This does beg the question... what criteria should we use when deciding whether an AI has progressed enough to have standing in a discussion about whether to turn it off or not?
I might argue that if the AI has a state vector that cannot be replicated by simply re-running the training and experiential data of the system through the same algorithms - then that AI has a value that might justify some sentimentality.
But then again, maybe it would take something more than complexity/irreproducibility.
I run an R-based URL fetcher to run Google searches all night long and during idle time.
The searches are all for unicorns, leprechauns, and UFOs. I use the 42nd displayed word, whatever it is, to pick the next search. I vary the depth of this recursion when I am whimsical.
I figure flooding the system is better than trying to block all the accesses.
A carbon tax (a Pigovian tax) is the method preferred by most economists.
Keeping the tax to pick the winners is a sure fire way to create the political will to block it.
Citizens' Climate Lobby (Carbon fee and dividend) and others (Climate Leadership Council) propose a revenue-neutral carbon tax and dividend similar to the Alaska Permanent Fund - but such that the tax collected is fully refunded. This makes the tax progressive, revenue neutral and politically sustainable.
Australia's carbon tax scheme is often held up as an example of what happens if the crony capitalist types hold out some of the tax to fund their pick-the-winner favorite businesses.
Once again, the crony capitalists (this time on the left) demonstrate their basic failure to understand human nature (and the politics that derive from that nature). Once again, they are selling unicorns.
I'll promote this from an AC - we need a filter that lets me hide all AC, even though I would have missed this.
The initial comment was
Because of red state socialism. In other words, most of the red states (except Texas) could not afford to exist without the blue states transferring vast amounts of wealth to them.
I've removed the link because I have no reason to trust an unverifiable source.
Th AC then noted correctly...
It depends on what "wealth transfers" you choose to count.
For example, if you counted the Federal state-tax income credit, then California jumps to the #1 beneficiary of Federal transfers, with New York #2 right behind them.
But nah, you choose to count only those tax credits and payments that produce the outcome you like.
I use my statistician chops often to make this point - the superficial analyses that our politicians feed us are often biased, and seeing the full data only helps if you are willing to replicate the analysis in question (it's called, scientific method peer-review).
We used to be able to count on the press to do this, but now that they are all about speed and no vector, they only add to the general disinformation.
We need to figure out how to see that deeper data all the time, but bumper stickers are so much easier, and they make great dog whistles.
WHEREAS: Winning elections takes 51% of the submitted votes
WHEREAS: Only a tight group of wingnuts on either side have disproportionate power because elections are won at the margins, nobody cares about the 90% in the middle.
WHEREAS: Social media will be used to motivate those 10% who really matter, first to enrage them, then to engage them.
THEREFORE: If ALL the tech workers run away, then who owns the field?
As a data miner, I found the human side of this story as explained in Wired, January 2018 worth having a paper copy for, just so I can roll it up and use it to beat some sense into my politicians.
So long there exist those who can remind that world that it doesn't matter so much what some nation says, as it matters what that nation does, China will fail to convince folks that it has any goal other than putting ever-more power over others into the hands of its selfish leaders.
But the whole point of weaponized social media is that it makes you think you are reading what they are doing, you cannot separate the fake information that says that
China is ending pollution in its slave colonies in Afrika
from some whiny truth-pretender who says
China is poisoning millions of acres in Africa
Think of social media manipulation as analogous to CGI, two things have happened to our viewing of movies now that CGI rules the screen.
We are not as amazed by flashy space-opera fights between X-wings and Star destroyers
and
We now sometimes assume that an expensive flashy screen event is CGI when in fact it was real, and we really don't care about that distinction
This incident suggests we should start a 10 year transition during which:
* Self-driving cars are required to have three green flashing lights, front and both sides (LED green)
* Newly produced human driven cars are required to have purple flashing lights, same configuration (LED purple)
* By 2025, everyone will be driven nuts by the flashing lights, and we'll all figure out how to ride bikes while wearing pajamas on our way to the broom factories.
A story like this "Uber car kills!" only serves to feed data to our "availability heuristic" memory. We will hear about this story many times over the next few days... and even the statisticians among us cannot override that exposure to put this in the correct context.
The correct context is probably passenger-miles per equivalent death (pedestrians hit by vehicles). But finding that data is waaaaay beyond the abilities of a 24x7 news cycle.
If Slashdot is so smart, why aren't we able to help with this problem?
Okay, so I add this diagnostic Ubisoft Commit Assist to a Neural Net (NN) that is being trained, using a Genetic Algorithm (GA) to change the topology of the NN. Part of the Neural Net is dedicated to the Commit Assist subnet. This combination means that the entire system gets better exponentially because the diagnostic gets better as the system uses it to find errors in both the base problem and the subnet problem, while the GA improves the ability of the NN to do both the base problem and the subnet problem.
I'd call the made-for-VR movie, "Do Androids evolve to become veterinarians specializing in sheep?"
Actually, as a person who has researched and used machine learning, I enjoyed the article and think it is an adequate addition to the demystifying the black arts of the AI field.
This comment is so perfect that I took a screen shot so I can post it in my Facebook Closed Group" Republican Liberty Caucus - Heavy.
I highlighted the best parts.
I then added that the Universal part of the equation is key-most-critical if you want to really push back on the divisiveness of the post-modernist scum who are running and ruining the public supported information channels (PBS, NPR, MPR). See Michael Shermer of Skeptic column in SciAm.
PS. I photoshopped (actually using Paint) the "are not free to have" to become "now free to have". If I were quoting this in an article I would have used something like "are not [Ed. now] free to have".
Reminds me of the time I ruined a contractor's demo. He had written a ballistic missile program that I had prototyped, and I was ticked that I did not get to write it myself (petty, I know). When they let me try it out, I deliberately picked a target that his software could not find the solution for - locking up the demo.
I just shrugged and gave the "French salute" (at the time, the French salute was a form of shrug with both hands held at shoulder level) and departed the demo.
The Citizens Climate Lobby, Conservative Caucus, is looking for conservatives who are willing to stand up and suffer the slings from the left (your party sucks) and the right (you're a RINO).
They will be using this latest report to (politely) beat some sense into their fellow conservatives. The CCL training is for politeness, the Chair of the Conservative Caucus likes to tell conservative members of Congress that others will rant and chant and block your doors, but CCL will walk with you, talk with you and open doors to new solutions.
Libertarians often lurk within the conservative caucus, since they are often more liberty-oriented and individualistic than the liberals.
No, the models are much more sophisticated than this.
Gender/sex/affectation are way more complex than you can imagine.
Sex is pretty simple - (XY vs XX) x (metabolic response to (testosterone family, endocrine family) x (environmental exposure to various chemicals) x Gender x Affectation
Gender is equally simple - Biological presentation (see Sex) x (cultural presentation) x (physical features) x Sex x Affectation
Affectation is also pretty simple - (How the person perceives their sexuality/gender) x (How the society perceives their affectation) x Sex x Gender
Of course, all of these are f(t)'s, so go ahead and apply a simple logistic regression to predict (Sex(t), Gender(t), Affectation(t)), then get back to me.
Slashdot Mirror
By the time we average schmoes heard about cryptocurrency, the big guys were exiting taking their profits from our enthusiasm.
I might argue that if the AI has a state vector that cannot be replicated by simply re-running the training and experiential data of the system through the same algorithms - then that AI has a value that might justify some sentimentality.
But then again, maybe it would take something more than complexity/irreproducibility.
The searches are all for unicorns, leprechauns, and UFOs. I use the 42nd displayed word, whatever it is, to pick the next search. I vary the depth of this recursion when I am whimsical.
I figure flooding the system is better than trying to block all the accesses.
It's called "chaff".
It's about lynch-mob NGOs that come in the night, with the tacit support of sheriffs and courts who look the other way.
It's about dying with dignity rather than dying like victims, thinking here of Warsaw and Auschwitz.
It's about standing against collectivist fanatics who do not value the individual, thinking here of Cambodia's killing fields.
Keeping the tax to pick the winners is a sure fire way to create the political will to block it.
Citizens' Climate Lobby (Carbon fee and dividend) and others (Climate Leadership Council) propose a revenue-neutral carbon tax and dividend similar to the Alaska Permanent Fund - but such that the tax collected is fully refunded. This makes the tax progressive, revenue neutral and politically sustainable.
Australia's carbon tax scheme is often held up as an example of what happens if the crony capitalist types hold out some of the tax to fund their pick-the-winner favorite businesses.
Once again, the crony capitalists (this time on the left) demonstrate their basic failure to understand human nature (and the politics that derive from that nature). Once again, they are selling unicorns.
The initial comment was
Th AC then noted correctly ...
I use my statistician chops often to make this point - the superficial analyses that our politicians feed us are often biased, and seeing the full data only helps if you are willing to replicate the analysis in question (it's called, scientific method peer-review).
We used to be able to count on the press to do this, but now that they are all about speed and no vector, they only add to the general disinformation.
We need to figure out how to see that deeper data all the time, but bumper stickers are so much easier, and they make great dog whistles.
And you won't solve the social media problem by opting out.
WHEREAS: Only a tight group of wingnuts on either side have disproportionate power because elections are won at the margins, nobody cares about the 90% in the middle.
WHEREAS: Social media will be used to motivate those 10% who really matter, first to enrage them, then to engage them.
THEREFORE: If ALL the tech workers run away, then who owns the field?
As a data miner, I found the human side of this story as explained in Wired, January 2018 worth having a paper copy for, just so I can roll it up and use it to beat some sense into my politicians.
So long there exist those who can remind that world that it doesn't matter so much what some nation says , as it matters what that nation does , China will fail to convince folks that it has any goal other than putting ever-more power over others into the hands of its selfish leaders.
But the whole point of weaponized social media is that it makes you think you are reading what they are doing, you cannot separate the fake information that says that
China is ending pollution in its slave colonies in Afrika
from some whiny truth-pretender who says
China is poisoning millions of acres in Africa
Think of social media manipulation as analogous to CGI, two things have happened to our viewing of movies now that CGI rules the screen.
We are not as amazed by flashy space-opera fights between X-wings and Star destroyers
and
We now sometimes assume that an expensive flashy screen event is CGI when in fact it was real, and we really don't care about that distinction
* Self-driving cars are required to have three green flashing lights, front and both sides (LED green)
* Newly produced human driven cars are required to have purple flashing lights, same configuration (LED purple)
* By 2025, everyone will be driven nuts by the flashing lights, and we'll all figure out how to ride bikes while wearing pajamas on our way to the broom factories.
The correct context is probably passenger-miles per equivalent death (pedestrians hit by vehicles). But finding that data is waaaaay beyond the abilities of a 24x7 news cycle.
If Slashdot is so smart, why aren't we able to help with this problem?
I'd call the made-for-VR movie, "Do Androids evolve to become veterinarians specializing in sheep?"
Do neural nets dream of electric sheep? - the link should have been better presented, however.
Neither do the Courts
In my essay I postulate that I (as a conscious being) am a fractal image of the universe as a whole.
I highlighted the best parts.
I then added that the Universal part of the equation is key-most-critical if you want to really push back on the divisiveness of the post-modernist scum who are running and ruining the public supported information channels (PBS, NPR, MPR). See Michael Shermer of Skeptic column in SciAm.
PS. I photoshopped (actually using Paint) the "are not free to have" to become "now free to have". If I were quoting this in an article I would have used something like "are not [Ed. now] free to have".
Scientists are starting to push back too. As are skeptics like Micheal Shermer.
I back Google on this one.
(No irony between my post and my sig.)
But if the source of the carbon is freshly grown wood, then burning it is considered a zero-carbon footprint use. Citizens Climate Lobby.
Reminds me of the time I ruined a contractor's demo. He had written a ballistic missile program that I had prototyped, and I was ticked that I did not get to write it myself (petty, I know). When they let me try it out, I deliberately picked a target that his software could not find the solution for - locking up the demo. I just shrugged and gave the "French salute" (at the time, the French salute was a form of shrug with both hands held at shoulder level) and departed the demo.
Part of FidoNet raison d' etre was to get around long distance charges. It was to be augmented with radio to jump artificial political and telephone boundaries.
They will be using this latest report to (politely) beat some sense into their fellow conservatives. The CCL training is for politeness, the Chair of the Conservative Caucus likes to tell conservative members of Congress that others will rant and chant and block your doors, but CCL will walk with you, talk with you and open doors to new solutions.
Libertarians often lurk within the conservative caucus, since they are often more liberty-oriented and individualistic than the liberals.
Sex is pretty simple - (XY vs XX) x (metabolic response to (testosterone family, endocrine family) x (environmental exposure to various chemicals) x Gender x Affectation
Gender is equally simple - Biological presentation (see Sex) x (cultural presentation) x (physical features) x Sex x Affectation
Affectation is also pretty simple - (How the person perceives their sexuality/gender) x (How the society perceives their affectation) x Sex x Gender
Of course, all of these are f(t)'s, so go ahead and apply a simple logistic regression to predict (Sex(t), Gender(t), Affectation(t)), then get back to me.
Don't use averages when comparing ... these are not "normally distributed values", so averages are poor measures.