Hacker News Reader

This is pretty much what the ACCESS Act proposes: https://www.eff.org/deeplinks/2020/07/legislative-path-inter...

I actually wish that proposal was leaner… unless they intentionally put in extra parts they could sacrifice during negotiation.

“Prove that a third-party client with feature parity to yours can function on top your platform” could probably be the sole requirement, taking care of both back-end interoperability and ‘delegability’ (which to be honest I suspect is an unnecessarily obscure term).

Data portability could be a good next step, but I’m not sure it’s strictly necessary to fundamentally shift incentives.

> ... social media and instant messaging giants of the world are incentivised to prefer the captive user over the paying user.

What's the incentive though? That approximately zero user would ever pay for a social media or messaging app? I think they simply have no other viable business model whatsoever.

I don’t think people are any less willing to pay for social media than they are for their phone and it’s time we recognized that.

The largest player in social wants to normalize free service and lock-in: this way its users can come but can hardly leave, competition is impossible since no one can undercut it on price, and its leverage with the advertiser approaches infinity.

Now, the moment big social is legally obligated to enable feature-complete third-party clients to exist so that users can easily switch social service providers, a company will pop up and say “we don’t make money from advertising; we’re very transparent about all our records; we provide a full API so you can keep using your favourite client app; our subscription starts at $X”. You can count that this new provider will tear out a chunk of WhatsApp’s user base.

The next day yet another provider pops up and advertises cheaper subscriptions at $X/2, further eating away at Facebook’s advertising power. FB understands that the earlier it revamps its business model, the more revenue it can generate from actually paying users; so it revises the ToS, adds paid tiers to its services and becomes very good at upselling them.

FB understands that the earlier it revamps its business model, the more revenue it can generate from actually paying users

Facebook will exploit paying users even more, since a demonstrable willingness to pay makes you a more valuable set of eyeballs for advertising. That’s the fly in the ointment.

They would either be legally liable, or must reflect this in ToS—benefitting the next provider with simpler ToS.

And remember, they would pretty much lose the ability to show ads (where mined data is used for targeting), since getting rid of them becomes as easy as switching to another client.

[Am author] The best alternative is not to need a business model for anything besides sponsoring development; there shouldn't be a single organization that needs a large sum of money to keep a platform alive. From the article:

> Not all software needs to rake in billions. Federation allows services and networks like the Fediverse and XMPP to scale up to large numbers of users without forcing a single behemoth to sell its soul to foot the bill. Although anti-domestication business models are less profitable, they still allow the creation of the same technologies that were enabled by user domestication. All that’s missing is an advertising budget; the biggest advertising some of these projects get is long unpaid blog posts.

Instead of building a platform that requires one actor to scale up to millions of users, build a platform that's simple enough for it to require less development effort and scale to only a few hundred/thousand users per federated server. Delegate hosting to the 0.05% of users willing and technical enough to run a server.

The Fediverse has scaled to millions of users with a grand total of $0.00 VC money.

Related: when people first envisioned the ideas that led to networks like the Web, sceptics would ask questions such as "Who's going to pay to make all those hyperlinks? What's the financial incentive to upload a document?"

If a system is designed right, there doesn't need to be a financial incentive (i.e., an ROI) to grow.

> if a business that connects people to people—such as a social network, an instant messenger, etc.—grows larger than X users, it could be required to allow fully-featured third-party clients to exist

I might get some of the details wrong, but AT&T was forced to let smaller ISPs or telcos into their exchanges so they could set up equipment and act as an independent ISP on AT&T's copper. This wasn't huge win for small ISPs; AT&T still got most of the DSL orders, and it had a rough time competing with cable.

Line shared DSL had a lot of problems in the US, but it worked well overseas.

For one, AT&T got most of the orders because their retail prices were often lower (sometimes quite a bit lower) than the wholesale prices they charged to smaller ISPs. It's really hard to compete like that. Add in fun things like initially line shared services were often over dry pair (no dial tone) and telephone company lineman often used dial tone to see if a pair was in use; leading to loss of service if a pair was repurposed for another customer.

The alternative for small ISPs was somehow find a way to overbuild a 3rd network in the area, or sell out to Earthlink, or to someone who sold out to Earthlink.

They never should have let the same company own the wire and provide retail service. The retail part of AT&T should have had to pay the same rates to the wire owner as everyone else. And penalties should have been applied to the lineman's employer if they caused a loss of service without properly checking the database of the wire owner company.

Alas, as you say, that's not how it happened.

As much as people like to talk up the AT&T breakup, it wasn't all that effective because it just created regional monopolies with the silly local/long distance split.

Splitting up network companies and retailers might have been more interesting, but there's still the issue of having a monopoly on the last mile of copper.

It can be different (better) in case of social—I don’t think forcing the largest player to let competitors use its infrastructure is required.

[Am author] We kinda already have this:

> If big players are required to offer a fully featured API, and the user can switch freely, they would be incentivised to stop taking their money from the advertiser and charge the user directly, becoming a commodity like electricity.

> That way you could switch between messaging providers freely, while using your preferred client app. Messaging providers will be competing on which one is the most reliable, ethical, has the most complete API, and so on.

You've...uh...literally described the Fediverse.

In addition to having multiple hosting providers/instances, there are different implementations; Mastodon and Pleroma are twitter-likes, WriteFreely is for long-form blogging, PeerTube is for video (like YouTube), PixelFed is for photos (like Instagram)...the list goes on. All these implementations share an API; someone's PeerTube video can show up in my Pleroma stream, and a Mastodon user can comment under a WriteFreely blog entry.

Mastodon-to-Mastodon migrations are quite seamless; you can keep your followers and everything. Other implementations have a bit of work to do.

EDIT: I'm afraid I worded this very poorly. I don't want to take away from what the parent comment was trying to say (far from); I'm saying that we have a useful example of what that comment describes.

Some of those third party clients will abuse user data, and then the original platform will be held responsible for allowing it.

If a client is shown to be connecting to endpoints it shouldn’t connect to, you can switch to another client.

Holding messaging service provider accountable for what a third-party client does in this scenario would be similar to holding mobile service provider accountable for actions of a phone manufacturer.

True though, in this scenario popular client apps (as well as their runtime, including OS) could become higher-value targets, and infosec research would grow more valuable.