I am often impatient with arguments that assume we live in a free-market world, because we don’t – people don’t always think or behave rationally, and competition isn’t always perfectly fair. But in some cases, imagining that the way things are is just a reflection of market forces can be helpful in understanding a situation. What if, rather than being marketing victims duped by Apple, Google, or any other software company promoting software-as-service, tethered appliances, or locked-down devices (which you might believe to be the case if you read the comments on tech blogs), people use these services because they want them? Call me crazy, but maybe the shift toward non-generativity stems from the fact that non-generative devices suit people’s needs, or at least do so better than generative ones.
I’ll suggest that for the vast majority of the population, that is true. Grandma doesn’t need a generative PC. Grandma needs a device that will help her get the job done, even though she has no idea how a computer works. It needs to be intuitive and safe, so that even when she clicks the wrong button or performs the wrong procedure, nothing truly bad happens. She doesn’t want to deal with software that is potentially ineffective or malicious. The fact that both the number of programs she can run and the number of in-program adjustments she can make is greatly limited is probably of no concern to her. She just wants secure software that gets the basic job done.
I’ll also suggest that where we are now is a fairly normal and expected place on the timeline of a new technology, rather than the splitting-train-tracks situation depicted on the front of Zittrain’s book.
New technology often starts out generative and ends up non-generative. The automobile, for instance, started out much more generative than it is today – around the turn of the twentieth century, there were hundreds of hundreds of manufacturers tinkering around with their machines. While it wasn’t quite as generative as the PC or the Internet, the engine was fairly accessible and the automobile owner could make his or her own changes as desired. That’s still true today, but to a much lesser extent. Auto technology has improved to such an extent that most car owners would rather take their car to the mechanic than look under the hood themselves, and manufacturers are increasingly finding ways to monopolize repairs and prevent owners from making their own adjustments.
So what happened? Around the turn of the century, everybody and their mom was trying their hand at making cars. This led to a lot of innovation, and some problems: breakdowns were frequent and car safety was a novel concept. The industry started cleaning itself up around 1930 as people became concerned about reliability and safety. The number of manufacturers greatly decreased and names like Ford and Chrysler came to the forefront. These days, most people would not even consider buying a car from an amateur. But there are still isolated tinkerers out there.
Am I crazy for thinking that what’s happening to software is essentially the same thing? When a new type of technology shows up, it’s often basic enough that anyone can get involved. Lots of people do, and some of what results is awesome and pushes the frontiers of the technology. And some of it sucks. This vanguard is not particularly perturbed with the suckiness, but that all changes once the technology is spread to people who aren’t interested in the technology itself, but rather the benefits it might carry for them. These people don’t want potential. They want current value. They are interested in what the technology can do, not what it could do. Obviously we’ve seen this with the PC and with the Internet.
To cater to this population, hackers started making software and hardware for non-hackers. It was user-friendly and reliable, and not as generative. In fact, the more user-friendly and reliable the software, the less generative it was. This is essentially unavoidable. Most software is only useful if its capabilities are spelled out and finite. Make it any more generative than that, and it’s easy for the user to get overwhelmed.
But the reason that we shouldn’t get concerned about this shift is that it represents the expansion of software (or any technology) into a new segment of the population rather than a transformation of the technology itself. The hackers still want generative software and hardware, and they’re still getting it (more on this in a second). The people who want a reliable experience are getting that. So everyone wins. It’s not a zero-sum game, because the software industry is not monolithic. What is happening now is essentially consumer discrimination – Apple and other giants are writing software for n00bs, and smaller developers (but also to some extent the giants) are writing software for hackers.
Of course, Zittrain gets this (he quotes critic David Post: “a grid of 400 million open PCs is not less generative than a grid of 400 million open PCs and 500 million locked-down TiVos”). He has two main concerns: first, that through market dominance non-generative software will eventually swallow generative software; and second, that a generative software world is more advantageous than one that is primarily non-generative.
I don’t think the first consideration is as threatening as Zittrain makes it out to be. The fact is that the software market is very, very large, and there are many, many niches. Since hackers write software, and hackers are the kinds of people who will generally want generative and unlimited software, you have to construct a pretty convoluted hypothetical in order to get a world where literally all software comes through corporate giants.
The second concern should be taken very seriously though. How important is creativity? Is it more important than efficiency? Is it more important than security? Is it more important than giving people what they, on some level, want? These are obviously big questions.
Ultimately, I just can’t side with Zittrain here. The primary argument for bolstering the generativity of software is that the advantages it provides in innovation outweigh the negatives. That argument has a lot of merit, but only for some segments of the population. How much innovation is Grandma realistically going to produce? Innovations are always driven by a very small portion of a given population: the curious, creative types, the inventors. Most people just consume and don’t innovate, and that’s okay. The 400 million open PCs referred to by Post are probably just as generative as the most innovative 25 million-PC subset. Most of those PCs are, generatively speaking, dead weight.
So what’s important is that these different groups have different environments. The inventors want and need a sandbox to play around in, and if given one, they’ll continue to push the frontiers, develop new products, and make cool new toys. The consumers want and need the toys. They don’t want the sandbox; they’re not in it for the mucking around. If you give them one, they’ll get frustrated, they won’t do anything, and they might even get hurt.
Given that, we don’t want a purely generative software environment. We want a segmented environment in which the software that is generative is incredibly generative – it is essentially unbounded in its capabilities (with all of the associated problems and complications) – and the software not in that category is safe and efficient first, and generative second. In this environment, people can choose where they belong, depending on the task, and adjust as necessary. In other words, we want an environment that is roughly the result of market forces. Most purchasers will look for safe and non-generative software, so most software will be safe and non-generative. And enough people are interested in generative software to keep that niche healthy and the innovations flowing.
In order to bring about and maintain this kind of environment, generative software needs to be absolutely sacred. We’re putting all of our innovation eggs in one basket, so in the generative environment needs to allow for as much creativity as possible. Jailbreaking most devices should be legally protected. Of course, the developers of those devices can terminate their warranties and withdraw support, but if users want to experiment with their hardware or software, they should have the right to. You can purchase a tethered device if you want to, but you ought to be able to “de-tether” it at any time (manufacturers can determine whatever non-legal consequences that will entail from their end).
While these changes would definitely help, I think we’re doing just fine now, all things considered. Yes, software, the PC, and the Internet are all becoming less generative for most of the population, but until they become less generative for everyone, I won’t be sweating bullets, and on the other hand, the security that comes with non-generativity may be just what Grandma is looking for.