5 ms·

I linked a video below of Jensen Huang founder of NVIDIA speaking at Stanford. He says at one point in the 90s there were 100s of companies in their space and h

I linked a video below of Jensen Huang founder of NVIDIA speaking at Stanford. He says at one point in the 90s there were 100s of companies in their space and he goes on to list a series of unique insights/perspectives they had that led to become the last one standing. NVIDIA understood the insights that mattered. I would say a more modern example of this the insight that to build the best mobile phone product, you need to design the hardware, operating system and APIs (and thus first party apps) together as a single thing. Apple brought this insight from their past experience building less sophisticated computers. They knew a phone given it's size and use case would require this to an even _greater_ degree. Google on the other hand, based on their experience of working in open ecosystems and linux machines from server to desktop, thought the phone market would extend from an open OS platform. This was their perspective/insight. Ultimately, the company that won this argument was Apple. We know this because Google has finally adopted this truth that Apple knew all along with the Pixel hardware lineup. Furthermore, the Android phones that succeeded today have nothing to do with the fact they are "Android". Samsung phones are about the Samsung ecosystem. Choosing Android is more important to the OEM than the user because of the app ecosystem that cannot be bootstrapped from scratch any more. Android is the developer/oem platform, not the user platform. Now, think about how long it took Google to finally give in to that insight? These perspectives lead to very stubborn decision making. This perspective was often a debatable thing for a long time. It's no longer. Hence I think everything you said here is right and aligned what what we see in one of the biggest product categories on the planet right now.

robocat 4y ago
https://www.youtube.com/watch?v=Xn1EsFe7snQ https://www.youtube.com/watch?v=Xn1EsFe7snQ Amazing video thank you. From ~2009 that keeps hammering the unique perspective topic, especially during formation in 1993 and change of focus in 1998. I would be interested to know which of the failed hundreds of competitors had the same “unique perspective”, but maybe just failed because another leg of the company was too weak. Successful companies have a few core perspectives that they nail with A grades, but they have to get a C grade in all other areas of the business. It can take only one F grade in one topic to make a business fail. The meta-skill of recognising and fixing a critical failing is hard to learn. Edit: I love Jen-Hsun’s definition of a startup starting about 57:50: a startup is an enterprise that is nearly out of business all the time.
- enos_feedler 4y ago
  Yes it would be interesting. 3DFX was the big 3D chip company at the time when NVIDIA was still in fight mode. I know this because I remember wanting a 3DFX Voodoo 2 card so badly. I just checked wiki and it sounds like 3DFX died due to mismanagement. In the chip industry missing timing or delivery of a product cycle or two at that stage would be deadly. So in the case of 3DFX the execution leg might just have been weak. Would be interesting to go back and analyze all of these company cases though.
  - robocat 4y ago
    Jen-Hsun interviews Morris Chang https://youtube.com/watch?v=u-x7PdnvCyI https://youtube.com/watch?v=u-x7PdnvCyI Long interview, but I think it has moments where it reinforces the same concept that it is very difficult to copy a business, even if the CEO is directly telling you their secrets. Interesting interview - although exponentially above my pay grade.
sangnoir 4y ago
> Ultimately, the company that won this argument was Apple. We know this because Google has finally adopted this truth that Apple knew all along with the Pixel hardware lineup. I'm not convinced that Pixel hardware is anything other than reference designs: each device is developed from a clean slate and none sell in appreciable numbers > Samsung phones are about the Samsung ecosystem. What Samsung ecosystem? Samsung makes fantastic stand-alone hardware, but it's no ecosystem. No one buys Samsung earbuds because they work particularly well with the Galaxy 22, or their calendar and music are already integrated with Bixby
- akiselev 4y ago
  Does Google make the design files for the Pixel phones available to Android partners under NDA? The value of a reference design is basically zero without the schematics/layout and BSP. Or do you just mean upstreaming BSP drivers?
  - nine_k 4y ago
    Schematics can be reverse-engineered in a few days, I suspect; there's no point to hide it. The interesting part is the choice of components, the rationale behind that choice, and the software support. I suppose that Google makes available the Android OS with support for all the new capabilities as they emerge (OLED, NFC, fingerprint scanners, etc) which OEMs can leverage with their hardware and closed-source software (both apps and drivers). The Pixel line is the Google's vision of the platform, with Google's control over hardware implementation and a commitment to support it (longer than many OEMs but not nearly as great as iPhones).
    - akiselev 4y ago
      “Reference design” is a term of art in electrical engineering, not aspirational product marketing. Reverse engineering 10-20 layer mixed signal PCBs with buried microvias is nontrivial and is the exact opposite of what reference designs are supposed to be.
      - 4y ago

robocat 4y ago

https://www.youtube.com/watch?v=Xn1EsFe7snQ https://www.youtube.com/watch?v=Xn1EsFe7snQ Amazing video thank you. From ~2009 that keeps hammering the unique perspective topic, especially during formation in 1993 and change of focus in 1998. I would be interested to know which of the failed hundreds of competitors had the same “unique perspective”, but maybe just failed because another leg of the company was too weak. Successful companies have a few core perspectives that they nail with A grades, but they have to get a C grade in all other areas of the business. It can take only one F grade in one topic to make a business fail. The meta-skill of recognising and fixing a critical failing is hard to learn. Edit: I love Jen-Hsun’s definition of a startup starting about 57:50: a startup is an enterprise that is nearly out of business all the time.

enos_feedler 4y ago
Yes it would be interesting. 3DFX was the big 3D chip company at the time when NVIDIA was still in fight mode. I know this because I remember wanting a 3DFX Voodoo 2 card so badly. I just checked wiki and it sounds like 3DFX died due to mismanagement. In the chip industry missing timing or delivery of a product cycle or two at that stage would be deadly. So in the case of 3DFX the execution leg might just have been weak. Would be interesting to go back and analyze all of these company cases though.
- robocat 4y ago
  Jen-Hsun interviews Morris Chang https://youtube.com/watch?v=u-x7PdnvCyI https://youtube.com/watch?v=u-x7PdnvCyI Long interview, but I think it has moments where it reinforces the same concept that it is very difficult to copy a business, even if the CEO is directly telling you their secrets. Interesting interview - although exponentially above my pay grade.

sangnoir 4y ago

> Ultimately, the company that won this argument was Apple. We know this because Google has finally adopted this truth that Apple knew all along with the Pixel hardware lineup. I'm not convinced that Pixel hardware is anything other than reference designs: each device is developed from a clean slate and none sell in appreciable numbers > Samsung phones are about the Samsung ecosystem. What Samsung ecosystem? Samsung makes fantastic stand-alone hardware, but it's no ecosystem. No one buys Samsung earbuds because they work particularly well with the Galaxy 22, or their calendar and music are already integrated with Bixby

akiselev 4y ago
Does Google make the design files for the Pixel phones available to Android partners under NDA? The value of a reference design is basically zero without the schematics/layout and BSP. Or do you just mean upstreaming BSP drivers?
- nine_k 4y ago
  Schematics can be reverse-engineered in a few days, I suspect; there's no point to hide it. The interesting part is the choice of components, the rationale behind that choice, and the software support. I suppose that Google makes available the Android OS with support for all the new capabilities as they emerge (OLED, NFC, fingerprint scanners, etc) which OEMs can leverage with their hardware and closed-source software (both apps and drivers). The Pixel line is the Google's vision of the platform, with Google's control over hardware implementation and a commitment to support it (longer than many OEMs but not nearly as great as iPhones).
  - akiselev 4y ago
    “Reference design” is a term of art in electrical engineering, not aspirational product marketing. Reverse engineering 10-20 layer mixed signal PCBs with buried microvias is nontrivial and is the exact opposite of what reference designs are supposed to be.
    - 4y ago