Publicly traded companies have an Embedded Growth Obligation (EGO) due to the expectations of shareholders and the market. Nothing can grow forever, but the market seems to think that sustaining certain customer level for decades is equivalent to death. This will change eventually, due to the laws of physics, but it is likely to be a rough ride.
I believe (as I said back in 2014) that Tesla's plan was to modulate the conductivity of the ionosphere, effectively turning it into a MASER, and thus capturing a great deal of the energy imparted by the solar wind and making it available for use.
At the time, it would have seemed unlimited, but long ago I did the math, and if I recall correctly, it would be about 1 Terawatt of power, which is about 8% of our current worldwide power demand.
So, yes... I think it would have worked, but we would have outgrown it quickly enough.
Computers are pretty damned fast and cheap compared to the days of the 1 MIPS $3000 IBM PC with 3 360k floppy disks and a monochrome monitor. The rise of Arduino and Raspberry Pi make it possible to stick computing in almost anything for less than $20, and in some cases under a buck!
Bandwidth is far MORE expensive than I predicted... I expected full duplex gigabit for $50/month by now... it amazes me that cable TV is still a dominant way of delivering data to the masses. I predicted that you'd be able to have a full duplex video feed (Facetime anyone) between any 2 points in the world for $50/month.... we never made it.
Operating Systems are now far LESS reliable and secure than the days of MS-DOS. You could always write protect your OS disk, and easily make copies of it. You could trust copies to work years later, and everyone understood how to make them. You didn't have to worry about your hardware getting bricked.
Video and Cameras are amazing, I had no idea how cool things could get.
Wireless / Cellular networks are way better than I expected, but again the monopoly pricing structures are weird.
There are lots of cool surprises, Wikis, Blogging, Video Sharing, Podcasts, Ebay, Amazon, 3d printers and milling machines for cheap. Open source software and hardware,
Assume for the moment that our memories are stored in superconducting qubits, and the critical temperature really is 2000K, what is the critical field required to erase someone's memories completely at body temperature(310k)? Surely this is a simple matter for someone good at physics to figure out.
There's a lot more to do, it's not just about scaling down transistors. More layers of transistors, different ways of structuring logic to better use the transistors... there are a lot of ways left unexplored to keep going, even if transistor size remains fixed.
I don't have a smartphone now.... I've got a candy bar phone with no camera, and no internet. I text about once per week, and make about 5 calls per day, mostly related to coordinating with a friend I give rides home to. I get that people find those things useful, the only really compelling need for me is I wish I had google maps at times... but yes, you can live without it.
Imagine if you had a array of these clocks spread out in a 2d grid, level with each other to within less than a millimeter, It is possible some would run faster or slower than the others because of subterranean variations in density, and thus slightly different amounts of gravity.
It struck me as very odd to see how supportive the CEO of Google was of this walkout.... most of the left/right world just sees it as caving in to snowflake pressure, or the workers bringing about positive change through collective action.... but I have a different theory.
Normally, the hands of management are bound by lots of rules, shareholder pressure, the SEC, etc... I'm sure the CEO was aware of the issues, but too bound up by the rules and social pressures from above (shareholders, the 0.001%, etc) to effectively deal with it.
if the workers happen to "organize" a strike demanding something that the CEO would like to do, but can't.... you get the aforementioned weird reaction. Moral dilemma on the part of the CEO is solved, workers are happy that they have some power, and shareholder blame gets deflected safely away from management.
I expect this to happen more, as it might be a new corporate cultural norm.
This is a false choice. Any code added to address this non-issue is going to make the overall system LESS safe for everyone, as it will add complexity, and the likelihood of failure.
In the job shop / small manufacturing world I now inhabit, it takes about 1/2 hour or so to get everything going in the morning, and about the same time to shut it all down at the end of the day. So, we'd get about 70% of our current productivity if we took this approach. There are many other types of work, as stated above (ER, Medical care, Service industry), where you'd have hire 33% more workers to get coverage. Where's all that money going to come from to pay all and train all these new hires?
Some old white dude (like me) probably wrote this in a comfy office.
CGP Grey came up with idea that initially sounds silly, until you think about it...
Get rid of publicly seen view counts, subscriber counts and thumbs up/down counts for all videos and channels. The creators would still see them privately. It would take a lot of the pressure off.
Stop using systems that are based on ambient authority to get work done, and this problem goes away... we figured this out in the 1970s, but seem to have forgotten that lesson.
It is possible to build systems that are much, much harder to hack, thus making herd immunity high enough that worms and virii won't spread, no matter how clever.
10 more years until people wake up enough to take action.
On a small fire that has just started, this might work.... but once you've raised the temperature of everything to the point where oxygen supply is the rate limiting factor, I doubt this would work for more than a few minutes.
The fact is that thanks to Ambient Authority, nothing is safe, and can't be made safe. Anyone who works in infosec and thinks otherwise is nuts. The shitstorm is going to come, just hope it doesn't happen on your watch, or that you can deflect the blame enough to survive.
This is India where completely bogus rumors on social media have resulted in angry hordes of people burning or stoning innocents to death. I could easily imagine stores selling these products being looted and burned or the people working in their factories being killed.
Pikers - This is America where bogus rumors spread by social media resulted in the election a 70 year old reality TV star as our President. The death and destruction this will cause could potentially be of Biblical proportions.
There's a huge collective blind spot in the programming community as a whole... they somehow believe that Ambient Authority is an acceptable basis for writing applications in a world of persistent networking and mobile code.
This is really an argument about externalities, costs shoved off to society, instead of being paid for up front. There are costs to HTTPS, and a great deal of technical debt would be incurred in forcing older sites to deploy it. HTTPS is a set of trade offs, one of which involves centralizing trust (and thus the ability to censor) in the top level certification sites. Using HTTPS also prohibits the development of other options, any of which may actually be far superior, in other words, premature optimization.
There's no really good reason to force old web sites to change everything for your latest version of security kool-aid, and again in 6 months, and again in 6 months, ad hoc, ad nauseum. It won't actually do much good, and as stated above, does much harm by potentially removing history.
Grow up, kids.... HTTPS is like beta software... it's not done yet. Get back to me in when it hasn't undergone a revision in at least 5 years.
Ambient Authority is a design decision which only appears once you have multiple users sharing a computer. As a result, everyone just kept using it without much thought... until we find ourselves in a world of persistent networks, mobile code, no system administrators, and multiple layers of firmware and OS from various hardware and software vendors.
In such a system, any code runs with the full authority of the user who started the task, and the users have no effective means of limiting the side effects of running a given program. This in turn means we have to try to guess the intent of code (which is equivalent to solving the halting problem, and is thus impossible). The band-aid is to then try to enumerate all the bad code in the world (virus scanners), and to enumerate all the code bugs in all our programs (security updates), and to eliminate the trust of users (DRM, forced updates, "safety" filters in our browsers). None of these band-aids will work against a determined individual, let alone a nation-state.
Running tasks with the least possible privilege, the "Principle of Least Authority" (POLA) allows a user in such a system to decide ahead of time exactly what files the program is allowed to read, write, etc. Because we're all used to dialog boxes, and drag to drop GUI elements, this doesn't even require any special training of users to accomplish.
Of course, rebuilding our infrastructure to fix a design flaw of the size and scope of using 2 digit years (the Y2K problem we once faced), isn't going to be easy... especially when there's no deadline to make the need for action obvious. It's just going to remain an insidious vulnerability instead for decades to come.
If you think EAL certifications address this, they don't. 8(
No, AI can't be made to follow vague rules, You can't make rules explicit enough to be computed. This is like the conversation a while ago trying to apply "the trolley problem" to self driving cars... any solution just makes the code less reliable and thus more likely to kill people.
Slashdot Mirror
Publicly traded companies have an Embedded Growth Obligation (EGO) due to the expectations of shareholders and the market. Nothing can grow forever, but the market seems to think that sustaining certain customer level for decades is equivalent to death. This will change eventually, due to the laws of physics, but it is likely to be a rough ride.
I doubt you could extract it all at one place, you'd have to spread the load out a bit.
I believe (as I said back in 2014) that Tesla's plan was to modulate the conductivity of the ionosphere, effectively turning it into a MASER, and thus capturing a great deal of the energy imparted by the solar wind and making it available for use.
At the time, it would have seemed unlimited, but long ago I did the math, and if I recall correctly, it would be about 1 Terawatt of power, which is about 8% of our current worldwide power demand.
So, yes... I think it would have worked, but we would have outgrown it quickly enough.
Computers are pretty damned fast and cheap compared to the days of the 1 MIPS $3000 IBM PC with 3 360k floppy disks and a monochrome monitor. The rise of Arduino and Raspberry Pi make it possible to stick computing in almost anything for less than $20, and in some cases under a buck!
Bandwidth is far MORE expensive than I predicted... I expected full duplex gigabit for $50/month by now... it amazes me that cable TV is still a dominant way of delivering data to the masses. I predicted that you'd be able to have a full duplex video feed (Facetime anyone) between any 2 points in the world for $50/month.... we never made it.
Operating Systems are now far LESS reliable and secure than the days of MS-DOS. You could always write protect your OS disk, and easily make copies of it. You could trust copies to work years later, and everyone understood how to make them. You didn't have to worry about your hardware getting bricked.
Video and Cameras are amazing, I had no idea how cool things could get.
Wireless / Cellular networks are way better than I expected, but again the monopoly pricing structures are weird.
There are lots of cool surprises, Wikis, Blogging, Video Sharing, Podcasts, Ebay, Amazon, 3d printers and milling machines for cheap. Open source software and hardware,
Assume for the moment that our memories are stored in superconducting qubits, and the critical temperature really is 2000K, what is the critical field required to erase someone's memories completely at body temperature(310k)? Surely this is a simple matter for someone good at physics to figure out.
There's a lot more to do, it's not just about scaling down transistors. More layers of transistors, different ways of structuring logic to better use the transistors... there are a lot of ways left unexplored to keep going, even if transistor size remains fixed.
I don't have a smartphone now.... I've got a candy bar phone with no camera, and no internet. I text about once per week, and make about 5 calls per day, mostly related to coordinating with a friend I give rides home to. I get that people find those things useful, the only really compelling need for me is I wish I had google maps at times... but yes, you can live without it.
Imagine if you had a array of these clocks spread out in a 2d grid, level with each other to within less than a millimeter, It is possible some would run faster or slower than the others because of subterranean variations in density, and thus slightly different amounts of gravity.
Interesting stuff, thanks for sharing.
It struck me as very odd to see how supportive the CEO of Google was of this walkout.... most of the left/right world just sees it as caving in to snowflake pressure, or the workers bringing about positive change through collective action.... but I have a different theory.
Normally, the hands of management are bound by lots of rules, shareholder pressure, the SEC, etc... I'm sure the CEO was aware of the issues, but too bound up by the rules and social pressures from above (shareholders, the 0.001%, etc) to effectively deal with it.
if the workers happen to "organize" a strike demanding something that the CEO would like to do, but can't.... you get the aforementioned weird reaction. Moral dilemma on the part of the CEO is solved, workers are happy that they have some power, and shareholder blame gets deflected safely away from management.
I expect this to happen more, as it might be a new corporate cultural norm.
This is a false choice. Any code added to address this non-issue is going to make the overall system LESS safe for everyone, as it will add complexity, and the likelihood of failure.
In the job shop / small manufacturing world I now inhabit, it takes about 1/2 hour or so to get everything going in the morning, and about the same time to shut it all down at the end of the day. So, we'd get about 70% of our current productivity if we took this approach. There are many other types of work, as stated above (ER, Medical care, Service industry), where you'd have hire 33% more workers to get coverage. Where's all that money going to come from to pay all and train all these new hires?
Some old white dude (like me) probably wrote this in a comfy office.
CGP Grey came up with idea that initially sounds silly, until you think about it...
Get rid of publicly seen view counts, subscriber counts and thumbs up/down counts for all videos and channels. The creators would still see them privately. It would take a lot of the pressure off.
Stop using systems that are based on ambient authority to get work done, and this problem goes away... we figured this out in the 1970s, but seem to have forgotten that lesson.
It is possible to build systems that are much, much harder to hack, thus making herd immunity high enough that worms and virii won't spread, no matter how clever.
10 more years until people wake up enough to take action.
On a small fire that has just started, this might work.... but once you've raised the temperature of everything to the point where oxygen supply is the rate limiting factor, I doubt this would work for more than a few minutes.
I'm not a fireman, but this sounds implausible.
The fact is that thanks to Ambient Authority, nothing is safe, and can't be made safe. Anyone who works in infosec and thinks otherwise is nuts. The shitstorm is going to come, just hope it doesn't happen on your watch, or that you can deflect the blame enough to survive.
This is India where completely bogus rumors on social media have resulted in angry hordes of people burning or stoning innocents to death. I could easily imagine stores selling these products being looted and burned or the people working in their factories being killed.
Pikers - This is America where bogus rumors spread by social media resulted in the election a 70 year old reality TV star as our President. The death and destruction this will cause could potentially be of Biblical proportions.
I assumed it (Google Reader, and thus RSS) was killed deliberately by Larry Page.
There's a huge collective blind spot in the programming community as a whole... they somehow believe that Ambient Authority is an acceptable basis for writing applications in a world of persistent networking and mobile code.
So, now we need to update the old Russian saying
""there's no truth in Pravda, and no You in YouTube"
This is really an argument about externalities, costs shoved off to society, instead of being paid for up front. There are costs to HTTPS, and a great deal of technical debt would be incurred in forcing older sites to deploy it. HTTPS is a set of trade offs, one of which involves centralizing trust (and thus the ability to censor) in the top level certification sites. Using HTTPS also prohibits the development of other options, any of which may actually be far superior, in other words, premature optimization.
There's no really good reason to force old web sites to change everything for your latest version of security kool-aid, and again in 6 months, and again in 6 months, ad hoc, ad nauseum. It won't actually do much good, and as stated above, does much harm by potentially removing history.
Grow up, kids.... HTTPS is like beta software... it's not done yet. Get back to me in when it hasn't undergone a revision in at least 5 years.
You've obviously not heard of the GreenArrays GA144 chip. Insane amounts of compute power in a small passively cooled package.
Ambient Authority is a design decision which only appears once you have multiple users sharing a computer. As a result, everyone just kept using it without much thought... until we find ourselves in a world of persistent networks, mobile code, no system administrators, and multiple layers of firmware and OS from various hardware and software vendors.
In such a system, any code runs with the full authority of the user who started the task, and the users have no effective means of limiting the side effects of running a given program. This in turn means we have to try to guess the intent of code (which is equivalent to solving the halting problem, and is thus impossible). The band-aid is to then try to enumerate all the bad code in the world (virus scanners), and to enumerate all the code bugs in all our programs (security updates), and to eliminate the trust of users (DRM, forced updates, "safety" filters in our browsers). None of these band-aids will work against a determined individual, let alone a nation-state.
Running tasks with the least possible privilege, the "Principle of Least Authority" (POLA) allows a user in such a system to decide ahead of time exactly what files the program is allowed to read, write, etc. Because we're all used to dialog boxes, and drag to drop GUI elements, this doesn't even require any special training of users to accomplish.
Of course, rebuilding our infrastructure to fix a design flaw of the size and scope of using 2 digit years (the Y2K problem we once faced), isn't going to be easy... especially when there's no deadline to make the need for action obvious. It's just going to remain an insidious vulnerability instead for decades to come.
If you think EAL certifications address this, they don't. 8(
Ambient Authority, the design flaw underlying most "modern" operating systems, strikes again.
No, AI can't be made to follow vague rules, You can't make rules explicit enough to be computed. This is like the conversation a while ago trying to apply "the trolley problem" to self driving cars... any solution just makes the code less reliable and thus more likely to kill people.
Stop asking the question, please. ;-)
I was always in awe of the POLYD instruction - "Evaluate Polynomial, Double Precision"