I love this article
Build Unix, Not Uber
51–60 of 117 posts
Re: Build Unix, Not Uber
#52Earlier quoted context omitted.
> Things like GPS and the internet would’ve been deployed much faster by a private corporation but then they would be dumb monopolies as opposed to the public good they are now. 1) at least in the USA, access to the internet is not a public good, it is provided by an oligopoly (and in many circumstances a functional monopoly, where service at a given address is only available from a single ISP) 2) why do you believe…
I would compare GPS to something like Starlink which went from idea to activation in 5 years while GPS took over 15. While of course, Starlink relied on government know-how, given away for free, the speed of deployment was 2-3x. In the end we’re left with a monopoly service which sucks. But, it came together quick. Why can’t we have both? Fast deploying public goods.
Isn’t that more to do with when GPS was deployed. Doing things early is harder and takes longer, even more so when you’re first. Comparatively Starlink has fewer novel problems to solve.
If you’re going to compare the two you need to control for confounding factors.
Re: Build Unix, Not Uber
#53It's always interesting to me how little credit is given to GPS in Uber's story. The existence of space-based, free and ubiquitous semi-precise navigation was an enormous subsidy.
I want something that doesn’t operate on scarcity (like the corporation), but can work at the velocity of a corporation. Things like GPS and the internet would’ve been deployed much faster by a private corporation but then they would be dumb monopolies as opposed to the public good they are now. Also we need a space to work on big dumb ideas and toys[1] and neither Unix, Uber or GPS fit in that category. What is that…
I think the answer is corporations — but with a different culture.
Social purpose corporation with crowdfunding and transparency seems like it would accomplish much of what you want.
Re: Build Unix, Not Uber
#54It's always interesting to me how little credit is given to GPS in Uber's story. The existence of space-based, free and ubiquitous semi-precise navigation was an enormous subsidy.
I want something that doesn’t operate on scarcity (like the corporation), but can work at the velocity of a corporation. Things like GPS and the internet would’ve been deployed much faster by a private corporation but then they would be dumb monopolies as opposed to the public good they are now. Also we need a space to work on big dumb ideas and toys[1] and neither Unix, Uber or GPS fit in that category. What is that…
The Internet already got out so fast it beat the Big Government Solution, which was OSI, so comprehensively that people these days don't even know that there was an OSI protocol suite, let alone that it got beaten by the TCP/IP stack. Back in the early days, the ARPANET stuff was seen as a sideshow, whereas OSI was going to be the Serious Network for Serious Users, with X.500 directory services supporting X.400 email and nobody would remember toys like LDAP and SMTP and POP.
Of course, these days the idea that TCP/IP somehow implements a purely conceptual OSI is so ingrained that people get angry when you tell them the accurate history. So it goes.
Re: Build Unix, Not Uber
#55The concentration of big tech has become so intense it feels like a terrible defeat of the free software principles espoused in this post. Freedom is the main casualty of the current computing environment, which consists of a phone, which legally prohibits its owner from controlling it, and a cloud which is always coaxing programmers to its most proprietary solutions.
Re: Build Unix, Not Uber
#56Earlier quoted context omitted.
There are a lot of alternative programming languages that those companies could have used. There's no GPS alternative for Uber.
Alt to php at that point in time was usually Perl. Don’t think Facebook would have succeeded with Perl.
What about Perl would have made Facebook fail? I don't particularly like Perl but I'm curious.
Re: Build Unix, Not Uber
#57Rather convolution article. First mentioning economic scarcity then moving onto 'good ideas' like Unix. I'm missing the central thread that ties it all together.
If anything it was the antithesis of Unix, namely GNU with its specific license that protects the users freedom, that was really the revolutionary idea, thank you RMS. Of course Unix, the work by Thompson, Ritchie etc brought extensibility and endless creativity to the development/app space. Ie write programs that do one thing well and able to connect those through pipes and ultimately through networks. The cli interface is ancient, but its here to stay for a very long time indeed.
Not sure of the point that my comment is making. Meh. Happy Friday.
Re: Build Unix, Not Uber
#58> because companies are only ultimately beholden to customers.
Excuse the narcissism, but MBAs disagree (shareholders) and Wall Street disagrees (profits) and activist employees disagrees
Re: Build Unix, Not Uber
#59Earlier quoted context omitted.
I would compare GPS to something like Starlink which went from idea to activation in 5 years while GPS took over 15. While of course, Starlink relied on government know-how, given away for free, the speed of deployment was 2-3x. In the end we’re left with a monopoly service which sucks. But, it came together quick. Why can’t we have both? Fast deploying public goods.
Isn’t that more to do with when GPS was deployed. Doing things early is harder and takes longer, even more so when you’re first. Comparatively Starlink has fewer novel problems to solve. If you’re going to compare the two you need to control for confounding factors.
Sure but they had full
access to 20 years of launch technology from NASA at that point.
More generally, I don’t think “Government works slower then private corporations” is that controversial a conclusion. If we’re gonna try to improve it, we have to admit the problem first.
Re: Build Unix, Not Uber
#60> UNIX was a good idea. It was free to spread widely across the tech ecosystem, and eventually some corporate players like Apple and Sun captured a lot of the value, but Linux stands out as the UNIX “winner” in 2022. Neither the core ideas of UNIX-style operating systems nor the Linux open-source project grew solely because of corporate stewardship. They spread and won because they were good ideas.
The author neglects to mention that before POSIX[0], the "core ideas of UNIX-style operating systems" were very much proprietary. For example, there was no easy way to get arbitrary System V[1] or BSD source to easily compile on the other, let alone lesser known Unix variants.
So I posit that the ideas spread not due to the merit of each version of *nix, but instead because a standard which allowed (reasonable) portability was agreed upon and largely supported.