You can call bullshit all you want, but if you look at the work from Cal (lab of Jack Gallant) and the dozens of others who have replicated his work, you'd find you're wrong.
If you go to her company's website (opnwatr.io) you can see that its approach is based on high precision infrared imaging. If you look at what is going on in infrared imaging (e.g. via conferences, etc.) you will see that this is very much the focus (pun intended) of a lot of current research. AND that BOLD (blood oxygen level detection), which is what is measured in fMRI, can be measured with similar or greater precision via near-infrared techniques. And on and on...
1. Neural lace creates digital I/O to the brain... 2. Neural lace kills neurons... 3. No problemo! Off board memory and proessors to the rescue - we're all backed up to and supported on some cloud service: AWS on steroids, almost literally. 4. Shared cloud service! We're all collaborative because we're all ants under the same control.
I'm not actually in favor of a/ injecting into my (admittedly defective in some ways) brain or b/ having my skull cut open, or c/ having probes jabbed into my grey-ish matter or... well, actually, any of this.
"How is this even legal?" Because the Internet Archive received a waiver from the DMCA for the purposes of archiving software that is no longer commercially available.
Taiwan is China? Same way that Japan is China and Vietnam is China and Korea (all of it) is China. Ethnically, Han; culturally not.
Taiwan was administered as a Chinese territory for only seven years of history - and spent rather longer (over 50 years) under Japanese rule. The only land you can see from any part of Taiwan (the island) is... Japanese. The dominant ethnic groups are Han, the dominant languages are Chinese, but the dominant culture is Japanese. The island's dominant population until early in the 20th century was of the aboriginal groups (tribes?) which are not Chinese at all, but Austronesian. Taiwan has a vigorous and messy democracy, parties vie energetically for power. Rule of law, mainly, pertains, with corruption at a level comparable to the US.
And Taiwanese of many ages now think of China as having both the past - an historic link to culture - and a present and future - China the nation as a terrifying borg-like entity, not at all as being "we're China, we just disagree on where the capital". They think the capital of China is Beijing, and the capital of what should be the independent nation-state of Taiwan is Taipei.
We're all grownups, many of us are nerds, technically literate and so are completely used to the idea of using reasonably long words for the precision they offer.
Can we please ban "pee" and "poo"? Always and forever.
I should have made more clear. I was talking about value to the customer / user. You're talking about costs. The incremental costs may become small (but they're not yet, see - for example - Gettys work on bufferbloat). Meanwhile, lots of folks buy Apple computers for 2x the price of a more powerful Windows machine because, well, perceived value. Apple has high margins. Perhaps, in your worldview, we should regulate Apple's prices and margins?
A rather silly over-simplification. Of course, voice is carried as data. However, it requires more than low latency - it requires that the latency be sustained as low. And it requires low error rates.
The reasons are buried deep in human behaviors. Delays easily realizable in IP networks with error correction are perceptible to the listener. Then, however, they're not ignored (as they are in a video stream being re-aggregated for playing) but are heard by the listener as hesitation. The Q&A: "do you want to go out for dinner on Friday?" A: "yes"... becomes "do you want to go out for dinner on Friday?" A: (slight pause) "yes". In human interaction, that silent pause is extra information.
(Of course, the degradation of voice quality on mobile networks means that the Q&A leads to answers like "huh? what did you say?")
There's a BIG difference between saying "voice is data" and the fact "voice is carried as data".
Some of this loss, doubtless, *is* caused by the copyright nuts. However, another part is caused by the decline and fall of curated media (good record stores, good radio stations) in favor of search engines whose algorithms are basically popularity contests, gamed to sell ads. These show the fallacy of the long tail argument (tm), because - as this/. thread shows - the search engines, and predictive algorithms (if you liked xx you should like yy) herd people into more tightly knit winner groups. The interwebs haven't hurt Lady Gaga; they've crushed lots of minor classical ensembles.
Naxos recordings are hit and miss. Some good, some poor, none stellar... it has breadth, but if you're looking for a breathtaking performance of Goldberg Variations, for example, you aren't going to find Glenn Gould or Murray Perahia or... on Naxos.
IMHO: iTunes is (or at least was... I stopped looking last year) pretty pathetic. Here's why. Suppose you want to listen to Bach suites for solo cello. Sure, they've got a version or two, but I want a version played by a master on a great instrument. Casals? Check, but old. Rostropovich? Nope, sorry.
Or, I want to listen to something (a lot) more current: Kronos Quartet? Some. Alarm will sound? ok. Bang on a Can? Nope, sorry.
It *does* seem to have both Glenn Gould recordings of Goldberg, which is an improvement (and, yes, they're very different).
This, especially the latter observation is surely connected to the recent/. discussion about use of computer-controlled instruments. It seems to have taken iTunes a very long time indeed to understand that two recordings of the same piece, by different ensembles or performers, using different instruments, under different circumstances, reveal the piece in entirely different ways. They're not the same thing.
There are two reasons firms do this: 1. The devices look prettier. This is the triumph of "industrial design" over function, similar to the way (it seems) Apple's industrial designers over-ruled the antenna / RF designers on the iPhone4. Same consequence: it's less easily usable, you have to learn to use the screen despite its failings. 2. Specsmanship. Glossy screens (called in the industry "glare screens", which really summarizes the issue) have higher contrast ratios - if the contrast ratio is measured in a perfectly dark room. Colors look nicely saturated. That way the vendors get to put very high contrast ratios on their specs and it's an arms race. Gottaproblemwiththat? Sit in a dark room, silly.
Of course, the only screens designed for reading (e-Ink, Pixel Qi, Sipix) do NOT use glare / glossy screens.
The patent troll reference in the parent article is irrelevant. Patent trolls are, in legal parlance, non-practicing entities. AUO is one of the big four makers of LCDs, with about $14B a year in LCD revenues. It has its own labs, does its own research, spends billions on fabs. Not sure why the first/. responses are ranting about the failure of the patent system - building an LCD fab is a huge financial risk, and finding a firm that is violating your IP and thus undercutting your market is a major challenge.
We sometimes forget just how heavy water is, or how much energy ocean waves carry.
Some time ago, I did some statistical analysis of wave heights in Scapa Flow, not far away from the site proposed here in northern Scottish waters. It has very steady, large swells.
Imagine a wave (or swell) of 10m peak height, extending 2 km across, and 50m front-to-back. That's a nice 0.3 * 10^6 kg of water... move it forward at 30kph... repeat every 10 or 20 seconds, and you've got 10^9 Joules/second, about 1GW. For the surface wave. (More energy is transferred more steadily by sub-surface currents.)
Lunar tidal flows are so much larger than these that the prospect of drawing enough energy from open waters to do anything to earth - moon movements seem to be off by many orders of magnitude.
Full disclosure: I used to be a pretty good physicist, but that was a long time ago.
Perhaps phone companies really ARE evil, don't know. But here's the way some of this works as a business: 1. Spectrum auctions (and landlords charging for antenna locations) are economically perfect mechanisms to drive the business case for wireless services to nearly non-existence. Spectrum auctions almost necessarily push telcos to pay nosebleed prices, just to participate. (The UK auctions were manically unhinged: they had a rider saying that BT would lose its GSM license unless it bought 3G spectrum. In consequence BT just about *had* to pay whatever it took, just to stay in business.) Auctions are not about valuing assets, they're a hidden tax. The cost of equipment is not nearly as critical a cost factor as the cost of cell sites (~100,000 per major carrier in the US) and the spectrum; both lack competitive supply/demand forces to contain them. Likewise, landlords are armed with economic models and consultants that drive every last red cent out of business models too. Hey, that's how business works. The cost of equipment is not nearly as critical a cost factor as the cost of cell sites (~100,000 per major carrier in the US) and the spectrum; both lack competitive supply/demand forces to contain them. Operating networks with tens of thousands of nodes in the USA's large landmass ain't cheap.
2. Along come smartphones and these and and apps, (and misleading marketing) create soaring basic demand;
3. Bloated apps (Skype, ugh), IP and (e.g.) the Van Jacobsen quickstart algorithm then take said traffic and inherently drive it to network saturation.
So: perhaps telco execs are satanic, let's get pitchforks and blazing torches.
But, the economics and technical dynamics of the marketplace are in inherent conflict. US gov't policies are at least as much to blame. And so are landlords.
The analysis can get much deeper - but without revealing a useful solution for the US, alas.
BUT the reason no Nobel Peace Prize was awarded in 1948, the year of Gandhi's assassination, was because he was the winner, and it's never awarded posthumously.
Some of the old data can be of great quality - so these exercises can be highly useful.
A couple of decades ago, I worked - as a student intern - at British institution. A question came in on wave heights in the North Sea... a firm was wondering about engineering tolerances for oil rigs and such. I had to go to the data: much from the last few decades was already computerized and I did a quite stats analysis - and was surprised at how many BIG waves were observed. This would be very costly to the rig builders... so I was told to go and re-sift the recent data and dig up older data. The recent data sift yielded the same output. The old data... going back to the 1700s... showed the same statistical patterns (so long as you squinted at it a bit - the responsible sailors either were not at sea and certainly were not taking measurements in big storms, or didn't get to survive). The outcome was - as I recall - that in this particular spot of the North Sea, you'd see a wave (or cluster of waves) over 40' high every two or so months.
The reason for the tight correlation, of course, is that the data was being taken the same way: sextants and the like, with data literally tabulated by hand: and - registered vessels had someone on board whose job it was to take and log the data - it wasn't something done ad hoc. The systemic errors were consistent for two-plus centuries. Data since the 1980s is automated and since the 1990s is from satellite maps.
... like network-based virus blockers bring several good things: * an entirely different set of algorithms can be used, leveraging data and traffic patterns not specific to the message contents * a team of engineers not tied to a single enterprise
And, indeed, major network operators like to do stuff like this - takes traffic off the network, and relieves enterprises of evil traffic forms (including DDOS)
BUT then, net neutrality purists, like 4chan, despise this and fight back, as recently when AT&T worked to thwart a large-scale DDOS attack.
About 5 years ago, I led a private project that looked at the terms most often searched over Limewire, Napster, etc. The results were most similar to those of an academic study that emerged a bit later: the most commonly-searched terms were NOT popular songs or p0rn. They were pings for bank account information, SSNs and passwords.
Of course this is utter BS. But lest we forget: the Wifi AP will be on 7/24, while the typical consumer micro will only be in use 30 mins per day, perhaps 5 days/week. Note: I'm not defending the nutty guy who alleges he's allergic to 2.4GHz from wifi, just saying the counterarguments should be watertight, and comparisons to consumer micros aren't inherently watertight. (And yes that is a pun, for those of you paying attention to why 2.4GHz works.)
Few tropes are as tired and tiresome as Edison's claims, such as 'I haven't failed; I've found 10,000 ways that don't work.'
Back then, Edison could (and did) hire hundreds of engineers, have them grind through the experiments, and then claim their inventions as his own. It was then legal for a single powerful executive to claim all patents as his own, even when he'd done none of the work. (Now, they just get to claim they're co-inventors.)
Switzerland is an extraordinary example of the challenges. Its economy, to a larger degree than any other country, depends on money laundering. So, the freedoms of the Swiss are propped up by the suffering of the populaces of countries whose dictators or corrupt magnates need to stash their ill-gotten gains somewhere safe. (Nazi theft of Jewish wealth all went to Swiss banks, for example.)
Slashdot Mirror
You can call bullshit all you want, but if you look at the work from Cal (lab of Jack Gallant) and the dozens of others who have replicated his work, you'd find you're wrong.
If you go to her company's website (opnwatr.io) you can see that its approach is based on high precision infrared imaging. If you look at what is going on in infrared imaging (e.g. via conferences, etc.) you will see that this is very much the focus (pun intended) of a lot of current research. AND that BOLD (blood oxygen level detection), which is what is measured in fMRI, can be measured with similar or greater precision via near-infrared techniques. And on and on ...
Oo, that's cool.
1. Neural lace creates digital I/O to the brain ... ...
2. Neural lace kills neurons
3. No problemo! Off board memory and proessors to the rescue - we're all backed up to and supported on some cloud service: AWS on steroids, almost literally.
4. Shared cloud service! We're all collaborative because we're all ants under the same control.
I'm not actually in favor of a/ injecting into my (admittedly defective in some ways) brain or b/ having my skull cut open, or c/ having probes jabbed into my grey-ish matter or ... well, actually, any of this.
"How is this even legal?"
Because the Internet Archive received a waiver from the DMCA for the purposes of archiving software that is no longer commercially available.
Taiwan is China? Same way that Japan is China and Vietnam is China and Korea (all of it) is China. Ethnically, Han; culturally not.
Taiwan was administered as a Chinese territory for only seven years of history - and spent rather longer (over 50 years) under Japanese rule. The only land you can see from any part of Taiwan (the island) is ... Japanese. The dominant ethnic groups are Han, the dominant languages are Chinese, but the dominant culture is Japanese. The island's dominant population until early in the 20th century was of the aboriginal groups (tribes?) which are not Chinese at all, but Austronesian. Taiwan has a vigorous and messy democracy, parties vie energetically for power. Rule of law, mainly, pertains, with corruption at a level comparable to the US.
And Taiwanese of many ages now think of China as having both the past - an historic link to culture - and a present and future - China the nation as a terrifying borg-like entity, not at all as being "we're China, we just disagree on where the capital". They think the capital of China is Beijing, and the capital of what should be the independent nation-state of Taiwan is Taipei.
And, while I'm white ... I live in Taiwan.
We're all grownups, many of us are nerds, technically literate and so are completely used to the idea of using reasonably long words for the precision they offer.
Can we please ban "pee" and "poo"? Always and forever.
I should have made more clear.
I was talking about value to the customer / user.
You're talking about costs. The incremental costs may become small (but they're not yet, see - for example - Gettys work on bufferbloat).
Meanwhile, lots of folks buy Apple computers for 2x the price of a more powerful Windows machine because, well, perceived value. Apple has high margins. Perhaps, in your worldview, we should regulate Apple's prices and margins?
A rather silly over-simplification.
Of course, voice is carried as data. However, it requires more than low latency - it requires that the latency be sustained as low. And it requires low error rates.
The reasons are buried deep in human behaviors. ... becomes "do you want to go out for dinner on Friday?" A: (slight pause) "yes".
Delays easily realizable in IP networks with error correction are perceptible to the listener. Then, however, they're not ignored (as they are in a video stream being re-aggregated for playing) but are heard by the listener as hesitation.
The Q&A: "do you want to go out for dinner on Friday?" A: "yes"
In human interaction, that silent pause is extra information.
(Of course, the degradation of voice quality on mobile networks means that the Q&A leads to answers like "huh? what did you say?")
There's a BIG difference between saying "voice is data" and the fact "voice is carried as data".
Some of this loss, doubtless, *is* caused by the copyright nuts. /. thread shows - the search engines, and predictive algorithms (if you liked xx you should like yy) herd people into more tightly knit winner groups. The interwebs haven't hurt Lady Gaga; they've crushed lots of minor classical ensembles.
However, another part is caused by the decline and fall of curated media (good record stores, good radio stations) in favor of search engines whose algorithms are basically popularity contests, gamed to sell ads. These show the fallacy of the long tail argument (tm), because - as this
Naxos recordings are hit and miss. Some good, some poor, none stellar ... it has breadth, but if you're looking for a breathtaking performance of Goldberg Variations, for example, you aren't going to find Glenn Gould or Murray Perahia or ... on Naxos.
IMHO: iTunes is (or at least was ... I stopped looking last year) pretty pathetic. Here's why. Suppose you want to listen to Bach suites for solo cello. Sure, they've got a version or two, but I want a version played by a master on a great instrument. Casals? Check, but old. Rostropovich? Nope, sorry.
Or, I want to listen to something (a lot) more current: Kronos Quartet? Some. Alarm will sound? ok. Bang on a Can? Nope, sorry.
It *does* seem to have both Glenn Gould recordings of Goldberg, which is an improvement (and, yes, they're very different).
This, especially the latter observation is surely connected to the recent /. discussion about use of computer-controlled instruments. It seems to have taken iTunes a very long time indeed to understand that two recordings of the same piece, by different ensembles or performers, using different instruments, under different circumstances, reveal the piece in entirely different ways. They're not the same thing.
There are two reasons firms do this:
1. The devices look prettier. This is the triumph of "industrial design" over function, similar to the way (it seems) Apple's industrial designers over-ruled the antenna / RF designers on the iPhone4. Same consequence: it's less easily usable, you have to learn to use the screen despite its failings.
2. Specsmanship. Glossy screens (called in the industry "glare screens", which really summarizes the issue) have higher contrast ratios - if the contrast ratio is measured in a perfectly dark room. Colors look nicely saturated. That way the vendors get to put very high contrast ratios on their specs and it's an arms race. Gottaproblemwiththat? Sit in a dark room, silly.
Of course, the only screens designed for reading (e-Ink, Pixel Qi, Sipix) do NOT use glare / glossy screens.
The patent troll reference in the parent article is irrelevant. Patent trolls are, in legal parlance, non-practicing entities. AUO is one of the big four makers of LCDs, with about $14B a year in LCD revenues. It has its own labs, does its own research, spends billions on fabs. Not sure why the first /. responses are ranting about the failure of the patent system - building an LCD fab is a huge financial risk, and finding a firm that is violating your IP and thus undercutting your market is a major challenge.
Surely: the logical thing for someone (not me!) to do is coordinate with Archive.org and then have them host it all in perpetuity?
We sometimes forget just how heavy water is, or how much energy ocean waves carry.
Some time ago, I did some statistical analysis of wave heights in Scapa Flow, not far away from the site proposed here in northern Scottish waters. It has very steady, large swells.
Imagine a wave (or swell) of 10m peak height, extending 2 km across, and 50m front-to-back. That's a nice 0.3 * 10^6 kg of water ... move it forward at 30kph ... repeat every 10 or 20 seconds, and you've got 10^9 Joules/second, about 1GW. For the surface wave. (More energy is transferred more steadily by sub-surface currents.)
Lunar tidal flows are so much larger than these that the prospect of drawing enough energy from open waters to do anything to earth - moon movements seem to be off by many orders of magnitude.
Full disclosure: I used to be a pretty good physicist, but that was a long time ago.
Perhaps phone companies really ARE evil, don't know.
But here's the way some of this works as a business:
1. Spectrum auctions (and landlords charging for antenna locations) are economically perfect mechanisms to drive the business case for wireless services to nearly non-existence. Spectrum auctions almost necessarily push telcos to pay nosebleed prices, just to participate. (The UK auctions were manically unhinged: they had a rider saying that BT would lose its GSM license unless it bought 3G spectrum. In consequence BT just about *had* to pay whatever it took, just to stay in business.) Auctions are not about valuing assets, they're a hidden tax. The cost of equipment is not nearly as critical a cost factor as the cost of cell sites (~100,000 per major carrier in the US) and the spectrum; both lack competitive supply/demand forces to contain them.
Likewise, landlords are armed with economic models and consultants that drive every last red cent out of business models too. Hey, that's how business works.
The cost of equipment is not nearly as critical a cost factor as the cost of cell sites (~100,000 per major carrier in the US) and the spectrum; both lack competitive supply/demand forces to contain them. Operating networks with tens of thousands of nodes in the USA's large landmass ain't cheap.
2. Along come smartphones and these and and apps, (and misleading marketing) create soaring basic demand;
3. Bloated apps (Skype, ugh), IP and (e.g.) the Van Jacobsen quickstart algorithm then take said traffic and inherently drive it to network saturation.
So: perhaps telco execs are satanic, let's get pitchforks and blazing torches.
But, the economics and technical dynamics of the marketplace are in inherent conflict. US gov't policies are at least as much to blame. And so are landlords.
The analysis can get much deeper - but without revealing a useful solution for the US, alas.
Were Douglas Adams still among us, he'd perhaps remind us that it could be a highway. Worse still, it could be a highway under construction.
BUT the reason no Nobel Peace Prize was awarded in 1948, the year of Gandhi's assassination, was because he was the winner, and it's never awarded posthumously.
Some of the old data can be of great quality - so these exercises can be highly useful.
A couple of decades ago, I worked - as a student intern - at British institution. A question came in on wave heights in the North Sea ... a firm was wondering about engineering tolerances for oil rigs and such. I had to go to the data: much from the last few decades was already computerized and I did a quite stats analysis - and was surprised at how many BIG waves were observed. This would be very costly to the rig builders ... so I was told to go and re-sift the recent data and dig up older data. The recent data sift yielded the same output. The old data ... going back to the 1700s ... showed the same statistical patterns (so long as you squinted at it a bit - the responsible sailors either were not at sea and certainly were not taking measurements in big storms, or didn't get to survive). The outcome was - as I recall - that in this particular spot of the North Sea, you'd see a wave (or cluster of waves) over 40' high every two or so months.
The reason for the tight correlation, of course, is that the data was being taken the same way: sextants and the like, with data literally tabulated by hand: and - registered vessels had someone on board whose job it was to take and log the data - it wasn't something done ad hoc. The systemic errors were consistent for two-plus centuries. Data since the 1980s is automated and since the 1990s is from satellite maps.
Where do you live? I was thinking of coming over and stealing your stuff, since I think I deserve it more.
... like network-based virus blockers bring several good things:
* an entirely different set of algorithms can be used, leveraging data and traffic patterns not specific to the message contents
* a team of engineers not tied to a single enterprise
And, indeed, major network operators like to do stuff like this - takes traffic off the network, and relieves enterprises of evil traffic forms (including DDOS)
BUT then, net neutrality purists, like 4chan, despise this and fight back, as recently when AT&T worked to thwart a large-scale DDOS attack.
About 5 years ago, I led a private project that looked at the terms most often searched over Limewire, Napster, etc. The results were most similar to those of an academic study that emerged a bit later: the most commonly-searched terms were NOT popular songs or p0rn. They were pings for bank account information, SSNs and passwords.
Of course this is utter BS. But lest we forget: the Wifi AP will be on 7/24, while the typical consumer micro will only be in use 30 mins per day, perhaps 5 days/week.
Note: I'm not defending the nutty guy who alleges he's allergic to 2.4GHz from wifi, just saying the counterarguments should be watertight, and comparisons to consumer micros aren't inherently watertight. (And yes that is a pun, for those of you paying attention to why 2.4GHz works.)
Few tropes are as tired and tiresome as Edison's claims, such as 'I haven't failed; I've found 10,000 ways that don't work.'
Back then, Edison could (and did) hire hundreds of engineers, have them grind through the experiments, and then claim their inventions as his own. It was then legal for a single powerful executive to claim all patents as his own, even when he'd done none of the work. (Now, they just get to claim they're co-inventors.)
Switzerland is an extraordinary example of the challenges.
Its economy, to a larger degree than any other country, depends on money laundering. So, the freedoms of the Swiss are propped up by the suffering of the populaces of countries whose dictators or corrupt magnates need to stash their ill-gotten gains somewhere safe. (Nazi theft of Jewish wealth all went to Swiss banks, for example.)