23 ms·

>> One thing that was unusual is that all the technical people there understood all aspects of the machine. Software people could talk about ASICs and CPU instr

>> One thing that was unusual is that all the technical people there understood all aspects of the machine. Software people could talk about ASICs and CPU instructions, and the hardware people understood the software stack. Every aspect of what it takes to make a computer work was represented in one building: analog hardware, chip design, motherboard design, compiler design (objective C), loader, operating system, windowing system, application layer, and applications. Where other companies had engineering teams, NeXT would have a single individual. This is in stark contrast to most of today's companies, where you have front-end engineers who don't know anything about the backend they are interacting with, backend engineers who don't care about the frontend they are serving data to, database engineers who care about neither, etc. And that's just software. The hardware might as well be a black box for the vast majority of software engineers working at your average software company today. Where did we go so wrong?

tclancy 6y ago
I don't think we did; there's a bit of selection bias at play for NeXT where they had a large network of stars to choose from and a reputation that would attract a large pool of other people worth choosing. It's just that a lot more people are involved in the industry now so there's a lot more entry-level/ grunt work to be had.
samhuk 6y ago
I think you are misunderstanding. Do you think that engineers designing the plumbing system of the F1 rocket engine knew pretty much anything about "compiler design" or "motherboard design"? What you are in fact observing is a human system's tendency to adapt to growing complexity. Human systems adapt to growing complexity by specializing it's members to particular skills (see https://en.wikipedia.org/wiki/Cognitive_specialization https://en.wikipedia.org/wiki/Cognitive_specialization). It's just a fact that way back then in the NeXT days, the computer was not as complex, the understanding of which was just about achievable by 1 human. Eventually, that metric exceeds 1. For the case of the first apple computer, it was ~1 (Wozniak), in the case of Apollo, it was >100. Cognitive Specialization is an aspect of all living things, but is very apparent in the Human species. Computers got more complex in just the same fashion as how farming went from a farmer, a bull, and a blacksmith, to gigantic conglomerates to make the fertilizer, tractors, watering system, etc., that comprise of >10,000 humans.
randomdata 6y ago
> Where did we go so wrong? By getting it right, which saw an explosion in the market, leaving us in a position now where there is more work to be done than the people who understand systems from top to bottom can handle alone. If those pioneering efforts had failed, the tech industry would now be insignificant and those superstars would be struggling to find work, never mind those who have a lesser understanding/care.
TheOtherHobbes 6y ago
Partly ad tech. It changed the reward profile away from inventing cool new stuff towards lowest-common-denominator monetisation. The real change was the change in the culture of computer use from original creation to consumption and distribution of certain limited kinds of creation - which are mostly imitative, nostalgic, and either backwards- or (at best) sideways-looking rather than genuinely original. Real invention is now actively disfavoured. Google did a fair amount in the 00s but has slowly abandoned most of it, Amazon does a bit of blue sky but is mostly focussed on consumerism, Musk's idea of blue sky is straight out of a 1950s Tom Swift novel, and Facebook and Twitter are both hopeless. Netflix is cable TV done right - finally. But it's still cable TV. There are some exceptions at Apple, which still has a kind of legacy tradition of doing cool new stuff (see also, M1) but even that is a mix of invention for the sake of it and strategic lock-in as a goal. The result is a landscape full of development geared to comfortable suburban consumerism and associated corporate bureaucracy. There's very little interest in game changing technical development for the sake of it - which was more or less what NeXT was about. And there's even less interest in computing as subversion and empowerment, which was - believe it or not - a big interest in the 70s. FOSS doesn't change this. (It likes to believe it does, but practically it really doesn't.) Quantum computing and AI may be on the cusp - but even if they do something interesting they're going to be coopted by ad tech as soon as the paint dries. So it's not about technical scope so much as imagination failure. The real loss is the loss of imagination - something that tech and media have both done a lot of damage to over the last couple of decades.
uxp100 6y ago
I think part of the problem is you comparing to a fairly different industry. Companies that employ "Front End" and "Back end" engineers are very different than NeXT, which had products from asic to high level. If you work at a hardware company, even in a role that's very far from hardware, you become aware of this stuff because it effects you. Even if you don't understand ASIC at all, you still know about, oh, such and such process node has this issue, because it impacted our schedule and somebody told me about it at a lunch table. And you may not even know how to solder, but you can say, oh this needs rework, 0 ohm resistor at point such and such to make the display work, because you need to know it to go to the lab and have the work done. Really the thing I understood the least working at a company like that was actual productization. You hear about something you worked on a year and a half ago being a tablet, or an embedded device, and you'd be like, oh, that old thing is just being released now? And it's in that form factor? huh.
triceratops 6y ago
Specialization is a natural response to increasing complexity.

tclancy 6y ago

I don't think we did; there's a bit of selection bias at play for NeXT where they had a large network of stars to choose from and a reputation that would attract a large pool of other people worth choosing. It's just that a lot more people are involved in the industry now so there's a lot more entry-level/ grunt work to be had.

samhuk 6y ago

I think you are misunderstanding. Do you think that engineers designing the plumbing system of the F1 rocket engine knew pretty much anything about "compiler design" or "motherboard design"? What you are in fact observing is a human system's tendency to adapt to growing complexity. Human systems adapt to growing complexity by specializing it's members to particular skills (see https://en.wikipedia.org/wiki/Cognitive_specialization https://en.wikipedia.org/wiki/Cognitive_specialization). It's just a fact that way back then in the NeXT days, the computer was not as complex, the understanding of which was just about achievable by 1 human. Eventually, that metric exceeds 1. For the case of the first apple computer, it was ~1 (Wozniak), in the case of Apollo, it was >100. Cognitive Specialization is an aspect of all living things, but is very apparent in the Human species. Computers got more complex in just the same fashion as how farming went from a farmer, a bull, and a blacksmith, to gigantic conglomerates to make the fertilizer, tractors, watering system, etc., that comprise of >10,000 humans.

randomdata 6y ago

> Where did we go so wrong? By getting it right, which saw an explosion in the market, leaving us in a position now where there is more work to be done than the people who understand systems from top to bottom can handle alone. If those pioneering efforts had failed, the tech industry would now be insignificant and those superstars would be struggling to find work, never mind those who have a lesser understanding/care.

TheOtherHobbes 6y ago

Partly ad tech. It changed the reward profile away from inventing cool new stuff towards lowest-common-denominator monetisation. The real change was the change in the culture of computer use from original creation to consumption and distribution of certain limited kinds of creation - which are mostly imitative, nostalgic, and either backwards- or (at best) sideways-looking rather than genuinely original. Real invention is now actively disfavoured. Google did a fair amount in the 00s but has slowly abandoned most of it, Amazon does a bit of blue sky but is mostly focussed on consumerism, Musk's idea of blue sky is straight out of a 1950s Tom Swift novel, and Facebook and Twitter are both hopeless. Netflix is cable TV done right - finally. But it's still cable TV. There are some exceptions at Apple, which still has a kind of legacy tradition of doing cool new stuff (see also, M1) but even that is a mix of invention for the sake of it and strategic lock-in as a goal. The result is a landscape full of development geared to comfortable suburban consumerism and associated corporate bureaucracy. There's very little interest in game changing technical development for the sake of it - which was more or less what NeXT was about. And there's even less interest in computing as subversion and empowerment, which was - believe it or not - a big interest in the 70s. FOSS doesn't change this. (It likes to believe it does, but practically it really doesn't.) Quantum computing and AI may be on the cusp - but even if they do something interesting they're going to be coopted by ad tech as soon as the paint dries. So it's not about technical scope so much as imagination failure. The real loss is the loss of imagination - something that tech and media have both done a lot of damage to over the last couple of decades.

uxp100 6y ago

I think part of the problem is you comparing to a fairly different industry. Companies that employ "Front End" and "Back end" engineers are very different than NeXT, which had products from asic to high level. If you work at a hardware company, even in a role that's very far from hardware, you become aware of this stuff because it effects you. Even if you don't understand ASIC at all, you still know about, oh, such and such process node has this issue, because it impacted our schedule and somebody told me about it at a lunch table. And you may not even know how to solder, but you can say, oh this needs rework, 0 ohm resistor at point such and such to make the display work, because you need to know it to go to the lab and have the work done. Really the thing I understood the least working at a company like that was actual productization. You hear about something you worked on a year and a half ago being a tablet, or an embedded device, and you'd be like, oh, that old thing is just being released now? And it's in that form factor? huh.

triceratops 6y ago

Specialization is a natural response to increasing complexity.