3 ms·

All the comments here are about journalism or politics so I'll start a thread about technology instead. The article says this: this kind of power hoarding exis

All the comments here are about journalism or politics so I'll start a thread about technology instead. The article says this: this kind of power hoarding exists only because of insufficiently farsighted design of the early web ... Were there a public protocol that allowed video to be shared as easily as hypertext, there would be no need for YouTube But this is a major misunderstanding of the problem, which is not due to the web's design but rather the basic incentive issues around the creation of decentralised, public infrastructure. Because of course there is such a public protocol that allows video to be shared as easily as hypertext. It's called the web. How does this author think YouTube works, exactly? Anyone can transcode a video, upload it to their personal web server and use the <video> tag. The codecs, the transports, the viewer software, it's all free and public. So why don't PewDiePie and others like them do that? Why do they all flock to YouTube and then create massive dramas over YouTube's policies? And by the way, I ask these rhetorical questions as someone who is no fan of Google's recent turn towards blocking and censorship of conservative views. Well the questions are rhetorical because the answers are obvious. They stick with YouTube and don't even go elsewhere if they get silenced because the value YouTube is providing them is staggeringly huge. YouTube isn't merely a way to publish video. It's a huge, no questions asked bandwidth and CPU subsidy. It's the technology to instantly handle serving to planet-scale populations. It's commenting and voting that's defending against spamming and other forms of DoS. It's discovery and sharing. It's editing tools and analytics and most of all it's advertising - monetisation is what we lack public infrastructure for, not video serving. And why do we lack public protocols for all these things? Because creating YouTube is a hell of a lot of work for many people and they want to get paid for their skills. Which inherently means they need to own the result of their work, and ownership implies control. Private property rights, that's capitalism and incentives 101. I've worked on decentralisation related software for a long time and still do, because there are certain areas where it can really help. But you can't inherently point the finger at existing technologies and say "if only they were better, everything would be decentralised". No, it doesn't work like that. It's a failed analysis of the problem. Create decentralised tech like the <video> tag or git, and a YouTube and GitHub will arise around it and re-centralise it all by providing value the public protocol does not. Society wants centralisation because most of the time, the abuses of power it allows are collectively considered low cost relative to the benefits of the extra features and competence the centralisation allows. This doesn't mean it's all useless. Git being decentralised is helpful even though everyone has centralised around GitHub: if you do want to move your code off of GitHub, you are just a "git push" away from having all your history and branches painlessly transferred to GitLab or your own server. Sure, that won't help your issues or pull requests, but that doesn't take away from the fact that source code is now properly decentralised and that's still better than before. So we should continue to develop decentralised technology, whilst understanding that it won't result in everyone using purely decentralised systems.

allenf13 8y ago
Hi all - this is Allen Farrington, the author of the piece (not sure how to prove that but oh well. If only we had a public protocol for identity :P ) I just wanted to say that of all the commentary on the article across the web that I am aware of, this is by far the most interesting and informed. I would really encourage the contributor of this post to copy and paste the whole thing onto the Quillette article itself as it may well get a great conversation going. In short, I am aware of all these issues and agree that the best way to frame the problem is not really in terms of 'technology' per se, but in terms of incentives. That folds in what I think is the all-important point about monetisation, which I completely agree with. Where I think I disagree slightly is the overly deterministic assertion that 'recentralisation will occur' due to social demands. I think this understanding underplays how much social interaction with technology is moulded by the technology itself. This is getting a bit subjective at this point, but I don't think society really does 'want' centralisation, at least not along the lines of such a sweeping declaration. I think they like the results of centralisation as they happen to exist on the web so far. But, as above, these are strongly determined by the available technology, and in turn what incentives are viable. As you say yourself lower down, there are cases at meaningful scale where the degree of decentralisation is increasing, and other more speculative cases in which the degree of decentralisation is a step change from current norms. My thought (/my hope) is that this provides enough leeway to enormously alter incentives such that social interaction with the technology can alter enormously as well. Fortunately for some, unfortunately for others, I covered these issues extremely briefly and overly rhetorically in the article as I am assuming that some 99% of the audience has never even considered these issues, never mind having the technical understanding to follow a discussion this in-depth. I was really using the hook of PewDiePie to meander towards this exact topic. That said, I am very happy that this discussion has got going here! Good job, and please spread it if you can!
- repolfx 8y ago
  Glad you liked it! :) You may well be right about the interplay of what society 'wants' vs technology choices. After all, git and GitHub competed in a market against totally centralised and proprietary version control systems, and won. The decentralised nature of git, making forking easy and so on, was surely a part of that, and is certainly the hard part relative to fairly trivial features like an issue or PR tracker. I'll copy and paste the comment into Quillette. One might say, I'll fork the conversation. Perhaps we need a git-like mechanism for discussing essays on the internet :)

allenf13 8y ago

Hi all - this is Allen Farrington, the author of the piece (not sure how to prove that but oh well. If only we had a public protocol for identity :P ) I just wanted to say that of all the commentary on the article across the web that I am aware of, this is by far the most interesting and informed. I would really encourage the contributor of this post to copy and paste the whole thing onto the Quillette article itself as it may well get a great conversation going. In short, I am aware of all these issues and agree that the best way to frame the problem is not really in terms of 'technology' per se, but in terms of incentives. That folds in what I think is the all-important point about monetisation, which I completely agree with. Where I think I disagree slightly is the overly deterministic assertion that 'recentralisation will occur' due to social demands. I think this understanding underplays how much social interaction with technology is moulded by the technology itself. This is getting a bit subjective at this point, but I don't think society really does 'want' centralisation, at least not along the lines of such a sweeping declaration. I think they like the results of centralisation as they happen to exist on the web so far. But, as above, these are strongly determined by the available technology, and in turn what incentives are viable. As you say yourself lower down, there are cases at meaningful scale where the degree of decentralisation is increasing, and other more speculative cases in which the degree of decentralisation is a step change from current norms. My thought (/my hope) is that this provides enough leeway to enormously alter incentives such that social interaction with the technology can alter enormously as well. Fortunately for some, unfortunately for others, I covered these issues extremely briefly and overly rhetorically in the article as I am assuming that some 99% of the audience has never even considered these issues, never mind having the technical understanding to follow a discussion this in-depth. I was really using the hook of PewDiePie to meander towards this exact topic. That said, I am very happy that this discussion has got going here! Good job, and please spread it if you can!

repolfx 8y ago
Glad you liked it! :) You may well be right about the interplay of what society 'wants' vs technology choices. After all, git and GitHub competed in a market against totally centralised and proprietary version control systems, and won. The decentralised nature of git, making forking easy and so on, was surely a part of that, and is certainly the hard part relative to fairly trivial features like an issue or PR tracker. I'll copy and paste the comment into Quillette. One might say, I'll fork the conversation. Perhaps we need a git-like mechanism for discussing essays on the internet :)