18 ms·

Software had the promise of massively decentralizing business; anyone with a computer and a bit of knowledge could start hacking away at a new product. It is a

Software had the promise of massively decentralizing business; anyone with a computer and a bit of knowledge could start hacking away at a new product. It is a bit sad to see big tech becoming a way of even more centralization of power. I guess it had the potential for either; individuals being able to easily capitalize on it, or businesses being able to replicate infinitely and have huge margins from very little input capital. But maybe both are happening at the same time. A double edged sword, like so many other things.

dredmorbius 5y ago
That's the myth we were sold. It's the myth many of us believed. I certainly did. In truth, greater efficiencies don't decentralise, they centralise. The market area expands. Suppliers or vendors, rather than competing locally, compete regionally or globally. In winner-take-all markets, there is only one winner, worldwide. There are some twists on this, having to do with regional customs, language, and law, mostly. There's also the time element, where tech monopolies tend to be ascendent, then (often but not always) replaced. Still, a technological advantage can last a very long time, as with IBM and AT&T. Two of today's FAANG date from the 1970s, another two from the 1990s. Dominance can be long-lived. There were some who saw this coming. Andrew Shapiro's The Control Revolution (1999) kinda-mostly saw this coming, and certainly called out the major aspects. Shoshana Zuboff recognised at least some of the potential in the 1970s and 1980s. The crowd still believes the lie, for the most part.
- api 5y ago
  In biology we understand fairly well that environmental complexity leads to greater biodiversity. If the Earth's environment were entirely uniform there would probably be only a few species and nothing would ever change. Global highly efficient markets are like a polished flat globe with a uniform climate everywhere. Evolution would produce only one or two species in permanent equilibrium.
  - TeMPOraL 5y ago
    The more I think of it, the more I realize, biology is very much unlike the market in such analogies. In a way, the market is biology flipped on its side, like a transposed matrix - where in nature, life forms mostly cooperate within their species (or at least ignore each other) and prey on other species, on the market, companies of the same type fight it out, while ignoring or cooperating with companies of different types. But I think the main difference is scale. The area of influence of a life form - where they live, hunt, reproduce - is very limited. Same is true with companies serving local markets. A local glass factory. A corner market. You say: > Global highly efficient markets are like a polished flat globe with a uniform climate everywhere. But it's not just a flat globe. It's a globe with a surface of couple dozen square kilometers! It's literally a single niche. Not even that - it's a cage. Imagine a zoo that decided to put all its animals in a single, large cage. You won't have to wait much before there's only one big animal standing (and a bunch of insects scattering about). On a more general note: biology is a good example that you can make a complex and robust system by giving imperfectly-self-reproducing agents a lot of space to play in, and then throwing calamities at them to keep things interesting. But I honestly don't like using it as a recipe for designing systems, because of all the waste that's involved. Waste of effort, waste of resources, waste of life. I feel we should avoid this kind of hard self-regulation whenever we can.
    - shkkmo 5y ago
      > where in nature, life forms mostly cooperate within their species (or at least ignore each other) This is blatantly not true. There are condiserable levels of intra-species competition in many species. Even in "cooperative" species, those animals cooperate in groups, but then those groups compete against each other. Most predators prey on a very limited subset of the ecology. You seemly cannot reliably predict inter vs intra species like you seem to think. Also "preying on" is pretty different that "competeing with" and at the ecological scale, predation looks much more like interactions between industries than competition within an industry. Thus the correct mapping of the analogy in this case is between ecological niches and industries. Regarding your more general note: I think there is a trade off between efficiency and robustness. Top down heirachical structures (most companies, militaries, etc) are generally more efficient in terms of resources used but are vulnerable to broken/inefficient/disengenous nodes, especially for the higher nodes. Bottom up solutions (markets, grassroots movements, etc) tend to use more total resources but do better at dealing with bad nodes and are thus less failure prone. I would argue that the best systems tend to layer combinations of these two types of systems. Edit: In the US for example, we (simplistically) have a bottom up democracy that governs a top down government that regulates bottom up markets that are participated in by top down companies that increasingly rely on outsourcing core functionality to other bottom up markets that that top down particpants.
coliveira 5y ago
People didn't think well about the other side of the coin: if anyone can hack away a product, then what you need to make a lot of money is somehow gather the best minds that can make this machine work for you and monopolize their ideas. This is exactly the strategy behind the FAANG.
api 5y ago
Software doesn't favor decentralization at all... at least not the current software and Internet ecosystem. The complexity of software coupled with its absolute rigidity makes it hard to build systems that talk to each other even when specifications are public. When things are proprietary take the difficulty of making software interoperable and add a total lack of documentation and a possibly adversarial relationship with the vendor where they'll try to deliberately shift things to keep you out. (There could also be legal barriers, but I'm just talking about technical ones.) Then consider the added difficulty of opening systems to interoperability with other systems while maintaining security. If you allow connections from untrusted systems, now you have to be particularly meticulous in the design of your software to ensure there are no exploitable bugs, undefined behaviors, or holes in your permission system. We still mostly program in language that are "insecure by default" and make this hard. Then there's the Internet itself which is intrinsically hierarchical. DNS, BGP, and the hierarchy enforced by NATs and firewalls that separate endpoints from the "real" Internet all make the network favor a "feudal" client-server model. We shouldn't be surprised that the Internet produces a lot of monopolies and closed silos; software built around the Internet takes on its shape. Last but not least there's the use of cryptography to enforce walled gardens: certificates, code signature requirements, closed app stores, etc. This adds an artificial barrier on top of everything I mentioned above. Compare this with older mechanical and analog systems. Even complex mechanical and analog electronic systems are straightforward to examine, measure, characterize, and understand. Ignoring patents and other legal restrictions it's straightforward to build replacement parts or interoperate with physical systems.
mikechen233 5y ago
It depends on the product. But software has stronger the ecosystem effect and data lockin more so than any previous technology. Think facebook, youtube, Windows, office, apple ecosystem of products. They have two market characteristics: 1. People need to be on the platform that others are in 2. Once you are in you it's hard to leave. Both creates tendency for winner takes all economics. And I think many people knows this. Part of the reason why software companies since the 90s have higher valuation than other industries.

dredmorbius 5y ago

That's the myth we were sold. It's the myth many of us believed. I certainly did. In truth, greater efficiencies don't decentralise, they centralise. The market area expands. Suppliers or vendors, rather than competing locally, compete regionally or globally. In winner-take-all markets, there is only one winner, worldwide. There are some twists on this, having to do with regional customs, language, and law, mostly. There's also the time element, where tech monopolies tend to be ascendent, then (often but not always) replaced. Still, a technological advantage can last a very long time, as with IBM and AT&T. Two of today's FAANG date from the 1970s, another two from the 1990s. Dominance can be long-lived. There were some who saw this coming. Andrew Shapiro's The Control Revolution (1999) kinda-mostly saw this coming, and certainly called out the major aspects. Shoshana Zuboff recognised at least some of the potential in the 1970s and 1980s. The crowd still believes the lie, for the most part.

api 5y ago
In biology we understand fairly well that environmental complexity leads to greater biodiversity. If the Earth's environment were entirely uniform there would probably be only a few species and nothing would ever change. Global highly efficient markets are like a polished flat globe with a uniform climate everywhere. Evolution would produce only one or two species in permanent equilibrium.
- TeMPOraL 5y ago
  The more I think of it, the more I realize, biology is very much unlike the market in such analogies. In a way, the market is biology flipped on its side, like a transposed matrix - where in nature, life forms mostly cooperate within their species (or at least ignore each other) and prey on other species, on the market, companies of the same type fight it out, while ignoring or cooperating with companies of different types. But I think the main difference is scale. The area of influence of a life form - where they live, hunt, reproduce - is very limited. Same is true with companies serving local markets. A local glass factory. A corner market. You say: > Global highly efficient markets are like a polished flat globe with a uniform climate everywhere. But it's not just a flat globe. It's a globe with a surface of couple dozen square kilometers! It's literally a single niche. Not even that - it's a cage. Imagine a zoo that decided to put all its animals in a single, large cage. You won't have to wait much before there's only one big animal standing (and a bunch of insects scattering about). On a more general note: biology is a good example that you can make a complex and robust system by giving imperfectly-self-reproducing agents a lot of space to play in, and then throwing calamities at them to keep things interesting. But I honestly don't like using it as a recipe for designing systems, because of all the waste that's involved. Waste of effort, waste of resources, waste of life. I feel we should avoid this kind of hard self-regulation whenever we can.
  - shkkmo 5y ago
    > where in nature, life forms mostly cooperate within their species (or at least ignore each other) This is blatantly not true. There are condiserable levels of intra-species competition in many species. Even in "cooperative" species, those animals cooperate in groups, but then those groups compete against each other. Most predators prey on a very limited subset of the ecology. You seemly cannot reliably predict inter vs intra species like you seem to think. Also "preying on" is pretty different that "competeing with" and at the ecological scale, predation looks much more like interactions between industries than competition within an industry. Thus the correct mapping of the analogy in this case is between ecological niches and industries. Regarding your more general note: I think there is a trade off between efficiency and robustness. Top down heirachical structures (most companies, militaries, etc) are generally more efficient in terms of resources used but are vulnerable to broken/inefficient/disengenous nodes, especially for the higher nodes. Bottom up solutions (markets, grassroots movements, etc) tend to use more total resources but do better at dealing with bad nodes and are thus less failure prone. I would argue that the best systems tend to layer combinations of these two types of systems. Edit: In the US for example, we (simplistically) have a bottom up democracy that governs a top down government that regulates bottom up markets that are participated in by top down companies that increasingly rely on outsourcing core functionality to other bottom up markets that that top down particpants.

coliveira 5y ago

People didn't think well about the other side of the coin: if anyone can hack away a product, then what you need to make a lot of money is somehow gather the best minds that can make this machine work for you and monopolize their ideas. This is exactly the strategy behind the FAANG.

api 5y ago

Software doesn't favor decentralization at all... at least not the current software and Internet ecosystem. The complexity of software coupled with its absolute rigidity makes it hard to build systems that talk to each other even when specifications are public. When things are proprietary take the difficulty of making software interoperable and add a total lack of documentation and a possibly adversarial relationship with the vendor where they'll try to deliberately shift things to keep you out. (There could also be legal barriers, but I'm just talking about technical ones.) Then consider the added difficulty of opening systems to interoperability with other systems while maintaining security. If you allow connections from untrusted systems, now you have to be particularly meticulous in the design of your software to ensure there are no exploitable bugs, undefined behaviors, or holes in your permission system. We still mostly program in language that are "insecure by default" and make this hard. Then there's the Internet itself which is intrinsically hierarchical. DNS, BGP, and the hierarchy enforced by NATs and firewalls that separate endpoints from the "real" Internet all make the network favor a "feudal" client-server model. We shouldn't be surprised that the Internet produces a lot of monopolies and closed silos; software built around the Internet takes on its shape. Last but not least there's the use of cryptography to enforce walled gardens: certificates, code signature requirements, closed app stores, etc. This adds an artificial barrier on top of everything I mentioned above. Compare this with older mechanical and analog systems. Even complex mechanical and analog electronic systems are straightforward to examine, measure, characterize, and understand. Ignoring patents and other legal restrictions it's straightforward to build replacement parts or interoperate with physical systems.

mikechen233 5y ago

It depends on the product. But software has stronger the ecosystem effect and data lockin more so than any previous technology. Think facebook, youtube, Windows, office, apple ecosystem of products. They have two market characteristics: 1. People need to be on the platform that others are in 2. Once you are in you it's hard to leave. Both creates tendency for winner takes all economics. And I think many people knows this. Part of the reason why software companies since the 90s have higher valuation than other industries.