The Web Is for User Agency
berjon.com
berjon.com
I’ve often mused on my role as a developer, thinking really that we’re just another set of gatekeepers. Within normal economic relations at work, it seems we’re often taking agency from users and gifting power to others.
This isn’t anything new, and we need to not be blinded by technocratic ideas of efficiency or technological progress, as if they’re not deeply political terms.
It’s very telling that we’re regurgitating Keynes’ idea, from the mid 20th century, that growth, technology and automation will liberate us from work, when it’s done nothing of the sort.
Lovely writeup from Robin. Great to have this so well elaborated.
The brief was that they have to propose a new feature of the web and explain how it would end up being something you can use.
Really fascinating to see what students with very little experience thinks about and suggest. The importance of agency really hit home with me and I’ll try to bring that to the class.
I studied computer science, and I never had anything about ethics (philosophy, but not ethics). I have been looking for a good book about that (some kind of "ethics for engineers 101"). Does somebody here have recommendations?
When I was in school for my engineering degree, a few required classes had ethics sections. It always wound up talking about an international standard like those from IEEE, understanding international business and supply chains, and preventing a conflict of interest.
> Does somebody here have recommendations?
Sadly I have no recommendation for you as everything I read as a student was very dull, but I can't overstate how easy a conflict of interest can arrive and cause huge problems. If you include that topic in your search for a book, it may have good application. I'd be interested in what you find as well.
(excuse my quoting/mistranslation from yet another language):
" part I
... Man find himself when he tries himself against some obstacle. And for that he needs a tool. ... So the airplane, the tool of air transport, pulls the human into all them old problems.
...
part iV
... For centuries our roads have been deceiving us.. avoiding badlands, rocks, sands, only going from one spring to another.... Airplane gave us the straight line. .. and we see the deep essential bottom, rocks in the base.. ... "
The web is a stage for all sorts of shit.
Your comment reads so generic so as to be essentially meaningless.
I am sorry my comment was so generic as to be meaningless. I am trying to see the glass half-full. If you disagree with that, it is your prerogative.
I didn't have a single job as a programmer in the last 22 years where I was able to dispute orders.
Though I'd love to work -- like actually work for good money -- for an organization like Wikipedia.
I can't say about your boss, but in my experience the boss never really get the details of what I am doing. That's precisely why the "performance evaluation" form is filled first by me, then given to my direct manager, who can read my answer and then fill their side of the form themselves. Because they need my help to understand what I am doing.
This means that I can't wait for them to see what I doing and value it (don't expect recognition...) but also that I actually can take quite a few decisions at the technical level. As long as I pretend that I am doing what they said, that's all that matters.
Engineers have some power. The problem is just that often they run for the shitty solution that makes the boss richer, because they think they will get recognition for that.
We just need better connections. It s already at the point where this comment could be served by my home network without problem , even videos et al. Servers failed us, they got extremely centralized and censorious. Decentralized platforms have the same flaws. Self-serving from self-owned hardware is the original vision of the internet , and the final solution
This isn't quite enough.
The web is built on top of the internet right now, and the internet assumes that identities are mostly static. To find you, I dial your address, and that gets routed to your computer.
The problem is, in the real world, addresses for humans change frequently unless you invest heavily in keeping them stable.
On the trivial side, an ISP will rotate an IP address out from under you.
On the less trivial side, you'll move from your laptop to your cellphone.
The internet is missing an overlay network that can stably route messages to users in the face of unstable identifiers.
Right now, I'm working on solving this with the libp2p overlay network paired with a scuttlebutt style gossip tree that supports multi-device concurrent writing to an append only database. You group together one or more databases under a "user" client side to model an account for a real world person.
If you can find that database, you can find the user. We locate the database by looking for every node that announces the sha256 hash of root node on a DHT.
This also pushes availability requirements down past the threshold of the persistent connection most devices enjoy these days. Any number of devices can mirror the database, so even if one device loses cell coverage another device is still mirroring it.
The next step is solving delegation. Once you have multiple keys that can write to a shared data structure for a user, you can issue capabilities to third party services to modify that data structure. So you can do fun things like delegate managing/backing-up blog posts to a 3rd party hosted SaaS while still retaining full control of your data.
We are walking around with super computers in our pockets with a near-HA 5G cell connection. The problem is, they aren't addressable in any meaningful way for the use cases users care about. If you want to host your website, you can't do it on your personal devices without figuring out a stable identifier for them on the internet.
Edit: If this stuff interests you, we have a good list of research papers inspiring what we are building listed on our open interview challenge - https://gitlab.com/webai-open/network/interview-challenge