No, I'm arguing that you need a certain level of complexity and that neurons are likely very close to the minimum level necessary as that is what we've observed in data densities in DNA and information capacity of proteins.
Not necessarily, vector processors existed in the 1980s and a GPU is just a very cut-down vector processor. It depends on what calculation you're doing.
I'm assuming that, in order to guarantee being able to create virtual synapses via SDN and be able to give hard real-time networking guarantees, you have to have fixed path lengths. That rules out trees. Your choice is a butterfly network or a hypercube topology.
I'm assuming that Amdahl's Law applies, that communication and task switching are your chief limits for parallelism.
And, finally, since we can't build smaller transistors but have to switch to quantum computing, I'm assuming that the scaling of QC systems will be bounded. Since each synapse is state-bound where the state is implied by other states, only certain combinations are valid. Since the synapses can move from one neuron to another (by shifting weights in a NN sim, by actually moving in a biological simulator), all the benefits of QC suddenly stop applying.
He does not have a vision for space and wants most space activities shut down. What remaining earth observatory work is still going on will be the first to go. There will be no further missions to Mars and space telescopes will be scrapped.
Well, one small problem. Those aren't generally the people who have STDs, so if you're ignorant then you'll allocate completely the wrong budget to the problem which IS an administrative issue.
Also, admins with a 'tude are generally what's known in the trade as "losers". As are ACs with a 'tude, but I doubt the AC in question can read. They certainly can't write worth a damn.
Based on Moore's Law, we won't have human-level AI until the year 3200. (We know the speed of simulating 100,000 neurons in 2014 and we know the speed of simulating 100 neurons in 1985, therefore we can fit the appropriate family of curves.)
Universal Income would work if the total cost of UI minus the increase in revenue generated through more people working minus the administrative cost of the systems you'd no longer need minus the cost of current benefits to be replaced is negative (ie: you're saving more than you're spending).
This is possible, and there's every reason to think it would work that way, but I have not seen any models based on the trials that have been done. Theory can only be based on fact and can only be tested by fact. Anything else belongs in the category of religion.
Chances are, I have far more valuable skills than the entire right-wing side of Slashdot.
And, you know what?
I don't give a damn about being paid. I have a choice, work for an idiot or work for me. I'm tired of idiots and right-wing losers. They can take their jobs and stuff 'em wherever the hell they want.
I can grow my own food and prepare it better than any fast food place.
Or I can program particle accelerator systems, devise aviation software or build high-rnd networking gear.
Done all those. Interesting, but not if you're going to shout. I don't need your money more than I need my hearing.
House? I'm negotiating to handle the upkeep of a 400 year old hall. The thinking is that if I can maintain it, I can stay in it. It doesn't fall apart and the owners don't have to pay for repairs.
No big. There's nothing I can't do there.
You lot get rid of me, I get rid of you lot, everyone's happy.
Part of the problem is the procurement procedure. Because it's minimum cost, regardless of whether it can be done for that or whether it would be any good, companies ALWAYS lowball.
The other part of the problem is economic mismanagement by the government. A huge multinational collaboration is going to rack up costs due to differing exchange rates, differing inflation rates, the overheads of tepid government investment, the price of melting down the entire global economy, the cost of replacing lost scientists and engineers due to downsizing from the occasional global meltdown, unexpected costs of doing business due to the occasional trade war, botched management by political appointees, dictating who you can buy from, stuff like that.
Giving a bunch of scientists and engineers a blank (but 100% refundable) check and telling them to come back with a working system or the money would give you better results for a lower price.
They keep lowering the definition of "broadband" and keep excusing companies for not supplying what customers pay for.
Meanwhile, in civilized countries, they're getting 90 mbps - 500 gbps rates in remote country villages halfway up the sides of mountains or in remote hills. For a fraction of the price. Which goes to show it's not about the money, it's about the profit the companies want to sponge off you.
If there's no encryption, the criminals will simply exploit other people's computers, just as they have in the past. This leaves no forensic traces on the criminal's computers. Want to talk darkness? Good place to start, the Involuntary Cloud.
If there is encryption, anything that would leave physical evidence still does so. Forensic labs are underfunded or privatised. In either case, incorrect results leading to false convictions and false rejection of suspects are commonplace. Getting those sorted would make defeating encryption less of a necessity and would solve far more crimes that have no relationship to encryption at all.
Let's focus on that second one. Crime labs are a mess. DNA analysis is regularly tainted with contamination and hundreds of thousands of serious crime scene swabs (rape kits, etc) are left abandoned, their evidence decaying. We all remember the false fingerprint match in the Spanish train bombing. These things are STUPID!
Fix this first. Then fix police attitudes. They are trained to be warlords, which is bad for community policing , bad for trust, bad for getting people in Starbucks arrested on suspicion of being black in a built-up area, and very very bad for SWATting victims.
Better community relations will solve a lot of problems. People know a lot but tell little to thugs in uniform. Give them police they can trust. Community police. Honest police.
Speaking of which, did you see the reports on the police unit that turned rogue and were using their uniforms to terrorize, burgle and extort? And the Chicago "black ops" jail where people were arrested without charge and tortured into confessions?
Are these reports true? Does it matter? As long as they're believed, people will respond with fear, hatred and disgust. 99.9% of your leads burned because you fancied the thumbscrews.
Fix the forensics and the police, and you'll find there simply isn't much you need to decrypt. But as long as you try to fight the darkness with the night, you will lose.
1. Are you running KDE, Gnome or Unity? If so, bury them with dignity at the bottom of the garden and install Enlightenment. 2. Are you running SystemD? Ditto and install OpenBSD's init or Gentoo's OpenRC. 3. Is the kernel tuned? It should not contain things you don't use. If power is a problem, that is the governor you should select. Run with the worst latency you can afford because you need more power to get equal performance with low latency. 4. You might want to use the hoard memory allocator over Glibc's malloc for some apps - you can set this using the dynamic library path. 5. You might want to use a lightweight libc to reduce system strain. 6. You might want to experiment with apps to see what causes netsnmp to scream in pain at the load levels. It's a good monitor. Once you've identified the problem apps, place them in a cgroup that is more restricted on resources.
Should you have to do this? No. Fact is, you do, because distro authors are going for mass appeal rather than usability. They always have, which is why MCC was the last distro I actually liked using.
Same as with any other kernel function - no verifiable use.
The problem is that this buggers up my plan to travel back in time and upload the latest Linux source into Linus Torvalds' head when he was a student. The architectures removed were the only ones around then. How's that going to convince him, it he can't run it?
For most purposes, it's not necessary to test it on hardware. We know the underlying functions work and that side-effects are minimal. The requirements are that the API presented matches the API the kernel uses, the semantics are as expected and that the code compiles. That requires a continuous integration tool and a way to describe the semantics.
You can get the software for free. The effort is a bit more, depending on how much it is worth.
It won't prove things work, but it will prove things will probably work and that they should work. If you're using obsolete hardware, you can expect to have to make mods, this would place bounds on that.
Ok, maybe from the mainstream kernel, but space missions use old hardware and you don't encourage recycling eWaste this way.
Maybe have a tree of obsolete architectures, where those who want old stuff can go.
Better yet, since architectures are split off anyway, have a subtree for each architecture (inline or otherwise). Let arch-specific distros merge the architectures they want and no others, let developers cleanly see what does what. It should make adding new architectures to (and removing old ones from) the mainline easier if the mainline had the arch section assembled.
Regardless of whether anyone thinks either is sensible or even sane, I might have a go at building a mega-arch (triumphal arch?) project just for kicks. Well, the DEC VAX arch needs to be less obscure, assuming there's still a copy out there.
Not change? On average, for virtually all of the last 250 million years.
The last century has seen more change than in any given ten million year period. I assume you've done calculus. Ok, maybe not.
Nobody is spending trillions on climate change. If they had done so, it wouldn't be a problem. It's because they're too busy spending it on known things like warships, cruise missiles, nuclear warheads and pay raises for bankers and the uber-wealthy that there's a problem.
No, I have absolutely bugger all sympathy for your argument.
Give universities a fixed fund for the year, then let them divide that between departments, then let departments divide that between researchers, then let researchers divide that between all the projects they want to do that year.
The way it traditionally was and still is in some countries.
This has the advantage that funding is then not a popularity contest, the market can't block research that is inimical to any given market, and politics is largely eliminated from science.
I have no problems with mass distribution of (some) information, be it by blockchain or other system. A federated database would be valuable in a lot of areas, but you need to eliminate the over-centralizing of power over the deep research, and deep research still needs deep pockets. Splitting those two apart is a serious challenge, one citizen science can't yet solve.
I've no problems with people trying things out, as long as there is a good flow of information back from patients to doctors and researchers when things either succeed or fail, where this includes genetic information and (in the case of failures) biopsies that can be tested using AMS to find out what was going on, and for patients where it's too early to really tell there's monitoring.
You won't be able to catch everything in time, and as I've noted elsewhere, there's a balance between cost and effectiveness. My opinion is founded on the idea that we don't know what the ideal balance is and therefore you want this to be dynamic and favouring slightly too much information. As time goes on, let a self-adjusting system find out what information is actually useful, when.
I have no illusions about modern medicine, but I do hold that if you give researchers more information, they're more likely to be able to do more with it and more likely to figure out what could be done if only they had X, Y or Z as well. Let the researchers do the fine-tuning, but give them enough information to figure out what that even means.
Slashdot Mirror
No, I'm arguing that you need a certain level of complexity and that neurons are likely very close to the minimum level necessary as that is what we've observed in data densities in DNA and information capacity of proteins.
Not necessarily, vector processors existed in the 1980s and a GPU is just a very cut-down vector processor. It depends on what calculation you're doing.
I'm assuming that, in order to guarantee being able to create virtual synapses via SDN and be able to give hard real-time networking guarantees, you have to have fixed path lengths. That rules out trees. Your choice is a butterfly network or a hypercube topology.
I'm assuming that Amdahl's Law applies, that communication and task switching are your chief limits for parallelism.
And, finally, since we can't build smaller transistors but have to switch to quantum computing, I'm assuming that the scaling of QC systems will be bounded. Since each synapse is state-bound where the state is implied by other states, only certain combinations are valid. Since the synapses can move from one neuron to another (by shifting weights in a NN sim, by actually moving in a biological simulator), all the benefits of QC suddenly stop applying.
Because human intelligence is emergent, you need a necessary and sufficient system for intelligence to emerge.
No, they don't. The NIH funds a lot of the work via your taxes and the rest is made up by hiking the prices you cough up. YOU pay. Not them.
He does not have a vision for space and wants most space activities shut down. What remaining earth observatory work is still going on will be the first to go. There will be no further missions to Mars and space telescopes will be scrapped.
The facts on the ground are, well, facts. Not opinions.
Thank you. Good day.
Rockets don't generally give a crap about politics. They work or they explode.
The people who make them not explode are almost invariably left-wingers because that's what you generally become when you're in maths and science.
So the election? Not worth a damn. Not to the rockets, not to Mars, not to the GPS systems. They don't give a shit.
You appoint people to get the job done, not to please some cattle rancher who thinks the world is flat and aliens live in Area 51.
Well, one small problem. Those aren't generally the people who have STDs, so if you're ignorant then you'll allocate completely the wrong budget to the problem which IS an administrative issue.
Also, admins with a 'tude are generally what's known in the trade as "losers". As are ACs with a 'tude, but I doubt the AC in question can read. They certainly can't write worth a damn.
If that is all there is to know about someone, because that's all there is in their brain, then you know 100%.
Based on Moore's Law, we won't have human-level AI until the year 3200. (We know the speed of simulating 100,000 neurons in 2014 and we know the speed of simulating 100 neurons in 1985, therefore we can fit the appropriate family of curves.)
Universal Income would work if the total cost of UI minus the increase in revenue generated through more people working minus the administrative cost of the systems you'd no longer need minus the cost of current benefits to be replaced is negative (ie: you're saving more than you're spending).
This is possible, and there's every reason to think it would work that way, but I have not seen any models based on the trials that have been done. Theory can only be based on fact and can only be tested by fact. Anything else belongs in the category of religion.
I don't believe in the existence of individuals, since the mind is emergent and therefore transitory.
I do believe in rates of return, fighters offer zero and telescopes offer - well, not much but still quite a bit.
Coastlines are fractal and have 1.4 dimensions. This does mean Cthulhu could break out at any moment.
Chances are, I have far more valuable skills than the entire right-wing side of Slashdot.
And, you know what?
I don't give a damn about being paid. I have a choice, work for an idiot or work for me. I'm tired of idiots and right-wing losers. They can take their jobs and stuff 'em wherever the hell they want.
I can grow my own food and prepare it better than any fast food place.
Or I can program particle accelerator systems, devise aviation software or build high-rnd networking gear.
Done all those. Interesting, but not if you're going to shout. I don't need your money more than I need my hearing.
House? I'm negotiating to handle the upkeep of a 400 year old hall. The thinking is that if I can maintain it, I can stay in it. It doesn't fall apart and the owners don't have to pay for repairs.
No big. There's nothing I can't do there.
You lot get rid of me, I get rid of you lot, everyone's happy.
I'd still take the telescope.
Part of the problem is the procurement procedure. Because it's minimum cost, regardless of whether it can be done for that or whether it would be any good, companies ALWAYS lowball.
The other part of the problem is economic mismanagement by the government. A huge multinational collaboration is going to rack up costs due to differing exchange rates, differing inflation rates, the overheads of tepid government investment, the price of melting down the entire global economy, the cost of replacing lost scientists and engineers due to downsizing from the occasional global meltdown, unexpected costs of doing business due to the occasional trade war, botched management by political appointees, dictating who you can buy from, stuff like that.
Giving a bunch of scientists and engineers a blank (but 100% refundable) check and telling them to come back with a working system or the money would give you better results for a lower price.
Congress is planning on killing the James Webb telescope as it can't now meet the absolute maximum budget authorized.
They keep lowering the definition of "broadband" and keep excusing companies for not supplying what customers pay for.
Meanwhile, in civilized countries, they're getting 90 mbps - 500 gbps rates in remote country villages halfway up the sides of mountains or in remote hills. For a fraction of the price. Which goes to show it's not about the money, it's about the profit the companies want to sponge off you.
If there's no encryption, the criminals will simply exploit other people's computers, just as they have in the past. This leaves no forensic traces on the criminal's computers. Want to talk darkness? Good place to start, the Involuntary Cloud.
If there is encryption, anything that would leave physical evidence still does so. Forensic labs are underfunded or privatised. In either case, incorrect results leading to false convictions and false rejection of suspects are commonplace. Getting those sorted would make defeating encryption less of a necessity and would solve far more crimes that have no relationship to encryption at all.
Let's focus on that second one. Crime labs are a mess. DNA analysis is regularly tainted with contamination and hundreds of thousands of serious crime scene swabs (rape kits, etc) are left abandoned, their evidence decaying. We all remember the false fingerprint match in the Spanish train bombing. These things are STUPID!
Fix this first. Then fix police attitudes. They are trained to be warlords, which is bad for community policing , bad for trust, bad for getting people in Starbucks arrested on suspicion of being black in a built-up area, and very very bad for SWATting victims.
Better community relations will solve a lot of problems. People know a lot but tell little to thugs in uniform. Give them police they can trust. Community police. Honest police.
Speaking of which, did you see the reports on the police unit that turned rogue and were using their uniforms to terrorize, burgle and extort? And the Chicago "black ops" jail where people were arrested without charge and tortured into confessions?
Are these reports true? Does it matter? As long as they're believed, people will respond with fear, hatred and disgust. 99.9% of your leads burned because you fancied the thumbscrews.
Fix the forensics and the police, and you'll find there simply isn't much you need to decrypt. But as long as you try to fight the darkness with the night, you will lose.
Checklist:
1. Are you running KDE, Gnome or Unity? If so, bury them with dignity at the bottom of the garden and install Enlightenment.
2. Are you running SystemD? Ditto and install OpenBSD's init or Gentoo's OpenRC.
3. Is the kernel tuned? It should not contain things you don't use. If power is a problem, that is the governor you should select. Run with the worst latency you can afford because you need more power to get equal performance with low latency.
4. You might want to use the hoard memory allocator over Glibc's malloc for some apps - you can set this using the dynamic library path.
5. You might want to use a lightweight libc to reduce system strain.
6. You might want to experiment with apps to see what causes netsnmp to scream in pain at the load levels. It's a good monitor. Once you've identified the problem apps, place them in a cgroup that is more restricted on resources.
Should you have to do this? No. Fact is, you do, because distro authors are going for mass appeal rather than usability. They always have, which is why MCC was the last distro I actually liked using.
Same as with any other kernel function - no verifiable use.
The problem is that this buggers up my plan to travel back in time and upload the latest Linux source into Linus Torvalds' head when he was a student. The architectures removed were the only ones around then. How's that going to convince him, it he can't run it?
For most purposes, it's not necessary to test it on hardware. We know the underlying functions work and that side-effects are minimal. The requirements are that the API presented matches the API the kernel uses, the semantics are as expected and that the code compiles. That requires a continuous integration tool and a way to describe the semantics.
You can get the software for free. The effort is a bit more, depending on how much it is worth.
It won't prove things work, but it will prove things will probably work and that they should work. If you're using obsolete hardware, you can expect to have to make mods, this would place bounds on that.
Ok, maybe from the mainstream kernel, but space missions use old hardware and you don't encourage recycling eWaste this way.
Maybe have a tree of obsolete architectures, where those who want old stuff can go.
Better yet, since architectures are split off anyway, have a subtree for each architecture (inline or otherwise). Let arch-specific distros merge the architectures they want and no others, let developers cleanly see what does what. It should make adding new architectures to (and removing old ones from) the mainline easier if the mainline had the arch section assembled.
Regardless of whether anyone thinks either is sensible or even sane, I might have a go at building a mega-arch (triumphal arch?) project just for kicks. Well, the DEC VAX arch needs to be less obscure, assuming there's still a copy out there.
Not change? On average, for virtually all of the last 250 million years.
The last century has seen more change than in any given ten million year period. I assume you've done calculus. Ok, maybe not.
Nobody is spending trillions on climate change. If they had done so, it wouldn't be a problem. It's because they're too busy spending it on known things like warships, cruise missiles, nuclear warheads and pay raises for bankers and the uber-wealthy that there's a problem.
No, I have absolutely bugger all sympathy for your argument.
Oh, and Get Off My Lawn!
No, the right to do and the right to be free are the same thing.
Why not take money out of the university system?
Give universities a fixed fund for the year, then let them divide that between departments, then let departments divide that between researchers, then let researchers divide that between all the projects they want to do that year.
The way it traditionally was and still is in some countries.
This has the advantage that funding is then not a popularity contest, the market can't block research that is inimical to any given market, and politics is largely eliminated from science.
I have no problems with mass distribution of (some) information, be it by blockchain or other system. A federated database would be valuable in a lot of areas, but you need to eliminate the over-centralizing of power over the deep research, and deep research still needs deep pockets. Splitting those two apart is a serious challenge, one citizen science can't yet solve.
I've no problems with people trying things out, as long as there is a good flow of information back from patients to doctors and researchers when things either succeed or fail, where this includes genetic information and (in the case of failures) biopsies that can be tested using AMS to find out what was going on, and for patients where it's too early to really tell there's monitoring.
You won't be able to catch everything in time, and as I've noted elsewhere, there's a balance between cost and effectiveness. My opinion is founded on the idea that we don't know what the ideal balance is and therefore you want this to be dynamic and favouring slightly too much information. As time goes on, let a self-adjusting system find out what information is actually useful, when.
I have no illusions about modern medicine, but I do hold that if you give researchers more information, they're more likely to be able to do more with it and more likely to figure out what could be done if only they had X, Y or Z as well. Let the researchers do the fine-tuning, but give them enough information to figure out what that even means.