than the local taxi system because I know how Uber works and I know it's fair. I trust Uber more than I trust an unfamiliar government.
You could fly into London and know you'd be able to get a black cab to your hotel at the designated rate.
Is the designated rate fair? I would rather fly to London and know I am getting a fair market rate. I know I am not getting ripped off, and I know I am not ripping off the cab (i.e. I know I'm not a fare that he would rather not take, but is forced to by some regulation at some fixed price).
Maybe, or is it the just the case that uber drivers are just cherry picking the best fares. Living the dregs to cabbies, who are mandated to take them. Maybe, or is it the just the case that uber drivers are just cherry picking the best fares. Living the dregs to cabbies, who are mandated to take them.
Uber drivers are supposed to cherry pick the best fares. It's Uber's job to ensure that the prices are set such that for every person who needs a ride, there is at least some drivers willing to take the job. If nobody wants to give you a ride because they can make more money on other jobs, that means that the price is set below market value.
Or is this a market will fix the problem situation?
This is *exactly* the kind of problem the market fixes.
And anyone who wants a ride somewhere off the beaten path, so there is no chance of a return fare, or a short haul which is less profitable will have pay a premium to get service?
Yes. That's how it's supposed to work. If no one wants to give you a ride for the price you are willing to pay, you are supposed to pay them more money. How would you feel if someone passed a law forcing you to do your job for less money than you are willing to do it for?
But that hearkens to the old days where cabbies simply refused fares outright, or would demand you pay double (ie pay a 'return fare') to make it worth their while.
Why do you feel entitled to a ride that costs less money than a taxi is willing to take? Why shouldn't people who need rides which require more resources pay more money? The fact that you live out in the boonies and taxis want you to pay a double fare is indicative of the fact that it is more costly to provide service to you.
I guess that would "solve" the problem. Not sure that's a good solution though.
It's a market solution where everyone pays a price that is based on how many resources they are consuming. Just like how you pay more money if you eat more food than a regular person, or use more gasoline than other people.
For all the problems that are best solved by government intervention, there is some shit like this happening. I don't know that much about taxi regulations, but I do know that hiring a guy off the internet to drive me to the airport can't be any more harmful than having a friend drop me off at the airport, so what the fuck is the problem?
In fact this is the only time I used Uber. The guy parked at the place where everyone else parks, I get my shit out of the trunk, I tried to tip him, he refused to take it, and he left. Does their really need to be a fucking regulation to prevent this from happening?
Dear California government, please don't make these tea party idiots right more often than they have to be.
Uber drivers and cherry pick and skip any job that doesn't seem worth it for them. One of the advantages of the Uber system is that they try to set prices so that there will always be *some* drivers willing to take the job for that price.
This in my opinion is a much better solution. It is more elegant. It doesn't force anyone to do anything they don't want to do. Rather than forcing taxis to take small jobs, why not allow them to be compensated enough to make them want to do those jobs?
Laws are just not versatile enough to adapt to these sorts of changing conditions. Maybe there was a time when we needed these sorts of laws to have the society we want. Now it appears we don't.
Mobile phones ring when they are called. Their are also ways to silence them, and this is what people are instructed to do at the start of a movie.
Do you know if Google glass does not have a similar feature to turn off lights and sounds in order not to disturb people (analogous to putting your phone on silent)?
This is a good example of how regulations can stifle the market and have unintended consequences. They all seem like reasonable regulations, but #3 and #4 make unnecessary assumptions about the way taxis work. The spirit of this regulation seems to be to make sure that customers are informed about how much they will need to pay before being obligated to pay it. Uber absolutely follows the spirit of this law, I would argue, even better than traditional taxis. You pay up front because with Uber it can calculate how far the journey will be before actually driving it. This eliminates the need to post rates and to have a meter.
But you are not allowed to follow the spirit of the law. You must follow the letter of the law. The letter of the law would require Uber to follow regulations designed for traditional taxis that for Uber just create unnecessary expenses that then drive up costs for consumers. Many regulations are necessary and work well. It is important to recognize when they become obsolete and to update or remove them. Many times it's better to have fewer regulations if possible, because it makes maintaining them easier and more cost efficient, and therefore more likely.
In many cases when regulations become too complicated, there is a temptation by both citizens and legislators to simply avoid the problems rather than fixing them (e.g. like allowing your garden to become overgrown). By keeping your garden small and manageable, it is more likely that people will be able and willing to keep it maintained.
It sounds like the swedish taxi regulation garden is certainly small enough to manage. So all that is left is to update a few rules for the 21st century. Good job Sweden!
Sure, nobody actually runs a Turing test. It's too hard. If a real Turing test were ever passed there wouldn't be any dispute. They're ALL restricted versions where the judges go easy on the computers.
I guess I am not really sure I see the benefit of calling anything but a real Turing test a Turing test. I think this concept of a "restricted Turing test" is a useful concept, but I think having the "Turing test" in the name, even though it's "restricted" carries a lot of baggage (i.e. in regards to sentience) , that makes people question the legitimacy of the Turing test in general as a test of sentience.
The 13 year old gambit isn't the problem though (it's the judges). In fact, it suggests all sorts of strategies for the judges to trip up the computer. I just had a quick conversation with Eugene where I told him a story about a pretty girl asking a guy to go to the movies, and the two of them sitting right at the back. He changed the subject. Obviously not a 13 year old boy.
It seems to me that changing the subject (even though humans do actually do it), should be considered a dead giveaway. In a real Turing test human subjects should actively try not to change the subject if possible, since this is a easy out for a computer.
I agree with you, a proper Turing test is the best, possibly the only way we currently have to assess an AI. But if you have a computer that you think is at the level of a five year old, for example, find some child psychologists and let them talk to it and some real five year olds. Or thirteen year olds. Or adults. Age doesn't matter.
I think age doesn't matter to a point. Yes you can get a child psychologist and train them to be Turing test judges, but it restricts the number of people that can be judges. Eventually you get to the point where young human children do not have enough mastery of any language (and typing) to be easily distinguishable from computers.
I am not sure exactly what that age is, but in the spirit of the turing test (i.e. removing all possible doubt), I think testing adult humans and "adult" AIs is better.
I do however think that a likely first real contender to passing a legitimate Turing test might be a "child", if the AI actually works in a way that it needs to learn language from scratch (like a baby), which seems quite probable. And I think we will probably create a sentient AI that is not able to pass the turing test before we make one that can.
I'm not talking about this specific chatbot. I am saying that this is how one could rig the turning test using the "13 year old boy angle", which is a just a special case of the "uncooperative subject angle".
I don't know exactly which tricks this chatbot used, but some were hinted at in the article, and thus far no computer has been able to legitimately pass the turing test (i.e. because machines are not sentient yet), and therefore all that appear to pass are using some set of dirty tricks.
Anybody who doesn't think intelligence is about memorizing trivia, the very thing the 13 year old gambit is supposed to help explain away, would see through it very quickly.
Any turing test run today should have nearly 100% of judges singling out the computers. If you are not getting these results, something is wrong with the way the test is being applied. IN the future this will not be the case.
The fact that so many judges were not able to figure out the chatbot was a computer, may even have just been a problem with the judges themselves. Most people don't actually know how to properly judge a turing test. They don;t know that you aren't supposed to "give the computer a fair chance". They don;t know you are supposed to mercilessly exploit any potential weaknesses a computer might have and the strengths of humans. You are supposed to give real humans every possible advantage to prove who they are (i.e. without cheating and knowing who they are).
My point was only that the currently we are seeing obviously non-sentient computer programs appearing to pass the Turing test, because it is applied incorrectly, not because the test itself is a bad one.
In fact I think it is probably the *only* legitimate test for human-like intelligence (when applied properly), and I think Turing was a brilliant man for coming up with it at the dawn of machine computing. He was also brilliant for many other reasons as well.
I don't have a problem with 13 year old kids in the test as long as they are mature enough to participate cooperatively. There are some kids that would not be able to, and even some adults that would not.
The problem I would have is when the "subject" tries to exploit a child's immaturity to pass the test. This wouldn't necessarily be a problem, as the judge could simply count immaturity as a trick that a computer program might employ, but if the pool of humans is also stacked with immature children, this might compromise the test because the judges would have a hard time distinguishing between the uncooperative human children and seemingly uncooperative chatbots.
You could do the exact same think with uncooperative adults, but I think when this is done, it is more easily recognized as the cheap trick that it is. None of the humans should be anything but cooperative and able participants. This is necessary to set an appropriately high bar for the computers. Simply lowering the bar for the computers by including terrible human subjects is basically rigging the test.
You say the device is distracting. How is it distracting? I notice people using their cell phones in a movie. How would I notice someone who is wearing a particular kind of eyeglass thing during a movie?
Another example I just thought of would be an adult English speak human who was just a very slow typer. Obviously slow typers are still intelligent, but it is really hard to judge the humanness of a test subject if they only type 4 sentences during the 5 minute time period.
The spirit of the Turing test is to actually engage the judge and actively convince the judge of it's humanity rather than simply providing plausible excuses for why it can't do that properly.
"I sorry, I don understand you say. I no speak good English" (poor english)
"You are boring, I only want to talk about pokemon and ignore whatever you say." (immature)
"They say I am mentally unstable and that I constantly change the subject and give irrational responses. I don't know what they are talking about. They are just voice in my head. I like icecream" (paranoid schizophrenic)
These are all things that real humans with real intelligence might say. They are also things that are tricially easy form computer programs to imitate without actually having any intelligence at all (i.e. they don't prove intelligence).
The whole point of the Turing test is to allow a computer to prove it is intelligent in an unbiased setting.
Not passing the Turing test does not mean you are not intelligent. Lots of human beings can't pass the test for various reasons (some of which I have used in examples). But legitimately passing a turing test is supposed to be really nearly impossibly hard, but if/when it happens should remove all doubt that whatever passed (human or machine) is intelligent.
We do consider 13 year olds intelligent. We also consider people who speak foreign languages to be intelligent. The problem is that you can't understand people who speak languages that we don't. We assume they are intelligent because nearly every other human we have interacted with was intelligent and we know that languages other than our native language are legitimate languages that we simply don't understand. We assume that these people that can't communicate with us are intelligent despite their inability to pass a legitimate turing test from our perspective.
My point was not to suggest that 13 year old Ukranian boys were not intelligent, but merely that they do not have the tools to pass a 5 minute turing test with English speaking judges. And this lack of ability was used as a way to influence the judges (i.e. a real 13 year old Ukranian boy might not be able to pass, so maybe this is a real 13 year old Ukranian boy).
This is cheating in the Turing test. This is like passing a class by calling in a bomb threat and not having to take the final exam. Sure you didn't even take the final, but none of the people who knew the material took the final either, so maybe you know the material as well. They are passing the Turing test by not really taking the test (by invoking a false language/maturity barrier).
Why didn't they make the chatbot an English speaking adult? Because English speakers are epxected to give normal human responses in English (where Ukranians are not). Adults are expected to be mature and give thoughtful responses (where 13 year old boys might be a bit more eratic and uncooperative if they are bored).
All these attributes of the chatbot are meant to provide plausible excuses for not passing the test rather than actually trying to pass the test.
A 5 minute test might actually be sufficient under the right situation (i.e. no communication/cooperation barriers), but this chatbot was actually trying to amplify barriers to make the 5 minute time limit as inadequate for determining "humanity" as possible.
I may as well make a chat bot that only says one sentence:
"I'm really sorry, but I have to go to the bathroom... BRB"
This is a very plausible thing for a real human to say, but if it is the only thing it says in the 5 minutes, then it makes it very hard to judge whether it is really a machine or a real human. It is creating a false plausible barrier to judgement.
Slashdot Mirror
When I was studying computer science, in a class of about 40 people we would have on average 2 or 3 girls. I'm shocked it's so high. I guess things have gotten better since when I was in college.
It's not about blaming men or women. If anyone is to blame, it is parents for helping to enforce gender stereotypes and not doing enough to make sure that their daughters have the tools necessary to compete in the tech sector. By the time these girls grow up to be women, bu they don't have these tools, it's too late. They don't even want to be computer scientists because they were never exposed to the sorts of things that would both inspire and train them to be attracted to and prepared for that field.
Why treat people as members of a group with the same skin color and genitals? Why not just treat them as individuals?
So if you deceive a company into hiring you by tricking their HR department, won't they just figure out you are incompetent at some point?
This doesn't seem like so much a problem of racial discrimination as it is with the fact that some people are not good interviewers. Are you suggesting that minorities are inherently bad at interviewing or something?
From a strategic point of view this is a clusterfuck. Why did Nokia put real money in the bag if they were planning to arrest the person that came to pick it up? If the police had succeeded then it wouldn't matter if the money was real. If the blackmailer gets away, then maybe, if you are lucky, he might keep his promise if he thinks you acted in good faith. But now I am reading a story on slashdot about how they tried to catch this guy and botched the plan, so now the blackmailer knows that Nokia was not acting in good faith. Now the blackmailer has no reason not to leak the keys, unless he plans to try to extort more money.
Lets say I decide to modify my ikea furniture myself. The modifications I made are dangerous. If my furniture causes me to injure myself, should I be allowed to sue ikea for selling me furniture that I was able to then turn into something dangerous? No obviously not. Does it matter if some (stupid) people perceived ikea's furniture as being safely modifiable? It seems like people's perception shouldn't really matter unless that perception is reasonable.
I would suggest that assuming that every modification presented in ikeahackers is safe because ikea allowed them to use their trademark (i.e. didn't sue the shit out of them), is also not a reasonable perception, and shouldn't be legitimized.
I find it ironic that someone who argues that everyone (even small children, and mentally incompetent and psychotic people) should have the right to own firearms, because freedom is infinitely more important than safety, would be defending trademark laws on the grounds of safety.
Considerate people may feel this way, but they shouldn't. Short term trips are a necessary part of the system; it may not be good for the one driver who takes you (on that day), but it is better for every other driver behind him in the queue, so it is better for everyone.
Actually the better solution is to not even use traditional taxis for short hauls. That way no taxis need to lose the lottery by picking you up. You can use Uber for short hauls, and pay a fair market price. You can also use Uber for long hauls to also get a fair market price for long hauls as well (i.e. where you won't get ripped off yourself), then the taxis go out of business as all the taxi drivers decide that their best option is to become an uber driver.
We're not quite there yet, actually. The problem is that we're in a transition period where traditional taxi companies need to coexist with services like Uber and the traditional companies need to serve people who aren't using smart phones. So there needs to be at least some temporary solution to keep things fair. E.g. an airport could have a short haul queue and a long haul queue. Or they could have the passengers line up at a kiosk where they are matched up with the best driver.
The temporary solution is airport kiosks that provide access to Uber for a premium. Maybe Uber could also provide a phone service for people without smartphones, where you talk to a live operator who help you set up an UberX ride (i.e. one that works more like a traditional taxi in that the driver is an uber employee driving an uber branded car, but with uber-style variable pricing)
In fact airports are places where a system like Uber might even be able to take advantage of the fact that lots of people might be going to the same place and have groups of people share a ride for better efficiency.
Holy crap, some things become usable after development! EVERYTHING must become usable after development!
Going to the airport as a taxi driver is like participating in a lottery with a positive expected value.
What is the benefit of having an airport trip be a lottery? Why not just charge people based on the actual cost of their ride?
Having a system where cabs must compensate for their losses on short hauls with disproportionate profits on long hauls, not only creates volatility (i.e. a driver can have really good days and really bad days), it means that long haul passengers are subsidizing the cost of short haul passengers. It also means that considerate people will feel guilty about going on short hauls as they are depriving a driver of his livelihood, and price conscious passengers are turned off from going on long hauls as they know they are being ripped off.
Taxi companies have fought hard to get the flag drop and first mile rates they already have. Good luck getting them to raise them further to implement your fix to the system.
The whole point of a free market is that you don't have to fight to raise prices. You can simply refuse to take jobs that are not profitable to you. If enough drivers refuse to take short hauls, then either you will have people willing to pay higher fares (due to lower supply), or maybe it just doesn't make economic sense to have taxis do short hauls from the airport at all.
At one point it may have been inefficient to exhaustively analyze the fair price for a cab fair or time consuming to bargain with each cab, and the benefit of having simple pricing schemes, outweighed the disadvantages of volatility and a distorted market. Now we have computers that can make price calculations fairly and quickly. Why not take advantage of them?
But taxi drivers, and Uber drivers are greed-driven.
The question is why taxi drivers are forced to take short hauls.
the city forces regulation on its taxi companies so that visitors know that taxis in Vegas are a reliable always-on-demand service.
When you must force someone to do something for money, it means they are not being paid enough to do it.
Letting the market "correct itself" so that short hauls are more expensive, the tragedy of the commons [wikipedia.org] happens and suddenly there are fewer and fewer long hauls, since there are fewer and fewer short-hauls, as more people rent cars or avoid Vegas entirely opting for Atlantic City, Foxwoods, or Macau.
I am well aware of the concept of the tragedy of the commons, and this not an example of it.
Why would allowing the price of short hauls to be dependent on it's actual cost cause fewer long hauls?
An unregulated free market doesn't always solve everything.
I never said it solves everything. I said it solves this particular problem.
Without laws forcing them to do short hauls, they simply won't do them. Uber drivers won't be any different in this regard.
I don't think you know how Uber works. If no drivers are willing to take a particular fare, then the price goes up until an Uber driver is willing to take the fare.
Furthermore, why is it fair to expect a really cheap short haul ride when a driver is required to wait in line at an airport for an hour as in your example?
Why shouldn't prices be more reflective of their true cost in resources like gasoline and people's time and energy?
If all these taxis only want to do long haul trips, then it simply means the long haul trips are overpriced and the short haul trips are underpriced. Uber solves this problem much better than government regulations forcing taxis to take fares they don't want.
Requiring cabbies to accept all fares at a fixed rate made a lot of sense years ago because it made visiting other cities far easier without being ripped off or stranded.
Why do you think requiring cabbies to accept all fares at a fixed rate is fair? If no cabbies want to accept certain fares because they are less profitable than other fares, or they all want to charge certain fares more money, it's not because they are greedy, it's because the government is bad at setting market prices.
Having a running meter is the most fair thing anyone could think of last century. Now we have better tools. No one is getting ripped off with Uber. Neither the customers nor the drivers. It's all just supply and demand based pricing. In fact as a tourist, I feel much more comfortable using Uber than the local taxi system because I know how Uber works and I know it's fair. I trust Uber more than I trust an unfamiliar government.
You could fly into London and know you'd be able to get a black cab to your hotel at the designated rate.
Is the designated rate fair? I would rather fly to London and know I am getting a fair market rate. I know I am not getting ripped off, and I know I am not ripping off the cab (i.e. I know I'm not a fare that he would rather not take, but is forced to by some regulation at some fixed price).
Maybe, or is it the just the case that uber drivers are just cherry picking the best fares. Living the dregs to cabbies, who are mandated to take them. Maybe, or is it the just the case that uber drivers are just cherry picking the best fares. Living the dregs to cabbies, who are mandated to take them.
Uber drivers are supposed to cherry pick the best fares. It's Uber's job to ensure that the prices are set such that for every person who needs a ride, there is at least some drivers willing to take the job. If nobody wants to give you a ride because they can make more money on other jobs, that means that the price is set below market value.
Or is this a market will fix the problem situation?
This is *exactly* the kind of problem the market fixes.
And anyone who wants a ride somewhere off the beaten path, so there is no chance of a return fare, or a short haul which is less profitable will have pay a premium to get service?
Yes. That's how it's supposed to work. If no one wants to give you a ride for the price you are willing to pay, you are supposed to pay them more money. How would you feel if someone passed a law forcing you to do your job for less money than you are willing to do it for?
But that hearkens to the old days where cabbies simply refused fares outright, or would demand you pay double (ie pay a 'return fare') to make it worth their while.
Why do you feel entitled to a ride that costs less money than a taxi is willing to take? Why shouldn't people who need rides which require more resources pay more money? The fact that you live out in the boonies and taxis want you to pay a double fare is indicative of the fact that it is more costly to provide service to you.
I guess that would "solve" the problem. Not sure that's a good solution though.
It's a market solution where everyone pays a price that is based on how many resources they are consuming. Just like how you pay more money if you eat more food than a regular person, or use more gasoline than other people.
For all the problems that are best solved by government intervention, there is some shit like this happening. I don't know that much about taxi regulations, but I do know that hiring a guy off the internet to drive me to the airport can't be any more harmful than having a friend drop me off at the airport, so what the fuck is the problem?
In fact this is the only time I used Uber. The guy parked at the place where everyone else parks, I get my shit out of the trunk, I tried to tip him, he refused to take it, and he left. Does their really need to be a fucking regulation to prevent this from happening?
Dear California government, please don't make these tea party idiots right more often than they have to be.
Uber drivers and cherry pick and skip any job that doesn't seem worth it for them. One of the advantages of the Uber system is that they try to set prices so that there will always be *some* drivers willing to take the job for that price.
This in my opinion is a much better solution. It is more elegant. It doesn't force anyone to do anything they don't want to do. Rather than forcing taxis to take small jobs, why not allow them to be compensated enough to make them want to do those jobs?
Laws are just not versatile enough to adapt to these sorts of changing conditions. Maybe there was a time when we needed these sorts of laws to have the society we want. Now it appears we don't.
I guess we are lucky that none of the places that our military fights in are dusty....
Mobile phones ring when they are called. Their are also ways to silence them, and this is what people are instructed to do at the start of a movie.
Do you know if Google glass does not have a similar feature to turn off lights and sounds in order not to disturb people (analogous to putting your phone on silent)?
This is a good example of how regulations can stifle the market and have unintended consequences. They all seem like reasonable regulations, but #3 and #4 make unnecessary assumptions about the way taxis work. The spirit of this regulation seems to be to make sure that customers are informed about how much they will need to pay before being obligated to pay it. Uber absolutely follows the spirit of this law, I would argue, even better than traditional taxis. You pay up front because with Uber it can calculate how far the journey will be before actually driving it. This eliminates the need to post rates and to have a meter.
But you are not allowed to follow the spirit of the law. You must follow the letter of the law. The letter of the law would require Uber to follow regulations designed for traditional taxis that for Uber just create unnecessary expenses that then drive up costs for consumers. Many regulations are necessary and work well. It is important to recognize when they become obsolete and to update or remove them. Many times it's better to have fewer regulations if possible, because it makes maintaining them easier and more cost efficient, and therefore more likely.
In many cases when regulations become too complicated, there is a temptation by both citizens and legislators to simply avoid the problems rather than fixing them (e.g. like allowing your garden to become overgrown). By keeping your garden small and manageable, it is more likely that people will be able and willing to keep it maintained.
It sounds like the swedish taxi regulation garden is certainly small enough to manage. So all that is left is to update a few rules for the 21st century. Good job Sweden!
Sure, nobody actually runs a Turing test. It's too hard. If a real Turing test were ever passed there wouldn't be any dispute. They're ALL restricted versions where the judges go easy on the computers.
I guess I am not really sure I see the benefit of calling anything but a real Turing test a Turing test. I think this concept of a "restricted Turing test" is a useful concept, but I think having the "Turing test" in the name, even though it's "restricted" carries a lot of baggage (i.e. in regards to sentience) , that makes people question the legitimacy of the Turing test in general as a test of sentience.
The 13 year old gambit isn't the problem though (it's the judges). In fact, it suggests all sorts of strategies for the judges to trip up the computer. I just had a quick conversation with Eugene where I told him a story about a pretty girl asking a guy to go to the movies, and the two of them sitting right at the back. He changed the subject. Obviously not a 13 year old boy.
It seems to me that changing the subject (even though humans do actually do it), should be considered a dead giveaway. In a real Turing test human subjects should actively try not to change the subject if possible, since this is a easy out for a computer.
I agree with you, a proper Turing test is the best, possibly the only way we currently have to assess an AI. But if you have a computer that you think is at the level of a five year old, for example, find some child psychologists and let them talk to it and some real five year olds. Or thirteen year olds. Or adults. Age doesn't matter.
I think age doesn't matter to a point. Yes you can get a child psychologist and train them to be Turing test judges, but it restricts the number of people that can be judges. Eventually you get to the point where young human children do not have enough mastery of any language (and typing) to be easily distinguishable from computers.
I am not sure exactly what that age is, but in the spirit of the turing test (i.e. removing all possible doubt), I think testing adult humans and "adult" AIs is better.
I do however think that a likely first real contender to passing a legitimate Turing test might be a "child", if the AI actually works in a way that it needs to learn language from scratch (like a baby), which seems quite probable. And I think we will probably create a sentient AI that is not able to pass the turing test before we make one that can.
I'm not talking about this specific chatbot. I am saying that this is how one could rig the turning test using the "13 year old boy angle", which is a just a special case of the "uncooperative subject angle".
I don't know exactly which tricks this chatbot used, but some were hinted at in the article, and thus far no computer has been able to legitimately pass the turing test (i.e. because machines are not sentient yet), and therefore all that appear to pass are using some set of dirty tricks.
Anybody who doesn't think intelligence is about memorizing trivia, the very thing the 13 year old gambit is supposed to help explain away, would see through it very quickly.
Any turing test run today should have nearly 100% of judges singling out the computers. If you are not getting these results, something is wrong with the way the test is being applied. IN the future this will not be the case.
The fact that so many judges were not able to figure out the chatbot was a computer, may even have just been a problem with the judges themselves. Most people don't actually know how to properly judge a turing test. They don;t know that you aren't supposed to "give the computer a fair chance". They don;t know you are supposed to mercilessly exploit any potential weaknesses a computer might have and the strengths of humans. You are supposed to give real humans every possible advantage to prove who they are (i.e. without cheating and knowing who they are).
My point was only that the currently we are seeing obviously non-sentient computer programs appearing to pass the Turing test, because it is applied incorrectly, not because the test itself is a bad one.
In fact I think it is probably the *only* legitimate test for human-like intelligence (when applied properly), and I think Turing was a brilliant man for coming up with it at the dawn of machine computing. He was also brilliant for many other reasons as well.
I don't have a problem with 13 year old kids in the test as long as they are mature enough to participate cooperatively. There are some kids that would not be able to, and even some adults that would not.
The problem I would have is when the "subject" tries to exploit a child's immaturity to pass the test. This wouldn't necessarily be a problem, as the judge could simply count immaturity as a trick that a computer program might employ, but if the pool of humans is also stacked with immature children, this might compromise the test because the judges would have a hard time distinguishing between the uncooperative human children and seemingly uncooperative chatbots.
You could do the exact same think with uncooperative adults, but I think when this is done, it is more easily recognized as the cheap trick that it is. None of the humans should be anything but cooperative and able participants. This is necessary to set an appropriately high bar for the computers. Simply lowering the bar for the computers by including terrible human subjects is basically rigging the test.
You say the device is distracting. How is it distracting? I notice people using their cell phones in a movie. How would I notice someone who is wearing a particular kind of eyeglass thing during a movie?
Another example I just thought of would be an adult English speak human who was just a very slow typer. Obviously slow typers are still intelligent, but it is really hard to judge the humanness of a test subject if they only type 4 sentences during the 5 minute time period.
The spirit of the Turing test is to actually engage the judge and actively convince the judge of it's humanity rather than simply providing plausible excuses for why it can't do that properly.
"I sorry, I don understand you say. I no speak good English" (poor english)
"You are boring, I only want to talk about pokemon and ignore whatever you say." (immature)
"They say I am mentally unstable and that I constantly change the subject and give irrational responses. I don't know what they are talking about. They are just voice in my head. I like icecream" (paranoid schizophrenic)
These are all things that real humans with real intelligence might say. They are also things that are tricially easy form computer programs to imitate without actually having any intelligence at all (i.e. they don't prove intelligence).
The whole point of the Turing test is to allow a computer to prove it is intelligent in an unbiased setting.
Not passing the Turing test does not mean you are not intelligent. Lots of human beings can't pass the test for various reasons (some of which I have used in examples). But legitimately passing a turing test is supposed to be really nearly impossibly hard, but if/when it happens should remove all doubt that whatever passed (human or machine) is intelligent.
We do consider 13 year olds intelligent. We also consider people who speak foreign languages to be intelligent. The problem is that you can't understand people who speak languages that we don't. We assume they are intelligent because nearly every other human we have interacted with was intelligent and we know that languages other than our native language are legitimate languages that we simply don't understand. We assume that these people that can't communicate with us are intelligent despite their inability to pass a legitimate turing test from our perspective.
My point was not to suggest that 13 year old Ukranian boys were not intelligent, but merely that they do not have the tools to pass a 5 minute turing test with English speaking judges. And this lack of ability was used as a way to influence the judges (i.e. a real 13 year old Ukranian boy might not be able to pass, so maybe this is a real 13 year old Ukranian boy).
This is cheating in the Turing test. This is like passing a class by calling in a bomb threat and not having to take the final exam. Sure you didn't even take the final, but none of the people who knew the material took the final either, so maybe you know the material as well. They are passing the Turing test by not really taking the test (by invoking a false language/maturity barrier).
Why didn't they make the chatbot an English speaking adult? Because English speakers are epxected to give normal human responses in English (where Ukranians are not). Adults are expected to be mature and give thoughtful responses (where 13 year old boys might be a bit more eratic and uncooperative if they are bored).
All these attributes of the chatbot are meant to provide plausible excuses for not passing the test rather than actually trying to pass the test.
A 5 minute test might actually be sufficient under the right situation (i.e. no communication/cooperation barriers), but this chatbot was actually trying to amplify barriers to make the 5 minute time limit as inadequate for determining "humanity" as possible.
I may as well make a chat bot that only says one sentence:
"I'm really sorry, but I have to go to the bathroom... BRB"
This is a very plausible thing for a real human to say, but if it is the only thing it says in the 5 minutes, then it makes it very hard to judge whether it is really a machine or a real human. It is creating a false plausible barrier to judgement.