Almost all of the software I've written is as useless outside of the company as air pockets taken out of a loaf of bread.
The big-rectangle diagram of what I do is fetch data from A, apply B(A), and send the result to Z. This is great when A is a proprietary API delivered over long-poll HTTP, B(A) converts it into a vocabulary common to just our products, and Z is Websockets to a mobile app. When we add a new customer who uses A2 all I do is write B2() and the rest stays the same.
Outside of the company, it's useless. There are thousands of products that do the same thing for different flavours of A, B and C. However, we can get our particular product to our target markets a year earlier than someone starting from scratch.
Compulsorily licensed to a competitor, they'd barely be able to install it because it's not packaged like a product. There's an old database that has evolved over more than a decade that's as crusty as an old fishing boat's keel. If a "good compulsory licensing scheme" would compensate the company for packaging it and writing manuals for it, then it's looking at millions of dollars just in the opportunity costs of taking the lead developers off other projects.
> It seems that making good backroom software, to use internally, has become something of a natural monopoly. Creating such IT has large fixed costs and big random factors.
I can't understand this. Google, Amazon, Facebook, Apple, Oracle, Microsoft, countless others, all have their own huge backroom IT software projects.
Which one of them is supposed to have a monopoly on it?
People just seem to throw around 'monopoly' to mean nothing more than 'some large number of companies are able to achieve some things that others aren't'. Some other people just seem to use it for 'company that I think is too successful'.
There's no point in the word any more.
I think part of the problem is people feel free to define the group that can have a monopoly in any way they want. Is the intention here that 'large talented companies' have a monopoly on backroom IT software? If we're going to define groups however we want I can say things like 'people who are good at football have a monopoly on playing football professionally'.
There is a point and your confusion is frequent one. Let me clarify.
When a economist like the author uses uses the word 'monopoly', it's shorthand for 'monopoly power'. It's obvious from the context of the discussion what he means. The confusion arises when people know the definition of pure monopoly (single supplier) but don't know what is monopoly power.
Between pure monopoly and perfectly functioning markets there is a large area with varying degrees of monopoly power.
There are different ways to quantify the monopoly power. One is using market power and Lerner index.
In perfect markets Lerner index would be zero. In practice it rarely is. https://en.wikipedia.org/wiki/Lerner_index
Internet economy generates large companies with monopoly power trough the network externalities. Dominating companies create barriers to entry for newcomers. If the value comes from being connected to others, the platform that connects gets most of it's value from the number of customers and economies of scale, not from being technically better. (of course, after you have the economies of scale you can hire the best people to keep up).
Classical Natural monopolies like railway systems, telephone and electric networks are often regulated to preserve markets. Platform economies have similar attributes.
Yeah, Hanson phrases this idea poorly. It might be better to say that creating large backroom software has something of a high barrier to entry. For instance, popular packages like DynamoDB (internal to Amazon.com before public release) and TensorFlow (internal to Google before public release) make certain types of work easier and more efficient; they’re also large enough that they couldn’t easily be reproduced by smaller competitors. Those companies had to pay the said “large fixed cost,” as well as a continuous cost for maintenance, to develop those tools. Economies of scale play a related role here—the more engineers you can afford to hire, the less you have to outsource and the more you can create in-house solutions that can be reused companywide and ultimately give you a competitive advantage long-term.
In fairness to Hanson, a natural monopoly is distinct from a “monopoly” in the popular sense. Public utilities are often natural monopolies—for instance, a city might have only one power company because it doesn’t make sense to build the redundant infrastructure for two power companies. Doing so would incur a high, unnecessary fixed cost. [1] I think Hanson’s trying to draw an analogy to such firms, though I find it a bit loose.
The better word is oligopoly (which is hard to argue that they're not). Just because there is no clear single winner doesn't mean that there isn't a problem. The biggest problem of any -poly (duopoly, oligopoly, etc) is that they're almost guaranteed to lead to monopolistic behavior without government intervention usually in the form of regulation.
The really good stuff tends to make it's way onto the market anyway. That's pretty much the entire business model of AWS. They make the tools that Amazon wants, and then sells them to everybody else too, and they go to extreme lengths to make sure they suit their AWS customers too.
I thought the article was pretty clear on what the issue is. A lot of capital is going into building proprietary IT and society would benefit if more people could make use of it.
It's a bit silly. That's like saying a lot of capital is going into proprietary plumbing and society would benefit if more people could make use of it. The software is implicitly linked to the people and processes.
Open source software development absolutely has a purpose. Licensing absolutely has a purpose. But generalization is a well-known cost for those in tech, and marketing is a well-known cost at general.
I don't think that anyone is claiming that there aren't tradeoffs. You seem to agree that there are times when licensing provides a benefit to society that outweighs the cost of generalization. I don't think it's silly to suggest that companies aren't always in the best position to weigh their own costs against the larger benefits to society.
> Addendum: "Okay, so far the pretty consistent answer I’ve heard is that it is very hard to take software written for internal use and make it available for outside use. Even if you insist outsiders do things your way."
Yes, largely these days process software is the business - it defines the structure, employee roles, how business units interact, how internal accountability works, and so on.
That's why modern startups can contract out in a pure Coase style both employees (Uber etc) and capital assets (AirBnB); only the software is the irreducible core of the business.
> Note that while compulsory licensing of patents is rare in the US, it is common worldwide, and it one of the reasons that US drug firms get proportionally less of their revenue from outside the US
Wow. I was totally unaware of this. Short term exclusivity is common but I didn't realise that in the US it was commonly for the duration of the patent!
This rings true to me. I think Google's biggest competitive advantage is not network effects, brand awareness, or eyeballs for advertisers. Their biggest advantage is that no one runs datacenters better than Google. Amazon is a close second but I would be willing to bet that Google is more efficient by potentially as much as an order of magnitude.
Similarly they invest heavily in development and IT tooling. It allows them to scale in ways that very few other companies can.
I think it practically would be very challenging to achieve something like this, unless the company is interested in selling the service/product it is probably very hard to force them to make a compelling offer to the market. The Walmart bar code reader for example, what was really the inovation/edge there? I guess that is was probably not the physical hardware but rather a deep and complex integration into the internal systems, which would be next to useless for anyone but Walmart.
We also see some voluntary licensing of at least software, not seldomly free of charge (and open source). It's probably easier to build something like Facebook today, thanks to frameworks open sourced by Facebook themselves (i.e React). Getting access to hosting at scale is also easier and importantly less capital intensive than before, thanks to the cloud providers, who at least partly, make their internal hosting infra accessible for outsiders.
However, it's not like the big players are open sourcing/selling their most important stuff. AirBnb is not successful thanks to their server side javascript rendering library, but them giving it away is nevertheless helpful for someone trying to disrupt them.
Patents ostensibly offer a tradeoff - you have the right to a temporary monopoly on your invention, on the condition that you publish an explanation of your invention sufficient for a third-party to re-create it. It is not a unilateral gift to inventors, but a quid-pro-quo to encourage the public dissemination of new technologies.
Modern patent practices have diminished the latter part of that tradeoff; many contemporary patents protect only one small part of an invention, or are sufficiently obscure and complex to provide little value to a third party. The rate of technological change has greatly accelerated, which has drastically changed the value of a twenty-year monopoly; many computer-related patents become obsolete long before the patent expires.
I think there is a serious need to review the basic mechanisms of the patent system, because it has gradually transmuted into something fundamentally different. In a great many cases, patents inhibit rather than encourage innovation.
Isn’t public IaaS and PaaS exactly what he is talking about? Amazon, Google, Microsoft selling the exact software and infrastructure they use to operate at scale. So maybe regulated prices for Cloud services? But they already seem pretty low because of competition.
Access to massive data is the real value that these companies have exclusive access to, not IT infrastructure
To create true competitors to Google, give them access to Google's running datacenters and codebase, just as Google employees have. They need the data even more than code, and the only practical mechanism is (readonly) access to the running systems. Let them hire Google employees too. Ditto for Facebook, Twitter, etc.
(I'm not advocating for this, just trying to interpret a practical means for what he's suggesting. This could be a direction for antitrust)
As stated, “backroom IT” sounds like random crap middleware plumbing. But I don’t think that’s what he means to refer to.
The big-rectangle diagram of what I do is fetch data from A, apply B(A), and send the result to Z. This is great when A is a proprietary API delivered over long-poll HTTP, B(A) converts it into a vocabulary common to just our products, and Z is Websockets to a mobile app. When we add a new customer who uses A2 all I do is write B2() and the rest stays the same.
Outside of the company, it's useless. There are thousands of products that do the same thing for different flavours of A, B and C. However, we can get our particular product to our target markets a year earlier than someone starting from scratch.
Compulsorily licensed to a competitor, they'd barely be able to install it because it's not packaged like a product. There's an old database that has evolved over more than a decade that's as crusty as an old fishing boat's keel. If a "good compulsory licensing scheme" would compensate the company for packaging it and writing manuals for it, then it's looking at millions of dollars just in the opportunity costs of taking the lead developers off other projects.