The up/down voting is rather terrible as it heavily favours clickbait and popular content (not to be confused with correct or good content). Slashdot's moderation system where your votes fall into categories (funny, insightful, etc.) is far more useful, as it tells you something about the content of the post, not just a number.
Moderation on Reddit also does often more harm than good as it actively works against the already existing scoring system and just makes mods power tripping and remove content that they don't like. There is nothing in place for users to override the mods or even just see what they are up to. Overreaching bans and deletions are common place on reddit.
The real strength of reddit is far basic: Everybody can create subreddits. This allows people do go different ways when they no longer like the state of a subreddit or when they have interests not covered by any of the existing ones. This allows reddit to grow far beyond the regular web forum that is limited to a few predefined categories that an admin has to invent.
Trolling is overrated. The amount of time a troll causes issues and isn't easy to ignore are pretty minimal. Meanwhile the problems caused by measures to deal with trolling tend to far outweigh the problem it causes in the first case, e.g. no anonymity, heavy handed moderation, shadow banning, censorship or just outright getting rid of comment sections altogether.
What makes this even more scummy is that Amazon is doing all they can to blur the lines between third party sellers and Amazon purchases. The marketplace is not a separate subsection of Amazon, it's completely integrated into the main shop and they will show you products sold by them and products sold by third parties interchangeably. To complicate things even further Amazon will stock some third party products in their own warehouses, so everything is handled by Amazon, except the liability. As a customer all you get is a line in tiny font saying "Sold by SomeCompany" to warn you that you are buying from some Chinese company that doesn't care about safety regulation, but unless you already had a few bad experiences with third party sellers that line is very easy to overlook.
Old consumer VR (VFX-1, Cybermaxx, i-glasses) used LCD micro-displays, only some of the professional ones used CRTs. The resolution also didn't even reach 320x200, but was instead some oddball number like 262 x 230. It wasn't quite the era of 800x600 yet, Descent, Quake and Co. where still mostly played in 320x200, while 640x480 was used in some strategy games (e.g. Warcraft 2), it's use in 3D games was still a few years away and only really started to go mainstream with the rise of 3D accelerator cards.
I think what killed VR back than was the lack of a standard 3D API due to DOS and most games like Doom not even being real 3D, which meant you couldn't have generic injection driver or headtracking. But probably even more important however was just the lack of progress. VR back when it was first released was quite impressive and held up well enough next to a 320x200 game, but it never really advanced beyond that initial point. The rest of the PC world made heaps of progress, 3D accelerators started appearing and VR just was kind of stuck. The VR companies started folding before they managed to put out a headset that could do 640x480 and those that survived a while longer shifted to business use instead of gaming. Took a good 15 years until the VR landscape started to advance again with the Rift.
A statistic wouldn't be all that useful at this point, as we don't have any real consumer self driving cars on the road, just experimental vehicles with safety drivers and those vehicles are only driven in condition that they are deemed to be able to handle. We don't even know how much or how little the safety driver had to intervene. Those cars could be terrible, but you still wouldn't notice since there is a human on the wheel helping out.
I'd be much more interested in seeing the self-driving software being made available to the public. Load it up into GTA and have the kids try to exploit it. I'd like to see how that software reacts to all those circumstances that fall outside the expected. I don't want self driving cars made safe by dead people on the road when many of those issues could be caught in simulation. It would also help the public trust those things when it can be shown that they work outside controlled conditions.
All the Kinect demos, Holo Lens demos, etc. were fake as fuck.
When Kudo demoed Kinect (then Project Natal) on stage at E3 his avatar turned into a pretzel. The fancy product videos of Project Natal were fake, but the state demo was pretty genuine at demonstrating the shortcomings of the product.
With Hololens they cheated on the FOV, but otherwise got pretty close to the actual product. All the room scanning and hand tracking is in the actual product. That RoboRaid game with the robots crawling out of the wall and over your furniture is available on Hololens, framerate looks however worse than the official demos.
This whole thing would have been so much easier if the FSF would have build a proper Linux distribution of their own and called it GNU, but they didn't (outside of that Hurd thing I booted up some 15+ years ago). So we have Debian, Ubuntu, Arch, Gentoo, SuSE, Redhat, Fedora and Co. instead and that's the names I tend to use when refering to my OS. GNU and Linux are just some small subset of software that is part of those distributions.
The name GNU/Linux still has a bit of value, as it differentiated the desktop Linux distributions from Linux distributions like Android that share almost nothing but the kernel, but it's not really a correct term, it's just the only common term I can think of that connects all those distributions.
The problem with a lot of safety measures is they make people behave less safe. Thus the benefit of the safety measure gets eaten up by people getting into more accidents. In the case of autopilot you don't just augment an attentive driver with additional features, you turn him into an inattentive driver when you give him autopilot abilities. See the recent Uber self-driving death, driver wasn't paying attention to the road and fumbling around with the phone, we can blame the driver, but that's ignoring the fact that that's just how humans behave when they get bored. So the technology doesn't just need to catch a few dangerous situations, it has to be able to fully replace an attentive driver. Furthermore people like to overestimate what the capabilities of the technology, a Tesla car is not a fully self driving vehicle, but many people treat it like one and go hands-free, despite the technology not being able to handle that.
All that said, the biggest issue in judging self-driving cars is the lack of data points. Not just in actual crashes, but also in what the self driving technology is actually capable off. We have a few videos of the technology behaving correctly, but hardly anything that explores the limits of their abilities. I'd much prefer it if those car companies would just stuff their self-driving software into a simulation and have the public have some fun with trying to exploit it, than slowly collecting data points by the means of dead people on the road. So far there also doesn't even seem to be an official driving test for self-driving cars, we just trust the manufacturer's to do the testing themselves.
Powerful hardware isn't necessary when they have eye tracking and foveated rendering. The human eye only sees high resolution in a tiny spot in the center, everything else can be rendered at much lower resolutions.
I don't quite get who the target audience of those are. All the social media seems to be focused on people posting photos of themselves, not of stuff they are looking at. So a selfie-stick makes sense for that audience. The camera on these glasses however is pointing in the wrong direction. These glasses might have more success if they'd changed the design into something more serious and make it a cheaper consumer focused alternative to a Taser Axon.
The whole loot-box thing started because gamers were beginning to complain and boycotting games. Politicians had nothing to do with that, they are just reacting to those complains now. And depending on how exactly the gambling laws are written, those games were already illegal to begin with, the law just wasn't enforced properly.
To me the OLPC had three major points where it fell short:
Lack of regular consumer availability. They only sold the thing to the government, not to regular consumers. The Give1Get1 program was time limited and overpriced, since you were buying two. Thus the EeePC stole the show, since it was hardware that you could actually buy instead of just read about. The OLPC might have fared much better if they had released an adult version, with a bigger keyboard, more RAM and better color scheme.
Lack of self-hosting. While the idea of allowing people to write programs on the OLPC was there, that was never really all that practical. The software wasn't up to snuff and the documentation was lacking. Thus the OLPC ended up feeling much more like a consumer-only device, like a modern Android tablet, than a machine you could build stuff with yourself.
Underpowered hardware. 1GB Flash and 256MB RAM just wasn't enough, especially when it comes to Web browsing. It would fit the core OS barely, but it would drastically limit what you could do with the device. Double the storage and RAM would have increased the price, but it would also have lead to a much more useful device.
All that said, the OLPC itself might not have been the success they wanted it to be, but it was still the starting point of cheap computing devices. The idea of a $100 PC was utopic back than. Now I can buy $50 Android tablets and $30 Raspberry Pis.
Tablets flat out didn't exist back when the OLPC launched (outside of some business PCs that cost like $2000). The iPad came three years later. Netbooks didn't yet exist either, the OLPC hype pretty much created that branch of the market.
For a recent event, see the banning of grid girls in F1. For something older, the banning of booth babes at E3. And that's not even prostitution, that's just attractive girls hired for money, but that's already considered demeaning so it must be banned.
With the possible exception of Uber, most autonomous vehicle companies can already beat those death rates.
The problem is that we don't know how those cars are tested. It's easy to not kill anybody when you have a safety driver on board and decide when you let the car on the road as that allows you to avoid all the situations that the car can't currently handle. But what when it runs into unexpected an blizzard, rain, hail, sandstorm or whatever? What about those situations where just stopping the car could get you killed, like for example in a forest fire or in the middle of nowhere in cold weather? Will it be able to deal with that? We can hope that the companies have prepared for all those freak situations, but the Uber crash clearly shows that they could also fail. What makes the Uber crash so troublesome is that this wasn't even a freak situation, this was automatic-braking 101, something that would have shown up even in the most basic of testing.
On top of that comes the issue that those cars don't just have to deal with natural occurrences, but also with adversarial attacks. What if somebody paints a few white stripes on the road that lead of a cliff? Will the cars just follow those lines and drive into the abyss?
That's the kind of stuff I'd like to see tested before we let them on the road in mass, not after the fact when they get people killed like in the case of Uber or the two Tesla crashes. I want to know when those cars fail, not just when they work.
The Web has been dying ever since its inception, one link at a time. The problem with the Web is that the links in it don't point to content, but to locations. So whenever a server goes down, be it temporarily or permanently, the content disappears and the links go dead. Even a bit of renaming on the the server site makes links go dead. That's one of the problems IPFS addresses by making links based on a content hash, not the current storage location.
That issue cascades into a whole lot of other issues and is one of the reasons why Facebook and Co. are so popular, since they can provide a stable content host, something that wouldn't be possible with regular HTTP and everybody renting their own server. With IPFS you can have a stable and self hosted web, since the storage location no longer is the thing that holds the web together.
In theory, sure, you could make detailed 3D scans of everything and do it all in VR. But for that to happen, somebody must first put in the effort to create those virtual 3D tours as well as get the permission to produce them to begin with. Neither of which are easy. Museums like to keep their things under lock, they want the visitors to come and spend money there, not make themselves obsolete by having some amazing digital reproduction. Producing those 3D scans, that are detailed enough to replace a real visit, would also need a lot of money and technology.
People have been making those same claims about revolutionizing teaching with every new technology. They did it with the radio, the TV, they did it with the multimedia CD-ROMs, with the Internet and so on. It never had all that much impact. If you really want to figure out how stuff works, you still need to go to a library and get a book. It's not that new technology couldn't do it better, but simply the result of there not being a viable business model to produce that kind of content on a scale that could impact teaching at large. People protecting their already established business models doesn't help either.
Torrent get you sued, while sharehosters generally don't and when you just stream they might even be completely legal in your jurisdiction. So there is good reason to switch away from torrent when you aren't using a VPN to conceal your identity.
And on top of that comes that Google isn't even all that relevant for product searches, Amazon and eBay are way more useful when you want to actually find stuff you can buy. Google product search on the other side has been a complete disappointment ever since it's inception.
Slashdot Mirror
The up/down voting is rather terrible as it heavily favours clickbait and popular content (not to be confused with correct or good content). Slashdot's moderation system where your votes fall into categories (funny, insightful, etc.) is far more useful, as it tells you something about the content of the post, not just a number.
Moderation on Reddit also does often more harm than good as it actively works against the already existing scoring system and just makes mods power tripping and remove content that they don't like. There is nothing in place for users to override the mods or even just see what they are up to. Overreaching bans and deletions are common place on reddit.
The real strength of reddit is far basic: Everybody can create subreddits. This allows people do go different ways when they no longer like the state of a subreddit or when they have interests not covered by any of the existing ones. This allows reddit to grow far beyond the regular web forum that is limited to a few predefined categories that an admin has to invent.
Trolling is overrated. The amount of time a troll causes issues and isn't easy to ignore are pretty minimal. Meanwhile the problems caused by measures to deal with trolling tend to far outweigh the problem it causes in the first case, e.g. no anonymity, heavy handed moderation, shadow banning, censorship or just outright getting rid of comment sections altogether.
What makes this even more scummy is that Amazon is doing all they can to blur the lines between third party sellers and Amazon purchases. The marketplace is not a separate subsection of Amazon, it's completely integrated into the main shop and they will show you products sold by them and products sold by third parties interchangeably. To complicate things even further Amazon will stock some third party products in their own warehouses, so everything is handled by Amazon, except the liability. As a customer all you get is a line in tiny font saying "Sold by SomeCompany" to warn you that you are buying from some Chinese company that doesn't care about safety regulation, but unless you already had a few bad experiences with third party sellers that line is very easy to overlook.
Old consumer VR (VFX-1, Cybermaxx, i-glasses) used LCD micro-displays, only some of the professional ones used CRTs. The resolution also didn't even reach 320x200, but was instead some oddball number like 262 x 230. It wasn't quite the era of 800x600 yet, Descent, Quake and Co. where still mostly played in 320x200, while 640x480 was used in some strategy games (e.g. Warcraft 2), it's use in 3D games was still a few years away and only really started to go mainstream with the rise of 3D accelerator cards.
I think what killed VR back than was the lack of a standard 3D API due to DOS and most games like Doom not even being real 3D, which meant you couldn't have generic injection driver or headtracking. But probably even more important however was just the lack of progress. VR back when it was first released was quite impressive and held up well enough next to a 320x200 game, but it never really advanced beyond that initial point. The rest of the PC world made heaps of progress, 3D accelerators started appearing and VR just was kind of stuck. The VR companies started folding before they managed to put out a headset that could do 640x480 and those that survived a while longer shifted to business use instead of gaming. Took a good 15 years until the VR landscape started to advance again with the Rift.
A statistic wouldn't be all that useful at this point, as we don't have any real consumer self driving cars on the road, just experimental vehicles with safety drivers and those vehicles are only driven in condition that they are deemed to be able to handle. We don't even know how much or how little the safety driver had to intervene. Those cars could be terrible, but you still wouldn't notice since there is a human on the wheel helping out.
I'd be much more interested in seeing the self-driving software being made available to the public. Load it up into GTA and have the kids try to exploit it. I'd like to see how that software reacts to all those circumstances that fall outside the expected. I don't want self driving cars made safe by dead people on the road when many of those issues could be caught in simulation. It would also help the public trust those things when it can be shown that they work outside controlled conditions.
All the Kinect demos, Holo Lens demos, etc. were fake as fuck.
When Kudo demoed Kinect (then Project Natal) on stage at E3 his avatar turned into a pretzel. The fancy product videos of Project Natal were fake, but the state demo was pretty genuine at demonstrating the shortcomings of the product.
With Hololens they cheated on the FOV, but otherwise got pretty close to the actual product. All the room scanning and hand tracking is in the actual product. That RoboRaid game with the robots crawling out of the wall and over your furniture is available on Hololens, framerate looks however worse than the official demos.
This whole thing would have been so much easier if the FSF would have build a proper Linux distribution of their own and called it GNU, but they didn't (outside of that Hurd thing I booted up some 15+ years ago). So we have Debian, Ubuntu, Arch, Gentoo, SuSE, Redhat, Fedora and Co. instead and that's the names I tend to use when refering to my OS. GNU and Linux are just some small subset of software that is part of those distributions.
The name GNU/Linux still has a bit of value, as it differentiated the desktop Linux distributions from Linux distributions like Android that share almost nothing but the kernel, but it's not really a correct term, it's just the only common term I can think of that connects all those distributions.
The problem with a lot of safety measures is they make people behave less safe. Thus the benefit of the safety measure gets eaten up by people getting into more accidents. In the case of autopilot you don't just augment an attentive driver with additional features, you turn him into an inattentive driver when you give him autopilot abilities. See the recent Uber self-driving death, driver wasn't paying attention to the road and fumbling around with the phone, we can blame the driver, but that's ignoring the fact that that's just how humans behave when they get bored. So the technology doesn't just need to catch a few dangerous situations, it has to be able to fully replace an attentive driver. Furthermore people like to overestimate what the capabilities of the technology, a Tesla car is not a fully self driving vehicle, but many people treat it like one and go hands-free, despite the technology not being able to handle that.
All that said, the biggest issue in judging self-driving cars is the lack of data points. Not just in actual crashes, but also in what the self driving technology is actually capable off. We have a few videos of the technology behaving correctly, but hardly anything that explores the limits of their abilities. I'd much prefer it if those car companies would just stuff their self-driving software into a simulation and have the public have some fun with trying to exploit it, than slowly collecting data points by the means of dead people on the road. So far there also doesn't even seem to be an official driving test for self-driving cars, we just trust the manufacturer's to do the testing themselves.
They do have a 'hands on steering wheel' sensor.
You mean the cars don't go into pull-over-and-park-safely mode if the driver is missing or seems to be asleep or incapacitated?
Tesla's will give a warning to put the hands back on the wheel and will slow down if that doesn't happen. However the warning can be worked around.
Powerful hardware isn't necessary when they have eye tracking and foveated rendering. The human eye only sees high resolution in a tiny spot in the center, everything else can be rendered at much lower resolutions.
I don't quite get who the target audience of those are. All the social media seems to be focused on people posting photos of themselves, not of stuff they are looking at. So a selfie-stick makes sense for that audience. The camera on these glasses however is pointing in the wrong direction. These glasses might have more success if they'd changed the design into something more serious and make it a cheaper consumer focused alternative to a Taser Axon.
The whole loot-box thing started because gamers were beginning to complain and boycotting games. Politicians had nothing to do with that, they are just reacting to those complains now. And depending on how exactly the gambling laws are written, those games were already illegal to begin with, the law just wasn't enforced properly.
That was the design phase, the crank never went into production.
To me the OLPC had three major points where it fell short:
Lack of regular consumer availability. They only sold the thing to the government, not to regular consumers. The Give1Get1 program was time limited and overpriced, since you were buying two. Thus the EeePC stole the show, since it was hardware that you could actually buy instead of just read about. The OLPC might have fared much better if they had released an adult version, with a bigger keyboard, more RAM and better color scheme.
Lack of self-hosting. While the idea of allowing people to write programs on the OLPC was there, that was never really all that practical. The software wasn't up to snuff and the documentation was lacking. Thus the OLPC ended up feeling much more like a consumer-only device, like a modern Android tablet, than a machine you could build stuff with yourself.
Underpowered hardware. 1GB Flash and 256MB RAM just wasn't enough, especially when it comes to Web browsing. It would fit the core OS barely, but it would drastically limit what you could do with the device. Double the storage and RAM would have increased the price, but it would also have lead to a much more useful device.
All that said, the OLPC itself might not have been the success they wanted it to be, but it was still the starting point of cheap computing devices. The idea of a $100 PC was utopic back than. Now I can buy $50 Android tablets and $30 Raspberry Pis.
Tablets flat out didn't exist back when the OLPC launched (outside of some business PCs that cost like $2000). The iPad came three years later. Netbooks didn't yet exist either, the OLPC hype pretty much created that branch of the market.
Care to point any out?
For a recent event, see the banning of grid girls in F1. For something older, the banning of booth babes at E3. And that's not even prostitution, that's just attractive girls hired for money, but that's already considered demeaning so it must be banned.
With the possible exception of Uber, most autonomous vehicle companies can already beat those death rates.
The problem is that we don't know how those cars are tested. It's easy to not kill anybody when you have a safety driver on board and decide when you let the car on the road as that allows you to avoid all the situations that the car can't currently handle. But what when it runs into unexpected an blizzard, rain, hail, sandstorm or whatever? What about those situations where just stopping the car could get you killed, like for example in a forest fire or in the middle of nowhere in cold weather? Will it be able to deal with that? We can hope that the companies have prepared for all those freak situations, but the Uber crash clearly shows that they could also fail. What makes the Uber crash so troublesome is that this wasn't even a freak situation, this was automatic-braking 101, something that would have shown up even in the most basic of testing.
On top of that comes the issue that those cars don't just have to deal with natural occurrences, but also with adversarial attacks. What if somebody paints a few white stripes on the road that lead of a cliff? Will the cars just follow those lines and drive into the abyss?
That's the kind of stuff I'd like to see tested before we let them on the road in mass, not after the fact when they get people killed like in the case of Uber or the two Tesla crashes. I want to know when those cars fail, not just when they work.
Is there any testing and certification done for those cars or do they trust the companies to handle that by themselves?
The Web has been dying ever since its inception, one link at a time. The problem with the Web is that the links in it don't point to content, but to locations. So whenever a server goes down, be it temporarily or permanently, the content disappears and the links go dead. Even a bit of renaming on the the server site makes links go dead. That's one of the problems IPFS addresses by making links based on a content hash, not the current storage location.
That issue cascades into a whole lot of other issues and is one of the reasons why Facebook and Co. are so popular, since they can provide a stable content host, something that wouldn't be possible with regular HTTP and everybody renting their own server. With IPFS you can have a stable and self hosted web, since the storage location no longer is the thing that holds the web together.
I wonder how hard it would be to crowdsource the same work.
Project Gutenberg has been at it since the 70's. But they currently only have 54.000 books, not a whole lot compared to Google's 25 million books.
In theory, sure, you could make detailed 3D scans of everything and do it all in VR. But for that to happen, somebody must first put in the effort to create those virtual 3D tours as well as get the permission to produce them to begin with. Neither of which are easy. Museums like to keep their things under lock, they want the visitors to come and spend money there, not make themselves obsolete by having some amazing digital reproduction. Producing those 3D scans, that are detailed enough to replace a real visit, would also need a lot of money and technology.
People have been making those same claims about revolutionizing teaching with every new technology. They did it with the radio, the TV, they did it with the multimedia CD-ROMs, with the Internet and so on. It never had all that much impact. If you really want to figure out how stuff works, you still need to go to a library and get a book. It's not that new technology couldn't do it better, but simply the result of there not being a viable business model to produce that kind of content on a scale that could impact teaching at large. People protecting their already established business models doesn't help either.
Torrent get you sued, while sharehosters generally don't and when you just stream they might even be completely legal in your jurisdiction. So there is good reason to switch away from torrent when you aren't using a VPN to conceal your identity.
And on top of that comes that Google isn't even all that relevant for product searches, Amazon and eBay are way more useful when you want to actually find stuff you can buy. Google product search on the other side has been a complete disappointment ever since it's inception.
Battery life makes them useless. Glass lasted only something like 30min when recording video, a Taser Axon can lasts 12 hours in comparison.