I tried really, really hard to 'cut the cord' (or in this case 'ditch the dish') but after careful study, I found with a family of four, including two children, this just won't work.
The reason why it won't work is that you haven't made the decision to cut your television watching, just thought about changing the means. If you decide you're going to be a no-TV family, then do that. Sure, watch your occasional movies, read your news articles online, buy a sports feed, buy or download some programs for the kids if you want, but the "flip on the tube and watch whatever's on for a while" is something that's worth looking at eliminating for its own benefit.
One of the things that the video seemed to imply is that they're simulating the expected procedure in their underwater tanks on the ground first, to make sure everything is reachable, and to know what tools they need to do the job.
The crew are not necessarily experts on every nut and bolt of the ISS. While they've certainly studied it, and have done the underwater sims themselves, the hands-on experience with each individual unit is quite rare compared to, say, a mechanic who works on cars every day for decades.
Chris Hadfield's personal mission is to communicate space to regular people and get them excited about it again. He's a hero even if just for that and everybody should watch his videos.
Okay, that's fine. But it's definitely not "standard process everyone follows". Most people seem to either watch them or block them.
However, I have heard that what YouTube does give to its paid content creators is based on views, irrespective of whether or not the ads were actually played, so both of your points could apply more to blockers than ignorers.
The 1970's climate scare was global cooling, which was ultimately dropped as incorrect. That gives a large precedent to not caring about (regardless of denying or not) the next climate scare.
You talk about a single right-wing channel among dozens to hundreds of other channels who fall in some non-conservative spectrum (super-left to moderate). That's pointless.
However, both the far right and far left are evil. Evils of the far right should not incur the promotion of the evils of the far left.
The majority of the arguments against building codes are for people building on their own property for their own use at their own expense.
Yes, building codes are far more important during purchase or rental. In that case, if building codes didn't exist, then full disclosure and understandable ratings of building materials and practices would at least be necessary. If a rental agreement was required to say "Electrical safety rating: 0 Stars, Completely Hazardous via ASDF rating 2013" it wouldn't prevent anybody from doing anything, but would strongly disincentivize people from renting there.
I'm not promoting either direction, but that's at least how a complete argument from the libertarian side would go. Many of them are big on contract law, trying to ensure that contractual exchanges and agreements are properly informed and consensual. Of course, there are the "Well, you should have known better" side of things, but when it gets down to details like many take the route of making sure the legal agreement is more sensible
The basic ANN model has only a single scalar state for its currently propagating value. Biological neurons seem to not just work on amplitude of signals, but also frequency and time patterns, and the current chemical balances in which they exist. This is the state that isn't captured in the single current "value" of a simple ANN node.
I guess you could consider the term "dynamic" to be superfluous, and I wouldn't disagree. I just used it to imply the time domain.
I'd doubt the claims of that IC manufacturer, especially as new properties about neuron electric & connection activity have been discovered within the last decade, as far as I'm aware.
If you'd look at the sentence again, "it's unknown if we can make up for that lack of dynamic state" is the important part. I'm not saying we can't increase simulated neuron complexity, I'm saying we can increase neuron complexity, or make larger nets to try to compensate for lack of current neural complexity. It seems we don't know which of these will be more fruitful yet, and both are being attempted.
but all those"oh you should just upgrade" are obviously people that have never actually done a large rollout because if they had they'd know that there is NO "just" when it comes to a large business, you are talking weeks to months of slow, tedious, headache inducing work and it is NOT a pleasant experience for anybody involved.
You're the one spouting nonsense you know nothing about. Only "weeks to months"? Things never happen that fast!;-)
If you look again it's really the first part (conceptualization & learning (mammalian-level)) that you wish to add to the second part (dynamic responses to environment (insectoid-level)) to complete the whole picture.
Getting just the former to work yields brains in a jar like the Star Trek ship computer. Getting just the latter to work gets you semi-autonomous "dumb but capable" things like Big Dog. These are the two main facets of we view as intelligence, and to get "real" AI or artificial life does seem to require a tight integration of both.
I'm in a hurry, but I'll dump some disconnected thoughts at you. I appreciate my having to take these still-vague thoughts and being more specific with them in discussion.
(perfect recall) Conflicts with prioritizing if you have provisions for priority zero (forgetting, irrelevant if the link goes away or the information is erased).
Recall of perceptions in particular is a very useful measure, considering "Oh, that's what you meant" style moments as new information is gained in the future, allowing reinterpretation of past input. This actual storage & recall is not a large technical challenge, especially when no continuous real-world sensors are involved (just dealing with operator statements, information you specifically point at it, etc).
Recall of complete system state would of course likely be an impossibility of recursion, but concepts and categorizations that it was instructed to remember should be remembered indefinitely, and that storage/recall should pose no technical challenge either. I think it's also interesting to view a system where the decision to allow to forget is an active choice, rather than our nature which seems to tend towards memorization (of otherwise primally non-impressing information) being an active choice. Forgetting could be an actively decided optimization parameter, as opposed to a byproduct of capacity.
(could converse about its knowledge and thought processes) Telling more than we can know (Nisbett &Wilson, 1977, Psychological Review, 84, 231–259), protocol analysis, expert interviews: evidence that this is at least not always possible. My hypothesis is that too much metaprocessing would lead to a deadlock.
I don't think that an artificial intelligence (not artificial life, or artificial modeled brain) is necessarily limited by those human shortcomings. There needs to be a bootstrap layer, of course, and from a practical development management perspective the software should be able to mostly write itself on top of that. The interest that lead me into AI in the first place is getting software to understand software as a developer tool, but that desired understanding quickly necessitates real AGI style conceptualization. However, as software there should be some configurable layer of diagnostics, heuristics, optimizations, and strategies available that it itself could maintain. There's nothing magic to non-AI software doing that; making it visible to the AI running within the software seems obvious. If the AI is enabled by some running algorithms, it can tune those, as well as introspect the current state of goals and decisions implemented by those algorithms. In order to avoid explosion, those would be intentionally perceived, not a constantly active information stream.
(conversational feedback would have immediate application without lengthy retraining ) Would imply that the system immediately trusts. Would probably be rather self destructive, thus not intelligent.
The tool would trust me, specifically. Or more generally, the owner and permission holder to the system. If there are weird issues with whatever information, commands, or boundaries I give it that cannot be resolved by whatever capacity of "common sense" it already has formed (required for true natural language interaction), then that would be a point of discussion and clarification; or of course the tool owner could escalate permission and hope the system doesn't screw up by acting without full understanding.
The other thought behind that statement is to contrast statistical learning methods with rational learning methods. Adding a rational statement to a theorem prover, for instance, takes (logically) immediate and absolute effect. Telling a car-recognizing neural net that a particular type of vehicle is now no longer classified for its purposes requires lengthy retraining, because it has no means of being "aware" of that distinction at all.
Right, that's why dealing with such things requires intelligence, and if it could do such a thing would generally be considered intelligent. It's also why symbolic AI has failed to produce any general intelligence, because it simply cannot scale. In order for a system to exhibit such behavior, it needs to adaptively and "intelligently" prioritize what it's doing and on what it's working, as well as to predictively index and preprocess information, in order to even begin to achieve any sense of tractability.
The snopes article still says that the complete change from sugar to HFCS was just months before the switch to New Coke. The two were not related, but the switch in sweetener was done at roughly the same time as the New Coke fiasco. Sugar-based "Old Coke" was no more right before New Coke came along.
We also don't know how many licenses Microsoft needs to break even for its Windows division. I'd suspect the vast majority of these are cheap OEM licenses, not corporate sales, as businesses are steering clear of Win8, so we're probably talking a couple dozen dollars of revenue per sale on average.
I'd be interested to know how much a couple billion dollars is compared to the Win8-related expenses Microsoft has had in pre-launch and continuing development, marketing, and support of the OS.
The improvement has been using multiple ANNs together as communicating units that can both communicate information and dynamically train each other, instead of trying to make a single large ANN and using external training sets. Of course, this isn't that new, as these guys have been working on such models since at least the early 90s.
We don't know what intelligence actually is, but we have an example of something that is unarguably intelligent: the mammalian brain.
To further this point, a brain is also not necessarily just a pile of neurons. There is specialization in the brain, and while other portions of the brain can make up for damaged parts, that only goes so far.
Any proposed mechanism of intelligence should be discounted unless it behaves the same way as a brain. Most AI research fails this test.
I would offer some more subjectivity to that. Mammalian brains tire, need sleep, do not have perfect recall, run things out of time order and convinces itself otherwise, take a long time to train, and has strong emotional needs.
With a view of "intelligence" that would include tools and helpers, and not just autonomous life forms, we do not want certain intelligence to behave the same way as the brain. We also want to extract what intelligence and certain components like rationality, creativity, intuitiveness, judgement, learning, and conceptualization are and be able to use those outside of the limitations that biology brings.
I'd mod you up if I could, but I think I can help out with a few points instead:
1) There is no concrete constructive definition of intelligence yet, and I think anybody at a higher level in the field knows that. Establishing that definition is a recognized part of AI research. Intelligence is still recognized comparatively, usually related to something like the capability to resolve difficult or ambiguous problems with similar or greater effect than humans, or can learn and react to dynamic environmental situations to similar effect as other living things. Once we've created something that works and that we can tangibly study, we can begin to come up with real workable definitions of intelligence that represent both the technological and biological instances of recognized intelligence.
4) Modern ANN research sometimes includes altering the morphology of the network as part of training, not just altering the coefficients. I would hope something like that is in effect here.
Biological neurons are far more complex than ANN neurons. At this point it's unknown if we can make up for that lack of dynamic state by using a larger ANN, or by increasing the per-neuron complexity to try to match the biological counterpart. I do have my doubts about the former, but that doubt is merely intuition, not science. We simply don't know yet.
as is whether this is what folks have in mind by "AI."
In my own studies, this isn't the branch of AI I'm particularly interested in. I don't care about artificial life, or structures based around the limitations of the biological brain. I'd love to have a system with perfect recall, could converse about its knowledge and thought processes such that conversational feedback would have immediate application without lengthy retraining, and could tirelessly and meticulously follow instructions given in natural language.
I don't see modeling biological brains as being a workable approach to that view of AI, except maybe the "tirelessly" part. I'm more interested in cognitive meta-models of intelligence itself than the substrate on which currently known knowledge happens to reside.
Most people read information on computer displays, reading web pages, emails, facebook updates, twitter feeds, wikipedia, and reference materials; and work in word processors, spreadsheets, and programming environments. All of these features are regularly constrained by vertical resolution.
For people watching cat videos and playing simple games (which comprises almost everybody else not doing the above), neither >1080p resolution nor fidelity matters.
For people doing high-end gaming and watching high-end media, your situation applies. However, it's a pretty tiny sliver of overall computer monitor time, all things considered.
Slashdot Mirror
I tried really, really hard to 'cut the cord' (or in this case 'ditch the dish') but after careful study, I found with a family of four, including two children, this just won't work.
The reason why it won't work is that you haven't made the decision to cut your television watching, just thought about changing the means. If you decide you're going to be a no-TV family, then do that. Sure, watch your occasional movies, read your news articles online, buy a sports feed, buy or download some programs for the kids if you want, but the "flip on the tube and watch whatever's on for a while" is something that's worth looking at eliminating for its own benefit.
I hope that you actually vote 3rd party instead of just sigh during elections.
One of the things that the video seemed to imply is that they're simulating the expected procedure in their underwater tanks on the ground first, to make sure everything is reachable, and to know what tools they need to do the job.
The crew are not necessarily experts on every nut and bolt of the ISS. While they've certainly studied it, and have done the underwater sims themselves, the hands-on experience with each individual unit is quite rare compared to, say, a mechanic who works on cars every day for decades.
Chris Hadfield's personal mission is to communicate space to regular people and get them excited about it again. He's a hero even if just for that and everybody should watch his videos.
Okay, that's fine. But it's definitely not "standard process everyone follows". Most people seem to either watch them or block them.
However, I have heard that what YouTube does give to its paid content creators is based on views, irrespective of whether or not the ads were actually played, so both of your points could apply more to blockers than ignorers.
The 1970's climate scare was global cooling, which was ultimately dropped as incorrect. That gives a large precedent to not caring about (regardless of denying or not) the next climate scare.
You talk about a single right-wing channel among dozens to hundreds of other channels who fall in some non-conservative spectrum (super-left to moderate). That's pointless.
However, both the far right and far left are evil. Evils of the far right should not incur the promotion of the evils of the far left.
The majority of the arguments against building codes are for people building on their own property for their own use at their own expense.
Yes, building codes are far more important during purchase or rental. In that case, if building codes didn't exist, then full disclosure and understandable ratings of building materials and practices would at least be necessary. If a rental agreement was required to say "Electrical safety rating: 0 Stars, Completely Hazardous via ASDF rating 2013" it wouldn't prevent anybody from doing anything, but would strongly disincentivize people from renting there.
I'm not promoting either direction, but that's at least how a complete argument from the libertarian side would go. Many of them are big on contract law, trying to ensure that contractual exchanges and agreements are properly informed and consensual. Of course, there are the "Well, you should have known better" side of things, but when it gets down to details like many take the route of making sure the legal agreement is more sensible
Standard process is having an ad blocker. It make not just YouTube but most web browsing faster and easier.
The basic ANN model has only a single scalar state for its currently propagating value. Biological neurons seem to not just work on amplitude of signals, but also frequency and time patterns, and the current chemical balances in which they exist. This is the state that isn't captured in the single current "value" of a simple ANN node.
I guess you could consider the term "dynamic" to be superfluous, and I wouldn't disagree. I just used it to imply the time domain.
I'd doubt the claims of that IC manufacturer, especially as new properties about neuron electric & connection activity have been discovered within the last decade, as far as I'm aware.
If you'd look at the sentence again, "it's unknown if we can make up for that lack of dynamic state" is the important part. I'm not saying we can't increase simulated neuron complexity, I'm saying we can increase neuron complexity, or make larger nets to try to compensate for lack of current neural complexity. It seems we don't know which of these will be more fruitful yet, and both are being attempted.
but all those"oh you should just upgrade" are obviously people that have never actually done a large rollout because if they had they'd know that there is NO "just" when it comes to a large business, you are talking weeks to months of slow, tedious, headache inducing work and it is NOT a pleasant experience for anybody involved.
You're the one spouting nonsense you know nothing about. Only "weeks to months"? Things never happen that fast! ;-)
If you look again it's really the first part (conceptualization & learning (mammalian-level)) that you wish to add to the second part (dynamic responses to environment (insectoid-level)) to complete the whole picture.
Getting just the former to work yields brains in a jar like the Star Trek ship computer. Getting just the latter to work gets you semi-autonomous "dumb but capable" things like Big Dog. These are the two main facets of we view as intelligence, and to get "real" AI or artificial life does seem to require a tight integration of both.
I'm in a hurry, but I'll dump some disconnected thoughts at you. I appreciate my having to take these still-vague thoughts and being more specific with them in discussion.
(perfect recall)
Conflicts with prioritizing if you have provisions for priority zero (forgetting, irrelevant if the link goes away or the information is erased).
Recall of perceptions in particular is a very useful measure, considering "Oh, that's what you meant" style moments as new information is gained in the future, allowing reinterpretation of past input. This actual storage & recall is not a large technical challenge, especially when no continuous real-world sensors are involved (just dealing with operator statements, information you specifically point at it, etc).
Recall of complete system state would of course likely be an impossibility of recursion, but concepts and categorizations that it was instructed to remember should be remembered indefinitely, and that storage/recall should pose no technical challenge either. I think it's also interesting to view a system where the decision to allow to forget is an active choice, rather than our nature which seems to tend towards memorization (of otherwise primally non-impressing information) being an active choice. Forgetting could be an actively decided optimization parameter, as opposed to a byproduct of capacity.
(could converse about its knowledge and thought processes)
Telling more than we can know (Nisbett &Wilson, 1977, Psychological Review, 84, 231–259), protocol analysis, expert interviews: evidence that this is at least not always possible. My hypothesis is that too much metaprocessing would lead to a deadlock.
I don't think that an artificial intelligence (not artificial life, or artificial modeled brain) is necessarily limited by those human shortcomings. There needs to be a bootstrap layer, of course, and from a practical development management perspective the software should be able to mostly write itself on top of that. The interest that lead me into AI in the first place is getting software to understand software as a developer tool, but that desired understanding quickly necessitates real AGI style conceptualization. However, as software there should be some configurable layer of diagnostics, heuristics, optimizations, and strategies available that it itself could maintain. There's nothing magic to non-AI software doing that; making it visible to the AI running within the software seems obvious. If the AI is enabled by some running algorithms, it can tune those, as well as introspect the current state of goals and decisions implemented by those algorithms. In order to avoid explosion, those would be intentionally perceived, not a constantly active information stream.
(conversational feedback would have immediate application without lengthy retraining )
Would imply that the system immediately trusts. Would probably be rather self destructive, thus not intelligent.
The tool would trust me, specifically. Or more generally, the owner and permission holder to the system. If there are weird issues with whatever information, commands, or boundaries I give it that cannot be resolved by whatever capacity of "common sense" it already has formed (required for true natural language interaction), then that would be a point of discussion and clarification; or of course the tool owner could escalate permission and hope the system doesn't screw up by acting without full understanding.
The other thought behind that statement is to contrast statistical learning methods with rational learning methods. Adding a rational statement to a theorem prover, for instance, takes (logically) immediate and absolute effect. Telling a car-recognizing neural net that a particular type of vehicle is now no longer classified for its purposes requires lengthy retraining, because it has no means of being "aware" of that distinction at all.
Right, that's why dealing with such things requires intelligence, and if it could do such a thing would generally be considered intelligent. It's also why symbolic AI has failed to produce any general intelligence, because it simply cannot scale. In order for a system to exhibit such behavior, it needs to adaptively and "intelligently" prioritize what it's doing and on what it's working, as well as to predictively index and preprocess information, in order to even begin to achieve any sense of tractability.
Why do we pour money and resources in building AI when we have so many people with under-utilized brains already?
Expected return.
The snopes article still says that the complete change from sugar to HFCS was just months before the switch to New Coke. The two were not related, but the switch in sweetener was done at roughly the same time as the New Coke fiasco. Sugar-based "Old Coke" was no more right before New Coke came along.
We also don't know how many licenses Microsoft needs to break even for its Windows division. I'd suspect the vast majority of these are cheap OEM licenses, not corporate sales, as businesses are steering clear of Win8, so we're probably talking a couple dozen dollars of revenue per sale on average.
I'd be interested to know how much a couple billion dollars is compared to the Win8-related expenses Microsoft has had in pre-launch and continuing development, marketing, and support of the OS.
The improvement has been using multiple ANNs together as communicating units that can both communicate information and dynamically train each other, instead of trying to make a single large ANN and using external training sets. Of course, this isn't that new, as these guys have been working on such models since at least the early 90s.
We don't know what intelligence actually is, but we have an example of something that is unarguably intelligent: the mammalian brain.
To further this point, a brain is also not necessarily just a pile of neurons. There is specialization in the brain, and while other portions of the brain can make up for damaged parts, that only goes so far.
Any proposed mechanism of intelligence should be discounted unless it behaves the same way as a brain. Most AI research fails this test.
I would offer some more subjectivity to that. Mammalian brains tire, need sleep, do not have perfect recall, run things out of time order and convinces itself otherwise, take a long time to train, and has strong emotional needs.
With a view of "intelligence" that would include tools and helpers, and not just autonomous life forms, we do not want certain intelligence to behave the same way as the brain. We also want to extract what intelligence and certain components like rationality, creativity, intuitiveness, judgement, learning, and conceptualization are and be able to use those outside of the limitations that biology brings.
I'd mod you up if I could, but I think I can help out with a few points instead:
1) There is no concrete constructive definition of intelligence yet, and I think anybody at a higher level in the field knows that. Establishing that definition is a recognized part of AI research. Intelligence is still recognized comparatively, usually related to something like the capability to resolve difficult or ambiguous problems with similar or greater effect than humans, or can learn and react to dynamic environmental situations to similar effect as other living things. Once we've created something that works and that we can tangibly study, we can begin to come up with real workable definitions of intelligence that represent both the technological and biological instances of recognized intelligence.
4) Modern ANN research sometimes includes altering the morphology of the network as part of training, not just altering the coefficients. I would hope something like that is in effect here.
Biological neurons are far more complex than ANN neurons. At this point it's unknown if we can make up for that lack of dynamic state by using a larger ANN, or by increasing the per-neuron complexity to try to match the biological counterpart. I do have my doubts about the former, but that doubt is merely intuition, not science. We simply don't know yet.
as is whether this is what folks have in mind by "AI."
In my own studies, this isn't the branch of AI I'm particularly interested in. I don't care about artificial life, or structures based around the limitations of the biological brain. I'd love to have a system with perfect recall, could converse about its knowledge and thought processes such that conversational feedback would have immediate application without lengthy retraining, and could tirelessly and meticulously follow instructions given in natural language.
I don't see modeling biological brains as being a workable approach to that view of AI, except maybe the "tirelessly" part. I'm more interested in cognitive meta-models of intelligence itself than the substrate on which currently known knowledge happens to reside.
The full Episode 4 using Disney iconic characters with the witch from Sleeping Beauty as Darth Vader
Her name is Maleficent, which is a pretty cool villain name to know.
Most people read information on computer displays, reading web pages, emails, facebook updates, twitter feeds, wikipedia, and reference materials; and work in word processors, spreadsheets, and programming environments. All of these features are regularly constrained by vertical resolution.
For people watching cat videos and playing simple games (which comprises almost everybody else not doing the above), neither >1080p resolution nor fidelity matters.
For people doing high-end gaming and watching high-end media, your situation applies. However, it's a pretty tiny sliver of overall computer monitor time, all things considered.
And if they called it 4xHD, I'd be fine with it.