For example, the programmer building a chess engine can understand why it performs each movement, but will always lose in a game against it.
So, having an algorithm delivering 92% accuracy would imply that people could detect these situations even more accurately than that(?!)
If a chess engine developer can be outperformed by his own algorithm, then a suicide predictor developer can also be outperformed by his own algorithm. It's the same concept.
I -- and everyone else -- don't know how HUMAN cognition/consciousness/self-awareness/actual THOUGHT/creativity works
If you don't know how it works, then you can't claim that someone didn't capture the essential elements in an algorithm. The only meaningful thing you can do is look at the output, and the output is pretty good.
We know how this fake AI works: Poorly, by comparison.
It seems to work better than a human psychologist.
But the question is: how are you expecting an algorithm, precisely developed by a person, to succeed where people will fail? It doesn't seem too logical, right?
Teach the algorithm by providing it with a list of properties from patients in the past, together with the patient outcome (suicide after N days, or no suicide). The algorithm then searches for patterns in the properties that have a high chance of resulting in suicide.
The developer doesn't even need to be educated in the field of psychology.
So, having an algorithm delivering 92% accuracy would imply that people could detect these situations even more accurately than that(?!)
No, it wouldn't. The different patterns of behavior could be so complicated and subtle that people can't pick them up, especially in an area where people tend to have biases.
If they are heavily taxed, and addicts can no longer afford them because they've lost their job from the side effects, they would still have to resort to crime to buy them.
Please vote for people who will end the "war on drugs." It is the very antithesis of liberty
The problem with this liberty is that people who abuse the drugs and damage their health, or their ability to hold a job, or become a burden in some other way, reduce the liberties of the rest of society.
Good example. After fixing the #include, that code can be compiled. For generated code, that could be an acceptable way to format it. For human readability, I can run it through a pretty-printer with my favourite settings. I can copy/paste it into any other project, and it would still work.
Now try pasting a piece of python code, and see if the whitespace survives the HTML formatting.
We need to remove CO2 from the atmosphere, the only viable way to do this is photosynthesis,
A much better way to remove next year's CO2 from the atmosphere is to leave the carbon in the ground. Close a coal plant, and replace it with PV panels. After we've done that, we can focus on last year's CO2.
I wonder, people who claim the landing was reversed footage, did they actually try reversing the video and see what it looks like ? Must be amazing engines to suck in all that smoke.
Even a rat brain is a far more sophisticated neural network machine than anything we will probably build from scratch in the next few hundred years.
A rat brain has about 500 billion synapses. Assuming a generous 1000 Hz firing rate, we're talking 0.5 peta synapse operations per second. Google's 2nd generation TPU ASIC can do 0.045 petaflops in a single chip.
I don't think it's going to take hundreds of years.
I am convinced that wetware is going to be the real future and not so much neural net ASICs like Google's TPU or whatever Nvidia is working on to run neural network architecture
Why ? Everything you can do in wetware, you can do better in an ASIC. For a lot of limited domain pattern recognition jobs, the ASICs are already outperforming the human brain both in speed and accuracy. ASICs are much more flexible (you can experiment with different topologies and functions), plus you can also easily combine ASICs with conventional memory and processing.
going to continue to be thinking in terms the same sort of snail pace of incremental improvements in specific problem domains that we have seen so far.
Current AI developments are anything but "snail pace". It's the fastest developing field, with amazing new things coming out almost every day.
After the heavy lifting is done, tools will be made so that people will be able to integrate AI into all kinds of applications (and not just as a front-end for a big company). There's still plenty of work to be done.
Slashdot Mirror
It's not what people complain about. It's whether his policies were left or right wing.
And then he tried to shoot himself to finish it off, but missed.
It's unlikely that a government would try to control its own people with nukes, so you don't need nukes to defend yourself.
There is a very obvious difference between someone talking about a group that they themselves are in
Dividing humans in different groups is a racist thing to do.
Mel Brooks is jewish
Being jewish gives him to right to make nagger jokes ?
Racism writ large right there.
Not necessarily. Could be simply hatred, unrelated to race. Trump isn't treated better.
For example, the programmer building a chess engine can understand why it performs each movement, but will always lose in a game against it.
So, having an algorithm delivering 92% accuracy would imply that people could detect these situations even more accurately than that(?!)
If a chess engine developer can be outperformed by his own algorithm, then a suicide predictor developer can also be outperformed by his own algorithm. It's the same concept.
I -- and everyone else -- don't know how HUMAN cognition/consciousness/self-awareness/actual THOUGHT/creativity works
If you don't know how it works, then you can't claim that someone didn't capture the essential elements in an algorithm. The only meaningful thing you can do is look at the output, and the output is pretty good.
We know how this fake AI works: Poorly, by comparison.
It seems to work better than a human psychologist.
Sounds like the results are better than what doctors can do:
https://www.scientificamerican...
But the question is: how are you expecting an algorithm, precisely developed by a person, to succeed where people will fail? It doesn't seem too logical, right?
Teach the algorithm by providing it with a list of properties from patients in the past, together with the patient outcome (suicide after N days, or no suicide). The algorithm then searches for patterns in the properties that have a high chance of resulting in suicide.
The developer doesn't even need to be educated in the field of psychology.
So, having an algorithm delivering 92% accuracy would imply that people could detect these situations even more accurately than that(?!)
No, it wouldn't. The different patterns of behavior could be so complicated and subtle that people can't pick them up, especially in an area where people tend to have biases.
Since you admit that you don't know how it works, how would you tell the difference between the "real thing", and a suitably advanced "ersatz" ?
Or can you write a formal, terminating, deterministic sequence of elementary steps for reliably generating "Eureka!" moments in humans?
People can't reliably generate Eureka moments, so it would be impossible to put that in an algorithm.
If they are heavily taxed, and addicts can no longer afford them because they've lost their job from the side effects, they would still have to resort to crime to buy them.
Please vote for people who will end the "war on drugs." It is the very antithesis of liberty
The problem with this liberty is that people who abuse the drugs and damage their health, or their ability to hold a job, or become a burden in some other way, reduce the liberties of the rest of society.
So, unfortunately, are zillions of easily mismatched, hard to read, bracket pairs.
You could have an editor compare indentation with bracket pairing, and notify you of a discrepancy. That way you can verify both.
Good example. After fixing the #include, that code can be compiled. For generated code, that could be an acceptable way to format it. For human readability, I can run it through a pretty-printer with my favourite settings. I can copy/paste it into any other project, and it would still work.
Now try pasting a piece of python code, and see if the whitespace survives the HTML formatting.
Whitespace between keywords is fine. C requires that too. We're talking about using white space to define program structure.
It's because the indentation happens automatically based on logical structure, not the other way around.
We already have moss, so we're done.
We need to remove CO2 from the atmosphere, the only viable way to do this is photosynthesis,
A much better way to remove next year's CO2 from the atmosphere is to leave the carbon in the ground. Close a coal plant, and replace it with PV panels. After we've done that, we can focus on last year's CO2.
Actually, rocket experts all did know that you could land a rocket on its tail. After all, the lunar module landed that way
The hard part is dealing with the erratic atmosphere at supersonic speeds. It's a lot easier in a vacuum.
I wonder, people who claim the landing was reversed footage, did they actually try reversing the video and see what it looks like ? Must be amazing engines to suck in all that smoke.
Even a rat brain is a far more sophisticated neural network machine than anything we will probably build from scratch in the next few hundred years.
A rat brain has about 500 billion synapses. Assuming a generous 1000 Hz firing rate, we're talking 0.5 peta synapse operations per second. Google's 2nd generation TPU ASIC can do 0.045 petaflops in a single chip.
I don't think it's going to take hundreds of years.
I am convinced that wetware is going to be the real future and not so much neural net ASICs like Google's TPU or whatever Nvidia is working on to run neural network architecture
Why ? Everything you can do in wetware, you can do better in an ASIC. For a lot of limited domain pattern recognition jobs, the ASICs are already outperforming the human brain both in speed and accuracy. ASICs are much more flexible (you can experiment with different topologies and functions), plus you can also easily combine ASICs with conventional memory and processing.
going to continue to be thinking in terms the same sort of snail pace of incremental improvements in specific problem domains that we have seen so far.
Current AI developments are anything but "snail pace". It's the fastest developing field, with amazing new things coming out almost every day.
After the heavy lifting is done, tools will be made so that people will be able to integrate AI into all kinds of applications (and not just as a front-end for a big company). There's still plenty of work to be done.