Thread · 10 tweets · 22 Nov 2022

Replying to @RickByers and @davidbaron, @hsivonen

Sure — I'm not interested in the moral high ground, I'm working on a very practical problem, namely (cf. upthread): mitigating the externalities and consequences of of browsers' business model. Allow me a schematic overview in an imaginary world to see if you find commonality!
In this parallel universe, people use bookmarks a lot, for everything. They're exactly like us, except they're all demented Marie Kondos. Every browser is integrated with a bookmarking service because it would use useless otherwise. People confuse the two.
Bookmarking services (BMSs) pay browsers to be the default. That's how browsers make money, and we want browsers to make money so that's good. It pays for infrastructure. We don't need to care how BMSs make money, it suffice that something like ARPU be scale-dependent.
The scale-dependence isn't a very strong constraint, this is pretty common in general and even more so in digital. We can immediately see that this creates a feedback loop. One BMS will have higher ARPU → bid more for default → more users/scale → higher ARPU.
Is that necessarily bad? No! We can absolutely have companies providing massive infrastructure without it being a problem (eg. Automattic). Also note that the higher ARPU in the loop can translate to more money for browsers. A win! So where do problems enter?
A potential problem is if the BMS has its own browser and uses bookmark strength to drive other browsers out or to bundle other services — but it could be that those could all be fine as monopolised/dominated, so I'm going to ignore this for now.
The externalities I'm most concerned about depend a lot on how the BMS itself works. If the BMS is "dumb" infrastructure and just offers great tools for people to organise their information (again, Automattic might be an example) then there is no problem.
If, however, the BMS exerts any form of editorial control over the bookmarks (eg. recommending some over others) then we have a single source of editorial control, which comes with a host of issues. Decreased intellectual diversity, loss of complexity, heterarchy, robustness.
Making it personalised is just a second-order version of the same problem. So no moral grandstanding! 😺 I just want a robust society → needs a healthy information ecosystem not managed like SLAS scientific forestry → need a browser business model to break that feedback loop.
I tried to keep this at a remove from reality because I would really like to find some ground for progress; but if it's too close for comfort it's fine if you don't answer. Maybe we can do this over drinks where it's not subject to discovery some day!