4 Console SKUs (Xbox One, Xbox One S, Xbox One X, Xbox Two) is a LOT fewer variations for a developer to target than then billion hardware\driver\utility\patch variations in Windows.
VW makes cars, too. EV and ICE. By the millions. And they tend to have a much better track record of launching vehicles on-time, and in all promised configurations, unlike Tesla.
The model 3 was 6 months late, which was 12 months earlier than they originally planned. (Model S and X were super delayed).
Model 3 might not be made by the "millions" but supposedly the current bottleneck is batteries not sheet metal. The world's largest battery producer by far... is battery constrained. Volkswagen is relying on third parties for their batteries. Chevy is using those same third parties and they can't even produce 3,000 cars a month let alone a week.
This is the story that nobody is pointing out about Tesla and GM. GM can make millions of cars per year too, but they can't even make a few thousand due to battery production constraints!
FCA and Tesla will sell the same amount of cars as if there wasn't a law in place.
FCA literally is paying Tesla hundreds of dollars per car that FCA sells.
If FCA sold 1 million cars and zero emission credits sold for $500 each that means Tesla would make $500m. That would allow Tesla to reduce the price of their 500k vehicles by $1k each.
That makes Tesla more competitive with FCA's polluting vehicles. The more cost competitive Tesla is, the sooner EVs appeal to more people. The more people that Tesla appeals to, the more cars they sell and the more efficiently (cheaper) they can sell cars. The cheaper their cars, the more they sell... etc... etc...
Meanwhile FCA's vehicles cost more. They pass on those costs to their consumers. So the EV gets cheaper and the ICE gets more expensive. Unless the car market works differently from the rest of the market, the result is fewer ICE sales and more EV sales.
Linus makes the mistake though of assuming that "Compile" is a step. Lots of Serverless applications are just javascript snippets. There are no compile bug concerns between platforms you just specify a.js block of code and execute. The platform is irrelevant. Serverless cloud providers don't even tell you what underlying CPU is executing your command.
There's plenty of serverless web tasks to justify ARM if it is cheaper/more efficient/has lower startup times.
Of course the Chinese grounded the plane. The Chinese want every Chinese Airline to buy exclusively Chinese built aircraft. Of course they would jump at the chance to ground their competitor. That's like saying "Airbus led the way in grounding all Boeing planes." The Chinese government needs to be viewed like a corporation not a certification authority.
As someone who works with raytracers you're spouting gibberish.
There are two components to raytracing: ray intersection and shading. No shit that Nvidia uses their SHADING hardware to shade raytraced rays. What they added was ray-intersection hardware which yes does take a good bit of processing power on non-trivial scenes.
Yes ray cohesion is important and no Nvidia didn't address it (like arguably Caustic\IMG did with their now-dead OpenRL raytracer) but they are in fact tracing rays.
As to raytracing needing "bouncy rays". You're confusing Global Illumination with "Ray Tracing". Ray tracing is just firing rays and returning the results. You are using a no-true-scotsman falacy to equate one with the other. If you raytrace only primary visible rays (like a rasterizer) you're still raytracing even if you have 0 bounces. And bounces is exactly where the RT cores do help boost the trace rate on the RTX GPUs.
As to why there are Tensor Cores? Because you already mentioned that you need millions of bouncing rays to deliver global illumination. Imagination Technologies almost delivered true ray counts high enough for GI but it was still too noisy. Nvidia without cohesion for its shaders said "how can we do global illumination without ray counts high enough?" and the answer was to apply a denoising neural net. So RT cores + a large tensor unit means they can denoise their low sample count raytracing into something useful.
As to "Rasterization can do anything raytracing can do but better". That's a load of bullocks. Raytraced shadows are infinitely more memory efficient. You also can't have self-reflections with reflection maps. There is a reason every single production renderer in existence today for high end visual effects is now a raytracer. Ray tracing is more efficient, it's faster and it produces far superior quality to rasterization hacks.
Dragon 2 is dead end tech for SpaceX now. They have a lucrative contract with NASA and they are going to do the bare minimum to deliver it safely. All development focus is now moving to Starship.
The article is full of shit. Here is the actual new language:
All new Tesla cars have the hardware needed in the future for full self-driving in almost all circumstances. The system is designed to be able to conduct short and long distance trips with no action required by the person in the driverâ(TM)s seat.
All you will need to do is get in and tell your car where to go. If you donâ(TM)t say anything, the car will look at your calendar and take you there as the assumed destination or just home if nothing is on the calendar. Your Tesla will figure out the optimal route, navigate urban streets (even without lane markings), manage complex intersections with traffic lights, stop signs and roundabouts, and handle densely packed freeways with cars moving at high speed. When you arrive at your destination, simply step out at the entrance and your car will enter park seek mode, automatically search for a spot and park itself. A tap on your phone summons it back to you.
The future use of these features without supervision is dependent on achieving reliability far in excess of human drivers as demonstrated by billions of miles of experience, as well as regulatory approval, which may take longer in some jurisdictions. As these self-driving capabilities are introduced, your car will be continuously upgraded through over-the-air software updates.
The only thing they changed was to put the imminently releasing features up front on the order page and to be more honest on the largest obstacles to release. They used to say "dependent on government approval" which was bullshit. Now they correctly say dependent on actually working. But the Summary claiming that they are only going to do L2 driver assist features is bullshit. Especially since Navigate on Autopilot will be L3 and Advanced Summon will be L3 and both are already being tested in public cars.
Tesla wanted there to be a reason to sell FSD now and that reason is Advanced Summon and NOA. Splitting the price and the features out of EAP and just releasing AP makes the most valuable EAP features available to more people for less money to be more in line with Honda and Audi pricing for L2 features.
it's fundamentally insane. Asking a member of the public to make connections that carry megawatts is Bonkers. And in fact it never will be anything but bonkers... ever. The only form of dense energy storage that we've come up with that is not so explosive is in fact gasoline.
Considering how many people light themselves on fire every year at gas stations that's a bold claim (The answer is about 50 with about 2 dying).
Except "businesses" are generally effectively run to benefit each employee in whatever position they hold. Generally, that means that in practice every employee seeks to do the minimum work while still keeping their job or being promoted. That means most decisions are pretty "lazy" and nothing is lazier than hiring someone you know.
The problem with that is you generally know people of the same gender\race\socioeconomic status. So many jobs are in reality not filled through an unbiased recruitment strategy but "hey my nephew is interested in your line of work and just graduated." "Oh yeah, send him on over we're currently hiring!"
Untested Emergency Automatic Braking systems are incredibly dangerous and require extensive validation. You slam on your brakes and a motorcyclist behind you dies.
And as of today vehicles aren't legally required to have one, so disabling the feature doesn't make your vehicle any less safe or less legal than the hundreds of millions of older vehicles on our roads without an EAB.
I think Linus is half-right. He's right that nobody is going to want to run an AWS Arm server. He's wrong in that a lot of the world is moving to serverless. There aren't weird bugs that are platform dependent between Node.JS on a Windows, Linux or ARM server. I just upload the code to the cloud and it could be running on PowerPC for all I care.
For serverless applications, if the cloud provider can get the javascript to run the user will be none the wiser. And they can develop at home.
Arguably though I would say that if VS Code could be run in the cloud as a website connected to your cloud provider directly that would be huge. Commit to git, work in a web browser (VSCode already is a neutron app), have your dev web server be on a shared cloud instance. Just switch from production to live targets. Keep your dev environment consistent between every computer you use since it's just a web browser interface.
And all of that could be running on ARM or x86 and nobody would care. Like it or not that's a huge portion of the world's dev work.
I've developed for iOS\Android and for Windows 10. Windows 10 was SOOOooooo much easier to develop for because of exactly what linus talks about. I could quickly figure out why my weird touch gesture wasn't working by hitting "run" not by compiling, transfering, launching, remotely debugging etc.
Many other cars have a channel around the inside of the boot lid that catches the water that runs off the lid when you open it,
So does the Model 3. But it's hip to ceramic coat their entire model 3 aka put on a full body hydrophobic coating. What happens is the water runs off so fast that it shoots several inches across the gap into the trunk (boot) and across the channel.
A stock Model 3 doesn't have this issue because the water doesn't evil kenivel the gap at highspeed.
The boot (trunk in US English) lets water in when you open it after its been raining too.
Only if you do what Tesla tells you explicitly not to do: ceramic coat\rainX the rear window.
If the water runs off at a regular pace with regular adhesion the water pours into the guides. If you put a hydrophobic coating on your rear window it runs off so fast that it overshoots the water guides and into the trunk.
And if they work, liek the tunnels, they end up being about the same as those made by others.
Musk's tunnel cost $10m for a 1.4 miles. For comparison my city just built a #)!@ bike lane for $12m/mile. They could have built an exclusive right-of-way bike tunnel for less with Boring Company.
By the way I assume this is what you're reffering to for "about the same"
For a better comparison Super Excavators (the previous owners of Godot) used the exact same machine to build a 1,640 ft sewer overflow tunnel for $12.4 million, or scaling up $38 million/mile
Hardly "about the same" and this was their very first attempt which was more about fact-finding than a finished product. Not to mention the competitor using the same boring machine is using it for sewers. Even at $40m a mile vs $10m or $1m a mile that's an innovation to repurpose a small bore tunnel for transit.
I would much rather pay by the GB than pay by the mbps.
I would much rather have 10gbe to my home and be able to upload 1TB a month as fast as the network allows than meter my usage via speed caps. Just put a QoS baseline on everyone that "guarantees" at least 50mbps and then let it run free. At 2am I can then upload my cloud backups for the night and have it done by the morning.
After all home\business internet naturally has different peaks throughout the day.
You're bending over backward though to assign sane and skilled leadership skills to someone who does not have them. I will illustrate it with two different questions.
Can we get someone to Mars before the end of my first term with infinite money?
With infinite money, how soon could we get someone to mars?
They're subtly different but one question is an intelligent question that identifies bottlenecks. The other is a vanity request.
The only reason that cable companies have given for Net Neutrality is that they will increase investment in infrastructure if they have a financial incentive.
When Tesla is unprofitable (even with "inflated" prices) they say Tesla is a sham that can't make money so we shouldn't be subsidizing a failed business model.
When those "inflated" (subsidized) prices finally help build a sustainable business suddenly it's proof that they never needed subsidies and Tesla was always going to succeed anyway.
The same is said of wind power and solar. We shouldn't offer solar\wind subsidies because they're a dead end waste of money. When solar\wind finally are cost effective then suddenly those subsidies are scams because solar\wind is cost competitive.
I guess expecting people to see cause\effect relationships is asking too much. Kind of like the person who said Bush and Obama shared "equal blame" for the financial crisis... of 2007 (a year before Obama took office).
Slashdot Mirror
4 Console SKUs (Xbox One, Xbox One S, Xbox One X, Xbox Two) is a LOT fewer variations for a developer to target than then billion hardware\driver\utility\patch variations in Windows.
VW makes cars, too. EV and ICE. By the millions. And they tend to have a much better track record of launching vehicles on-time, and in all promised configurations, unlike Tesla.
The model 3 was 6 months late, which was 12 months earlier than they originally planned. (Model S and X were super delayed).
Model 3 might not be made by the "millions" but supposedly the current bottleneck is batteries not sheet metal. The world's largest battery producer by far... is battery constrained. Volkswagen is relying on third parties for their batteries. Chevy is using those same third parties and they can't even produce 3,000 cars a month let alone a week.
This is the story that nobody is pointing out about Tesla and GM. GM can make millions of cars per year too, but they can't even make a few thousand due to battery production constraints!
FCA and Tesla will sell the same amount of cars as if there wasn't a law in place.
FCA literally is paying Tesla hundreds of dollars per car that FCA sells.
If FCA sold 1 million cars and zero emission credits sold for $500 each that means Tesla would make $500m. That would allow Tesla to reduce the price of their 500k vehicles by $1k each.
That makes Tesla more competitive with FCA's polluting vehicles. The more cost competitive Tesla is, the sooner EVs appeal to more people. The more people that Tesla appeals to, the more cars they sell and the more efficiently (cheaper) they can sell cars. The cheaper their cars, the more they sell... etc... etc...
Meanwhile FCA's vehicles cost more. They pass on those costs to their consumers. So the EV gets cheaper and the ICE gets more expensive. Unless the car market works differently from the rest of the market, the result is fewer ICE sales and more EV sales.
Linus makes the mistake though of assuming that "Compile" is a step. Lots of Serverless applications are just javascript snippets. There are no compile bug concerns between platforms you just specify a .js block of code and execute. The platform is irrelevant. Serverless cloud providers don't even tell you what underlying CPU is executing your command.
There's plenty of serverless web tasks to justify ARM if it is cheaper/more efficient/has lower startup times.
Of course the Chinese grounded the plane. The Chinese want every Chinese Airline to buy exclusively Chinese built aircraft. Of course they would jump at the chance to ground their competitor. That's like saying "Airbus led the way in grounding all Boeing planes." The Chinese government needs to be viewed like a corporation not a certification authority.
As someone who works with raytracers you're spouting gibberish.
There are two components to raytracing: ray intersection and shading. No shit that Nvidia uses their SHADING hardware to shade raytraced rays. What they added was ray-intersection hardware which yes does take a good bit of processing power on non-trivial scenes.
Yes ray cohesion is important and no Nvidia didn't address it (like arguably Caustic\IMG did with their now-dead OpenRL raytracer) but they are in fact tracing rays.
As to raytracing needing "bouncy rays". You're confusing Global Illumination with "Ray Tracing". Ray tracing is just firing rays and returning the results. You are using a no-true-scotsman falacy to equate one with the other. If you raytrace only primary visible rays (like a rasterizer) you're still raytracing even if you have 0 bounces. And bounces is exactly where the RT cores do help boost the trace rate on the RTX GPUs.
As to why there are Tensor Cores? Because you already mentioned that you need millions of bouncing rays to deliver global illumination. Imagination Technologies almost delivered true ray counts high enough for GI but it was still too noisy. Nvidia without cohesion for its shaders said "how can we do global illumination without ray counts high enough?" and the answer was to apply a denoising neural net. So RT cores + a large tensor unit means they can denoise their low sample count raytracing into something useful.
As to "Rasterization can do anything raytracing can do but better". That's a load of bullocks. Raytraced shadows are infinitely more memory efficient. You also can't have self-reflections with reflection maps. There is a reason every single production renderer in existence today for high end visual effects is now a raytracer. Ray tracing is more efficient, it's faster and it produces far superior quality to rasterization hacks.
LASIK is down to about $200 per eye
Citation needed, I keep checking prices and the bargain basement back alley LASIK is still $1k.
Dragon 2 is dead end tech for SpaceX now. They have a lucrative contract with NASA and they are going to do the bare minimum to deliver it safely. All development focus is now moving to Starship.
The article is full of shit. Here is the actual new language:
All new Tesla cars have the hardware needed in the future for full self-driving in almost all circumstances. The system is designed to be able to conduct short and long distance trips with no action required by the person in the driverâ(TM)s seat.
All you will need to do is get in and tell your car where to go. If you donâ(TM)t say anything, the car will look at your calendar and take you there as the assumed destination or just home if nothing is on the calendar. Your Tesla will figure out the optimal route, navigate urban streets (even without lane markings), manage complex intersections with traffic lights, stop signs and roundabouts, and handle densely packed freeways with cars moving at high speed. When you arrive at your destination, simply step out at the entrance and your car will enter park seek mode, automatically search for a spot and park itself. A tap on your phone summons it back to you.
The future use of these features without supervision is dependent on achieving reliability far in excess of human drivers as demonstrated by billions of miles of experience, as well as regulatory approval, which may take longer in some jurisdictions. As these self-driving capabilities are introduced, your car will be continuously upgraded through over-the-air software updates.
The only thing they changed was to put the imminently releasing features up front on the order page and to be more honest on the largest obstacles to release. They used to say "dependent on government approval" which was bullshit. Now they correctly say dependent on actually working. But the Summary claiming that they are only going to do L2 driver assist features is bullshit. Especially since Navigate on Autopilot will be L3 and Advanced Summon will be L3 and both are already being tested in public cars.
Tesla wanted there to be a reason to sell FSD now and that reason is Advanced Summon and NOA. Splitting the price and the features out of EAP and just releasing AP makes the most valuable EAP features available to more people for less money to be more in line with Honda and Audi pricing for L2 features.
it's fundamentally insane. Asking a member of the public to make connections that carry megawatts is Bonkers. And in fact it never will be anything but bonkers... ever. The only form of dense energy storage that we've come up with that is not so explosive is in fact gasoline.
Considering how many people light themselves on fire every year at gas stations that's a bold claim (The answer is about 50 with about 2 dying).
Except "businesses" are generally effectively run to benefit each employee in whatever position they hold. Generally, that means that in practice every employee seeks to do the minimum work while still keeping their job or being promoted. That means most decisions are pretty "lazy" and nothing is lazier than hiring someone you know.
The problem with that is you generally know people of the same gender\race\socioeconomic status. So many jobs are in reality not filled through an unbiased recruitment strategy but "hey my nephew is interested in your line of work and just graduated." "Oh yeah, send him on over we're currently hiring!"
Untested Emergency Automatic Braking systems are incredibly dangerous and require extensive validation. You slam on your brakes and a motorcyclist behind you dies.
And as of today vehicles aren't legally required to have one, so disabling the feature doesn't make your vehicle any less safe or less legal than the hundreds of millions of older vehicles on our roads without an EAB.
You mean to tell me that a relatively new customization of an open source project... still looks a lot like the original!? I'm shocked!
Lol yes. With a name like Captain Marvel, it's realllly important to not change anything lest it is mistaken for just some generic superhero. /s
I think Linus is half-right. He's right that nobody is going to want to run an AWS Arm server. He's wrong in that a lot of the world is moving to serverless. There aren't weird bugs that are platform dependent between Node.JS on a Windows, Linux or ARM server. I just upload the code to the cloud and it could be running on PowerPC for all I care.
For serverless applications, if the cloud provider can get the javascript to run the user will be none the wiser. And they can develop at home.
Arguably though I would say that if VS Code could be run in the cloud as a website connected to your cloud provider directly that would be huge. Commit to git, work in a web browser (VSCode already is a neutron app), have your dev web server be on a shared cloud instance. Just switch from production to live targets. Keep your dev environment consistent between every computer you use since it's just a web browser interface.
And all of that could be running on ARM or x86 and nobody would care. Like it or not that's a huge portion of the world's dev work.
I've developed for iOS\Android and for Windows 10. Windows 10 was SOOOooooo much easier to develop for because of exactly what linus talks about. I could quickly figure out why my weird touch gesture wasn't working by hitting "run" not by compiling, transfering, launching, remotely debugging etc.
Many other cars have a channel around the inside of the boot lid that catches the water that runs off the lid when you open it,
So does the Model 3. But it's hip to ceramic coat their entire model 3 aka put on a full body hydrophobic coating. What happens is the water runs off so fast that it shoots several inches across the gap into the trunk (boot) and across the channel.
A stock Model 3 doesn't have this issue because the water doesn't evil kenivel the gap at highspeed.
The boot (trunk in US English) lets water in when you open it after its been raining too.
Only if you do what Tesla tells you explicitly not to do: ceramic coat\rainX the rear window.
If the water runs off at a regular pace with regular adhesion the water pours into the guides. If you put a hydrophobic coating on your rear window it runs off so fast that it overshoots the water guides and into the trunk.
And if they work, liek the tunnels, they end up being about the same as those made by others.
Musk's tunnel cost $10m for a 1.4 miles. For comparison my city just built a #)!@ bike lane for $12m/mile. They could have built an exclusive right-of-way bike tunnel for less with Boring Company.
By the way I assume this is what you're reffering to for "about the same"
For a better comparison Super Excavators (the previous owners of Godot) used the exact same machine to build a 1,640 ft sewer overflow tunnel for $12.4 million, or scaling up $38 million/mile
Hardly "about the same" and this was their very first attempt which was more about fact-finding than a finished product. Not to mention the competitor using the same boring machine is using it for sewers. Even at $40m a mile vs $10m or $1m a mile that's an innovation to repurpose a small bore tunnel for transit.
I would much rather pay by the GB than pay by the mbps.
I would much rather have 10gbe to my home and be able to upload 1TB a month as fast as the network allows than meter my usage via speed caps. Just put a QoS baseline on everyone that "guarantees" at least 50mbps and then let it run free. At 2am I can then upload my cloud backups for the night and have it done by the morning.
After all home\business internet naturally has different peaks throughout the day.
Except you'll be reliving the late 2010s over and over and over.... Trump will your president forever.
You're bending over backward though to assign sane and skilled leadership skills to someone who does not have them. I will illustrate it with two different questions.
Can we get someone to Mars before the end of my first term with infinite money?
With infinite money, how soon could we get someone to mars?
They're subtly different but one question is an intelligent question that identifies bottlenecks. The other is a vanity request.
The only reason that cable companies have given for Net Neutrality is that they will increase investment in infrastructure if they have a financial incentive.
You might on some devices be typing on a screen... you do that already, it's called a smartphone.
But what this will enable is a smartphone sized device that could be unfolded or unrolled into a tablet sized device.
Further proof that conservatives just hate Tesla.
When Tesla is unprofitable (even with "inflated" prices) they say Tesla is a sham that can't make money so we shouldn't be subsidizing a failed business model.
When those "inflated" (subsidized) prices finally help build a sustainable business suddenly it's proof that they never needed subsidies and Tesla was always going to succeed anyway.
The same is said of wind power and solar. We shouldn't offer solar\wind subsidies because they're a dead end waste of money. When solar\wind finally are cost effective then suddenly those subsidies are scams because solar\wind is cost competitive.
I guess expecting people to see cause\effect relationships is asking too much. Kind of like the person who said Bush and Obama shared "equal blame" for the financial crisis... of 2007 (a year before Obama took office).