Actually, everything you've ever seen in movies about AI is probably implausible.
Some very common Hollywood implausibilities: 1. Sentience/consciousness emerges by accident. Hard AI is really difficult, and we are not just going to stumble onto it while trying to make, say, a lawnmower. 2. Most robots are humanoid. This is obviously done so human actors can play robots. But other than sexbots, there is no reason for them to be humanoid. 3. AIs have human attributes like jealousy, anger, selfishness, and ambition. Those are emergent properties of Darwinian evolution, and there is no reason to expect an AI to have those attributes unless they are intentionally programmed.
You don't need robots for that. All current wars are civil wars. Most countries have settled borders, no external threats, and their armies are used primarily for control of their own people. Some countries (Costa Rica, Panama, etc.) have abolished their armies with no detrimental effects.
Aside from this, anyone who watches local TV news sees frequent stories along the lines of "Entire family wiped out in car crash."
That isn't quite the same. People drive cars everyday. So when they see a story about a car crash, they can weigh it against their personal experience. People are also familiar with men and dogs, so they can dismiss a story about a man biting a dog as an anomaly. But people don't have personal experience with terrorism or self-driving cars, so when the media reports on rare events involving them, they should provide some context.
what users remain must be brimming with confidence.
Microsoft, Apple, and others, are doing the same thing. They have issued "non-denial denials" that were basically admissions with a wink-wink. It is unfair to pick on Yahoo for just obeying the law.
Maybe you can explain how these work in snow when they still need to see clear lane markings.
I have a Tesla. They work fine in the snow. In fact, Tesla specifically recommends engaging Autopilot on snowy roads because that is safer than driving yourself. Tesla Autopilot has driven several million miles on snowy and icy roads.
It is funny how people trying to point out weaknesses of SDCs, often focus on areas there they are particularly strong. On snowy roads, a human has only their eyes, so if they cannot see the lane markings, they have difficulty navigating. Tesla has cameras for vision, but also has GPS and radar, and can access a database of "landmarks" such as mileage markers, traffic signs, etc. that they can use as waypoints.
And yeah, there is a warning that goes off if the driver takes their hands of the wheel.
There is now. But originally, there was no warning. Tesla added the warning via a software update, because too many drivers were taking naps or watching movies.
I'll believe it when I see it. It seems like a strong claim to make that "we don't know how to do full class 5 driving yet, but we know this hardware is enough to meet the requirements of the thing we don't know".
Just because something isn't done yet doesn't mean they don't know how to do it. Tesla has self-driving software under test, that mostly works with their current cars. It is not yet ready for the public, but that isn't because of any deficiencies in the sensors.
Serious question, not an attempt to troll. What kind of people buy these 'smart TVs' and why?
Because you get additional functionality for zero cost. Compare the cost of a "smart" TV and a "dumb" TV of equal size and resolution. The smart TV will be the same price, or likely even cheaper. A TV already has a CPU, to handle decoding and clicker input. The extra software to make it "smart" has near zero marginal cost. The cost of a CAT-5 and/or WiFi interface is less than $1.
Doesn't almost everybody have a PC or tablet nowadays anyway?
Yes, and I don't want to stop using it because my kids want to watch Netflix. So I spent an extra $0 on a smart TV.
I have a good friend who is a manager at Google. Several times he has contacted me because he noticed that someone Google was considering hiring was in my LinkedIn network, and he wanted my recommendation. I usually give good recommendations, but in one case he later told me my input led to a candidate being rejected.
I have also been "tricked" into making unsolicited connections, thinking the other party had initiated it.
When I first created my LinkedIn account, they asked for my Yahoo and Gmail email addresses, and then ASKED FOR THE PASSWORDS. I saw no reason to provide that information, but my wife fell for it, and LinkedIn then, without her permission, logged into her accounts and sent a link request to every single person in her contact list, consisting of over 3000 people, many of whom she barely knew and hadn't heard from in years. Each email was phrased to imply that she was personally requesting the connection.
LinkedIn was a very sleazy company. They should fit right in at Microsoft.
There are laws that prevent employers from considering health
There are laws that prevent discrimination based on permanent disabilities, and even then there is only a requirement for "reasonable accommodation". So you can refuse to hire someone in a wheelchair to work in your warehouse, but not as a programmer. But there is no broad protection of "health". For instance, my company refuses to hire smokers, specifically because they are unhealthy, running up insurance costs and taking more sick days. That is legal. Nicotine addiction is not a protected disability.
I seriously doubt this will happen, unless the costs for this type of DNA analysis comes way down.
Full genome sequencing cost $100M in 2001, $10k in 2011, and is about $1000 today. If you just want to check for a few specific genes, rather than full sequencing, the cost can be under $100. These prices are expected to continue to fall. Larger companies can buy their own sequencer for about $5k, and do the analysis in-house for faster turn around.
DNA testing in hiring may or may not be a good idea, but cost will not be a significant barrier.
My prediction: The best predictor of future performance will continue to be past performance.
1. Do you have a driver's license or state ID card? 2. Have you ever been arrested (no conviction is needed)? 3. Have you ever served in the military? 4. Have you ever attended an anti-government protest? 5. Do you have a Facebook or Linkedin account?
That's a fine definition of productivity, but it's a meaningless way to compare workers or work styles in different places.
Your problem is that you have somehow conflated "more productive" with "better" or "nobler" or "more moral". Productivity inherently means none of those things.
Higher productivity simply means more is produced per unit of labor. That's it. That is all it means. So if my farm has richer soil, or more rain, than another farm, then I will be more productive, because I can grow more crops with the same amount of labor. It doesn't matter that I didn't "earn" that extra rain, and didn't "deserve" the richer soil.
Norway has high productivity because of their offshore oil. Luxembourg has high productivity because of their financial industry. America has higher productivity because of huge domestic demand, abundant natural resources, flexible labor policies, and deep/efficient capital markets.
That doesn't mean we are "better" than say, French people (We are, of course, but for different reasons).
It's a lot harder to quit your job and strike out on your own now that having health insurance is the law.
That is a lame excuse. But seriously, you should NOT try to start your own business. If you really think a molehill like this is a major impediment, you will never get over the mountains. Keep your day job.
I am very interested in how they defined the productivity per worker.
It is GDP (usually adjusted by PPP) divided by total hours worked by all workers in the country.
my gut feeling is that they simply divided the gross national product by the number of employees, which is a wildly inaccurate way
No. That is NOT what productivity means.
Norway is not significantly more productive than Sweden or Denmark - it just has a lot of oil and those two countries do not.
Productivity does NOT mean "working hard". Offshore oil workers may not "work harder" than, say, farmers, but they are likely to contribute far more to GDP, and are thus more productive.
Do you really want a politician out there who say "I was in favor eugenics in 1936, and I'm not going to change my mind just because it's 2016!"
Sure. Eugenics makes way more sense in 2016 than it did in 1936. Back then, the only way to change the gene pool was sterilization or extermination. Now, we can just directly edit defects out of our DNA.
This place was a huge hotbed of "Gore won!" and "BOOOSH is teh EVUL!"
Slashdot, like nearly every public forum, is used to bash whomever is in office. So during the Bush administration, we attacked Bush, and now we bash Obama. But that doesn't necessarily mean there was an underlying shift in ideology. People who are happy with the status quo usually aren't very vocal.
How long would it take the US to launch a manned moon mission?
Well, it is going to take 30 years to finish the BART extension from Fremont to San Jose, and a moon landing is more complex than that. So it could be a while.
China can send people into space and the USA cannot.
USA has rovers on Mars, just did a flyby of Pluto, and now has a probe heading into the Kuiper Belt. China does not.
Which will have a bigger long term impact, the 6.5 meter James Webb Space Telescope, or this manned Chinese space station that is just as pointless as the ISS?
Slashdot Mirror
Actually, everything you've ever seen in movies about AI is probably implausible.
Some very common Hollywood implausibilities:
1. Sentience/consciousness emerges by accident. Hard AI is really difficult, and we are not just going to stumble onto it while trying to make, say, a lawnmower.
2. Most robots are humanoid. This is obviously done so human actors can play robots. But other than sexbots, there is no reason for them to be humanoid.
3. AIs have human attributes like jealousy, anger, selfishness, and ambition. Those are emergent properties of Darwinian evolution, and there is no reason to expect an AI to have those attributes unless they are intentionally programmed.
They will then be used to attack civilians.
You don't need robots for that. All current wars are civil wars. Most countries have settled borders, no external threats, and their armies are used primarily for control of their own people. Some countries (Costa Rica, Panama, etc.) have abolished their armies with no detrimental effects.
Aside from this, anyone who watches local TV news sees frequent stories along the lines of "Entire family wiped out in car crash."
That isn't quite the same. People drive cars everyday. So when they see a story about a car crash, they can weigh it against their personal experience. People are also familiar with men and dogs, so they can dismiss a story about a man biting a dog as an anomaly. But people don't have personal experience with terrorism or self-driving cars, so when the media reports on rare events involving them, they should provide some context.
what users remain must be brimming with confidence.
Microsoft, Apple, and others, are doing the same thing. They have issued "non-denial denials" that were basically admissions with a wink-wink. It is unfair to pick on Yahoo for just obeying the law.
Unfortunately, they are planning on using neural net software
The NNs are for object recognition, like the proverbial "rock vs plastic bag". They aren't using NNs to control the brakes and steering.
a neural net, cannot be proven to be fool-proofed
There are NO systems that are "fool proof" at object recognition, including human brains.
Maybe you can explain how these work in snow when they still need to see clear lane markings.
I have a Tesla. They work fine in the snow. In fact, Tesla specifically recommends engaging Autopilot on snowy roads because that is safer than driving yourself. Tesla Autopilot has driven several million miles on snowy and icy roads.
It is funny how people trying to point out weaknesses of SDCs, often focus on areas there they are particularly strong. On snowy roads, a human has only their eyes, so if they cannot see the lane markings, they have difficulty navigating. Tesla has cameras for vision, but also has GPS and radar, and can access a database of "landmarks" such as mileage markers, traffic signs, etc. that they can use as waypoints.
And yeah, there is a warning that goes off if the driver takes their hands of the wheel.
There is now. But originally, there was no warning. Tesla added the warning via a software update, because too many drivers were taking naps or watching movies.
I'll believe it when I see it. It seems like a strong claim to make that "we don't know how to do full class 5 driving yet, but we know this hardware is enough to meet the requirements of the thing we don't know".
Just because something isn't done yet doesn't mean they don't know how to do it. Tesla has self-driving software under test, that mostly works with their current cars. It is not yet ready for the public, but that isn't because of any deficiencies in the sensors.
Serious question, not an attempt to troll. What kind of people buy these 'smart TVs' and why?
Because you get additional functionality for zero cost. Compare the cost of a "smart" TV and a "dumb" TV of equal size and resolution. The smart TV will be the same price, or likely even cheaper. A TV already has a CPU, to handle decoding and clicker input. The extra software to make it "smart" has near zero marginal cost. The cost of a CAT-5 and/or WiFi interface is less than $1.
Doesn't almost everybody have a PC or tablet nowadays anyway?
Yes, and I don't want to stop using it because my kids want to watch Netflix. So I spent an extra $0 on a smart TV.
I have a good friend who is a manager at Google. Several times he has contacted me because he noticed that someone Google was considering hiring was in my LinkedIn network, and he wanted my recommendation. I usually give good recommendations, but in one case he later told me my input led to a candidate being rejected.
I have also been "tricked" into making unsolicited connections, thinking the other party had initiated it.
When I first created my LinkedIn account, they asked for my Yahoo and Gmail email addresses, and then ASKED FOR THE PASSWORDS. I saw no reason to provide that information, but my wife fell for it, and LinkedIn then, without her permission, logged into her accounts and sent a link request to every single person in her contact list, consisting of over 3000 people, many of whom she barely knew and hadn't heard from in years. Each email was phrased to imply that she was personally requesting the connection.
LinkedIn was a very sleazy company. They should fit right in at Microsoft.
If a company cannot use genetic markers which signify greater intelligence (if such a thing even exists), why would they be allowed to use an IQ test
Using IQ tests, or any other test for general intelligence, is ILLEGAL in America.
There are laws that prevent employers from considering health
There are laws that prevent discrimination based on permanent disabilities, and even then there is only a requirement for "reasonable accommodation". So you can refuse to hire someone in a wheelchair to work in your warehouse, but not as a programmer. But there is no broad protection of "health". For instance, my company refuses to hire smokers, specifically because they are unhealthy, running up insurance costs and taking more sick days. That is legal. Nicotine addiction is not a protected disability.
I seriously doubt this will happen, unless the costs for this type of DNA analysis comes way down.
Full genome sequencing cost $100M in 2001, $10k in 2011, and is about $1000 today. If you just want to check for a few specific genes, rather than full sequencing, the cost can be under $100. These prices are expected to continue to fall. Larger companies can buy their own sequencer for about $5k, and do the analysis in-house for faster turn around.
DNA testing in hiring may or may not be a good idea, but cost will not be a significant barrier.
My prediction: The best predictor of future performance will continue to be past performance.
I don't know why people continue to buy their stuff.
Here's the reason: People buy Apple computers because they run Mac OS X.
... email though is too much of a diversion since at least some of it actually important.
An obvious solution is to use a reverse spam-filter to ensure that you only see the unimportant messages.
If you are in the database?
1. Do you have a driver's license or state ID card?
2. Have you ever been arrested (no conviction is needed)?
3. Have you ever served in the military?
4. Have you ever attended an anti-government protest?
5. Do you have a Facebook or Linkedin account?
That's a fine definition of productivity, but it's a meaningless way to compare workers or work styles in different places.
Your problem is that you have somehow conflated "more productive" with "better" or "nobler" or "more moral". Productivity inherently means none of those things.
Higher productivity simply means more is produced per unit of labor. That's it. That is all it means. So if my farm has richer soil, or more rain, than another farm, then I will be more productive, because I can grow more crops with the same amount of labor. It doesn't matter that I didn't "earn" that extra rain, and didn't "deserve" the richer soil.
Norway has high productivity because of their offshore oil. Luxembourg has high productivity because of their financial industry. America has higher productivity because of huge domestic demand, abundant natural resources, flexible labor policies, and deep/efficient capital markets.
That doesn't mean we are "better" than say, French people (We are, of course, but for different reasons).
It's a lot harder to quit your job and strike out on your own now that having health insurance is the law.
That is a lame excuse. But seriously, you should NOT try to start your own business. If you really think a molehill like this is a major impediment, you will never get over the mountains. Keep your day job.
I am very interested in how they defined the productivity per worker.
It is GDP (usually adjusted by PPP) divided by total hours worked by all workers in the country.
my gut feeling is that they simply divided the gross national product by the number of employees, which is a wildly inaccurate way
No. That is NOT what productivity means.
Norway is not significantly more productive than Sweden or Denmark - it just has a lot of oil and those two countries do not.
Productivity does NOT mean "working hard". Offshore oil workers may not "work harder" than, say, farmers, but they are likely to contribute far more to GDP, and are thus more productive.
It might just mean Europeans are 25% more efficient.
Nope. Americans not only work more, but also get more done per hour than any European countries except Luxembourgh and Norway.
Do you really want a politician out there who say "I was in favor eugenics in 1936, and I'm not going to change my mind just because it's 2016!"
Sure. Eugenics makes way more sense in 2016 than it did in 1936. Back then, the only way to change the gene pool was sterilization or extermination. Now, we can just directly edit defects out of our DNA.
This place was a huge hotbed of "Gore won!" and "BOOOSH is teh EVUL!"
Slashdot, like nearly every public forum, is used to bash whomever is in office. So during the Bush administration, we attacked Bush, and now we bash Obama. But that doesn't necessarily mean there was an underlying shift in ideology. People who are happy with the status quo usually aren't very vocal.
How long would it take the US to launch a manned moon mission?
Well, it is going to take 30 years to finish the BART extension from Fremont to San Jose, and a moon landing is more complex than that. So it could be a while.
China can send people into space and the USA cannot.
USA has rovers on Mars, just did a flyby of Pluto, and now has a probe heading into the Kuiper Belt. China does not.
Which will have a bigger long term impact, the 6.5 meter James Webb Space Telescope, or this manned Chinese space station that is just as pointless as the ISS?