4 ms·

And in the year 1990 you couldn’t buy a supercomputer that fit in your pocket for $1000 either. Today everyone has one. What exactly makes this case different?

xxs 5y ago
Dennard scaling/Moore's law. We pretty knew back then it'd happen. Most of the progress has been iterative but the basic process/lithography is still very similar. There is no indication or massive breakthrough in material science that would allow anything similar to what we have been seeing in the electronics development for the past 50 years.
- yazaddaruvala 5y ago
  > There is no indication or massive breakthrough in material science that would allow anything similar Then you may not be paying attention. Battery costs have been dropping exponentially and we are only starting to really invest capital and scale production. Meanwhile: 1. Solid state batteries, 2. 3D printed metal, 3. Algorithmically optimized design. Before 2018 battery production was basically negligible. We have only truly started to scale our need for storage in 2018. Today 3D printed metal is in its infancy, but has truly significant potential. I’m very excited to see how companies like Relativity Space, etc mature this tech and how it will disrupt current manufacturing. Even just giga molding is disruptive, let alone printing which will use significantly less metal and significantly reduce capital investment for manufacturing (i.e. something like AWS for manufacturing).
Ekaros 5y ago
You can find pretty cheap remote controlled electric cars today. Even for less than 50$... Though size might be somewhat issue. But hey aren't those in pocket also lot smaller?
flyinglizard 5y ago
Because progress is mostly done on value and not absolute costs. An IBM XT PC cost $5k in present day dollars back in 1985. Computers today are exponentially better but maintain a modest linear price adjustment. Going back to your post verbatim, super computers today still cost hundreds of millions and need an entire building; the definition of what makes a computer “super” is what evolved, not their cost or other burdens.
- yazaddaruvala 5y ago
  > Because progress is mostly done on value and not absolute costs. Oh.. am I reading you correctly? You don’t see the value in creating $1,000 cars that have subscription packages for self-driving, heated seats, navigation, etc? > the definition of what makes a computer “super” is what evolved, not their cost or other burdens. I feel like you’re intentionally deflecting. Clearly my point was that “we get significantly more, for significantly less”. There is no reason to assume we won’t do the same with cars, or all manufacturing with time. I’ve laid out some points why/how in a different response.
  - flyinglizard 5y ago
    No, you are not reading me correctly. I’m not saying that $1000 cars wouldn’t be great, I’m just saying they wouldn’t be possible. Any technological advancement goes into making better cars and not cheaper cars.
    - reginold 5y ago
      Here's a $930 EV on Alibaba, reviewed by Jalopnik. Seats 4: https://jalopnik.com/heres-what-the-worlds-cheapest-electric-car-is-like-to-1843904305 https://jalopnik.com/heres-what-the-worlds-cheapest-electric...

xxs 5y ago

Dennard scaling/Moore's law. We pretty knew back then it'd happen. Most of the progress has been iterative but the basic process/lithography is still very similar. There is no indication or massive breakthrough in material science that would allow anything similar to what we have been seeing in the electronics development for the past 50 years.

yazaddaruvala 5y ago
> There is no indication or massive breakthrough in material science that would allow anything similar Then you may not be paying attention. Battery costs have been dropping exponentially and we are only starting to really invest capital and scale production. Meanwhile: 1. Solid state batteries, 2. 3D printed metal, 3. Algorithmically optimized design. Before 2018 battery production was basically negligible. We have only truly started to scale our need for storage in 2018. Today 3D printed metal is in its infancy, but has truly significant potential. I’m very excited to see how companies like Relativity Space, etc mature this tech and how it will disrupt current manufacturing. Even just giga molding is disruptive, let alone printing which will use significantly less metal and significantly reduce capital investment for manufacturing (i.e. something like AWS for manufacturing).

Ekaros 5y ago

You can find pretty cheap remote controlled electric cars today. Even for less than 50$... Though size might be somewhat issue. But hey aren't those in pocket also lot smaller?

flyinglizard 5y ago

Because progress is mostly done on value and not absolute costs. An IBM XT PC cost $5k in present day dollars back in 1985. Computers today are exponentially better but maintain a modest linear price adjustment. Going back to your post verbatim, super computers today still cost hundreds of millions and need an entire building; the definition of what makes a computer “super” is what evolved, not their cost or other burdens.

yazaddaruvala 5y ago
> Because progress is mostly done on value and not absolute costs. Oh.. am I reading you correctly? You don’t see the value in creating $1,000 cars that have subscription packages for self-driving, heated seats, navigation, etc? > the definition of what makes a computer “super” is what evolved, not their cost or other burdens. I feel like you’re intentionally deflecting. Clearly my point was that “we get significantly more, for significantly less”. There is no reason to assume we won’t do the same with cars, or all manufacturing with time. I’ve laid out some points why/how in a different response.
- flyinglizard 5y ago
  No, you are not reading me correctly. I’m not saying that $1000 cars wouldn’t be great, I’m just saying they wouldn’t be possible. Any technological advancement goes into making better cars and not cheaper cars.
  - reginold 5y ago
    Here's a $930 EV on Alibaba, reviewed by Jalopnik. Seats 4: https://jalopnik.com/heres-what-the-worlds-cheapest-electric-car-is-like-to-1843904305 https://jalopnik.com/heres-what-the-worlds-cheapest-electric...