Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Build Unix, Not Uber

thesephist.com

Post reply on HN
Print view View original

71–80 of 117 posts

Re: Build Unix, Not Uber

#71

Post by throwaway14356 »

abathur wrote: post #5

At the risk of saying something dumb and meta... it's ideas all the way down (for a really long time now). The individual corporations are ideas, and then of course there's the idea of the corporation itself. As the post notes, corporations only work because of other ideas about property rights and scarcity. But more fundamentally they only work because they're an idea quite a few people already believe in. I'm a lit…

i applaud the simplicity

>can we better identify and allocate resources towards good ideas that don't make market sense?

no, but we shall die trying

Re: Build Unix, Not Uber

#72

Post by p33p »

m_mueller wrote: post #69
Animats wrote: post #25

An alternative to the corporation is the chaord. This is an idea that went nowhere. Except that Visa International was a chaord. A chaord is useful when you have a shared resource that's needed by competing entities. That's exactly what Visa International was set up to be. It's an operator of a data network and a standards body. It doesn't issue cards. It doesn't handle money. It just transmits transaction data betwe…

There’s also the cooperative. Here in Switzerland they are rather popular - some of the biggest banks, insurances and property owners, and more notably, the two largest supermarket chains (making up >80% of food / non food market) are organised this way. Your customers could be part of the ownership structure of your company, which creates a strong loyalty effect and a focus on organic & vertical growth rather than g…

These are called mutual companies in the US — at least insurance — New York Life, Northwestern Mutual, MassMutual, etc

Re: Build Unix, Not Uber

#73

Post by m_mueller »

p33p wrote: post #72
m_mueller wrote: post #69

Earlier quoted context omitted.

There’s also the cooperative. Here in Switzerland they are rather popular - some of the biggest banks, insurances and property owners, and more notably, the two largest supermarket chains (making up >80% of food / non food market) are organised this way. Your customers could be part of the ownership structure of your company, which creates a strong loyalty effect and a focus on organic & vertical growth rather than g…

These are called mutual companies in the US — at least insurance — New York Life, Northwestern Mutual, MassMutual, etc

I‘m not aware of any software company organised this way though. Wouldn’t it make sense for a company like Red Hat or Cannonical - FOSS based service companies?

Re: Build Unix, Not Uber

#74

Post by kortilla »

sklargh wrote: post #2

It'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.

That’s not a subsidy. A use of a non-rivalrous good that would exist with or without Uber is not a subsidy.

That’s like saying the English language and calculus are a subsidy. They are definitely shoulders we all stand on, but not a subsidy.

Re: Build Unix, Not Uber

#75

Post by kortilla »

meheleventyone wrote: post #52
pj_mukh wrote: post #50

Earlier 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.

>Comparatively Starlink has fewer novel problems to solve.

Lol, this is worse than the infamous Dropbox comment. There are no LEO providers except for Starlink for a reason. Starlink required novel launch techniques, novel antennas, novel network management, and that’s just what we know about publicly.

Re: Build Unix, Not Uber

#76

Post by niij »

bitwize wrote: post #40
Animats wrote: post #25

An alternative to the corporation is the chaord. This is an idea that went nowhere. Except that Visa International was a chaord. A chaord is useful when you have a shared resource that's needed by competing entities. That's exactly what Visa International was set up to be. It's an operator of a data network and a standards body. It doesn't issue cards. It doesn't handle money. It just transmits transaction data betwe…

Other examples: UL, arguably AT&T before the breakup, the Queensland public bus system which does privatization "right": the government establishes common routes and a common fare system, private carriers drive the routes and compete on quality and price. Chaordic entities are a good choice for government administration of public services without being too heavy-handed with state intervention.

Bangkok, Thailand has a similar bus system for many routes.

Re: Build Unix, Not Uber

#77

Post by wizofaus »

dylan604 wrote: post #29
foobarbaz33 wrote: post #22

Earlier quoted context omitted.

I don't think the programming language dependencies are comparable to tech like GPS. Reddit was originally written in lisp. It was later rewritten in Python with a trivial amount of effort. And it could have be trivially rewritten in any other language as well. The languages are not that important. Having access to high level languages in general is important, but they are a dime a dozen.

could you imagine the expense of a FAANG having to pay Microsoft Windows licensing for every machine they use? how would that affect a startup's runway? so instead of the languages, maybe open it up to Linux/FOSS in general

Startups I've worked at have struggled with various expenses but OS licenses weren't one of them. OTOH finding devs 100% comfortable working on pure linux machines would be a genuine blocker.

Re: Build Unix, Not Uber

#78

Post by krmboya »

beej71 wrote: post #33
EddieDante wrote: post #15

Earlier quoted context omitted.

GP should be grateful we didn't get stuck with Multics, ITS, or TOPS20 instead.

No kidding. The playing field wasn't super great, and Unix was and is still a top contender even today. CP/M. OS/2. MS-DOS. OS/2 Warp. Windows. Windows NT. Mac OS. VMS...

Yeah, maybe we optimistically speculate that Smalltalk or Lisp Machines would have won instead, but it may have been the poorer alternatives that would have won if not UNIX

Re: Build Unix, Not Uber

#79

Post by lokedhs »

m_mueller wrote: post #73
p33p wrote: post #72

Earlier quoted context omitted.

These are called mutual companies in the US — at least insurance — New York Life, Northwestern Mutual, MassMutual, etc

I‘m not aware of any software company organised this way though. Wouldn’t it make sense for a company like Red Hat or Cannonical - FOSS based service companies?

I think Dyalog was at least created that way. I don't know if they are an actual cooperative though.

The company itself was created by their customers to ensure support and development of the product that they used.

Re: Build Unix, Not Uber

#80

Post by shreyshnaccount »

Animats wrote: post #25

An alternative to the corporation is the chaord. This is an idea that went nowhere. Except that Visa International was a chaord. A chaord is useful when you have a shared resource that's needed by competing entities. That's exactly what Visa International was set up to be. It's an operator of a data network and a standards body. It doesn't issue cards. It doesn't handle money. It just transmits transaction data betwe…

another model is a highly autonomous company, co-owned by the central govt and the Delhi govt. for the Delhi metro.
Post reply on HN