Hacker News Reader

Comment by jmastrangelo | Hacker News Reader

I agree with your premise that breaking up FB, WA and IG doesn't really do anything (I made a much longer comment about this in a reply above). But what would breaking up facebook even mean then?

Fundamentally there's no equivalent to the geographic breakup, as physical natural monopolies emerge because you can only have one company per unit of natural resource. Facebook is a monopoly because of network effects, and by definition only a company as large as facebook can compete with facebook, and a facebook with only a subset of your friends becomes nearly useless.

If facebook has bad behaviour, I think regulation is the only actual solution that actually addresses that behaviour, and makes any sense given the market dynamics.

View on HN

Its absolutely possible to break facebook up horizontally. And if the government were to explore a breakup, its the way they should go.

Company 1 would be a Social Networking and Hosting PaaS from hardware, networking, bandwidth, databases, serverless workers to services like ads, anti-evil, localization, social graph, development tools, frameworks, runtimes. Company 1, would be THE place to build a social network, if you were an upstart. Company 1 could instantly be a tier 1 competitor to AWS and Azure. Imagine being able to tap into Facebooks Identity services, but running your own Directory Tenant, instead of tied to the FBID. Company 1 would compete with Firebase and restart Pulse, it would jump headfirst against Microsoft and its Visual Studio/GitHub offerings.

Company 2 become clients, consumers, of Company 1 services. They include Messenger, WhatsApp, FB, IG, and Occulus.

The other beauty of a horizontal breakup is that Company 1 lowers the barrier to entry for new upcoming social networks, without being a risk to them being eaten by Company 2. Mark then also can profit off selling his competitors service, similar to AWS profiting off Netflix.

I think the idea of your "company 1" is actually super interesting, essentially social Graph as a Service(GAAS?). But , I don't think that addresses any of the issues people actually have with facebook. What does facebook running both the front and backend of their site have to do with their market power?

That company 1 would still retain 90% of Facebooks value, as it would be the social graph where people acknowledge their real life connections. It would still have all the privacy liabilities, in fact Facebook's largest scandal (Cambridge Analytics) was due to the face they were too open with the social graph! And that it was misused! If we're selling the social graph now can't there be a million politically oriented "research operations" that spring up and use it?

Also, FB, IG and WA all have separate social graphs, them being frontend skins on one service doesn't reflect that. So they'd likely be completely separate companies full stop, maybe using a couple services from GAAS like spam algos. Sure new, potentially better frontends for FB could pop up with more specialized niches, but facebook is already exploring that with messenger and events getting their own apps and sites.

They use their market power against Company 2 competitors. Its waaay too hard to get a Company 2 type company off the ground, because you have to build/buy/lease the equivelant Company 1 services (from AWS etc) and assemble them in a logical way, all while getting a product off the ground before Company 2 can clone you. Giving people a huge head start, by eliminating the need to recreate Company 1, means the Public will see more competing social networks, and those social networks will have less incentive to sell to FB, because they can benefit from Company 1 engineering and anti-evil services without being gobbled up by the mothership.

Social Graph SERVICES are Company 1, the FB Social Graph would be Company 2.

I actually think Company 2 would retain most of the value, Company 1 would only gain value after non-Company 2 clients start buying their services. Company 1 has high fixed costs, and makes money at scale. Company 2 is where all the data about people is stored in a usable, actionable way. Company 1 can only see it in an abstract or encrypted set, they cant actually see what Company 2 or Company 2 competitors are holding. Company 1 builds the Ad infrastructure itself, allowing Company 2 competitors to spin up their own ad marketplaces, (like shopify for ads), but Company 2 is the one profiting off matching Ad customers to datasets and displaying the ads, and keeping peoples eyeball attention on them. It's like Azure, Microsoft isnt looking inside what people are doing at their data for product ideas or extracting the customer held data for its own advertising purposes (Amazon might be, but they shouldnt be.) Company 1 would be a completely agnostic company, it wouldnt know ANYTHING about the data it holds. Just database/graph/toolkit leasing service.

>FB, IG and WA all have separate social graphs

That wouldnt be the case, because Company 2 can compare, merge social graphs at will. What they cant do is look at competing social graphs hosted by Company 1. Company 2 could buy competitors to acquire their social graph or client app or userbase.

Essentially then, you're advocating for FB's core infra to become a competitor to AWS/Azure/GCP?

I guess I just don't see why this is necessary at all. Company 2 is essentially exactly the same as FB would be now if it switched its infra to AWS, with pretty much the same pros and cons as a company. Facebook would have a stressful couple years of transition, and then we'd all be right back at square one.

AFAIK nobody is offering PaaS for social networks. Youre missing the part where "ads, anti-evil, localization, social graph, development tools, frameworks, runtimes" are part of Company 1. If I want to spin up my own ad network, or social graph, identity service, spam blocking service, I can go to Company 1 and lease their technology, and plug it into my product. They wont have access to the data in the social graph, or in the ad network, but they provide the software infrastructure that makes it possible. They are the flux capacitor of Company 2 style social networks.

The barriers to entry into the social network realm (at scale) can be high (security, spam, compliance, censorship, localization.)

The government splitting them horizontally, means that by virtue of Company 1 existing, competing with Company 2 is infinitely easier, and a great way for consumers to benefit from a more diverse marketplace. Possibly the next best way besides the government funding a public/private hybrid version of Company 1, or the government enforcing some kind of interoperability standards.

Breaking up facebook this way leaves the consumer client side fb roughly how it is (and still at risk for privacy fines, and fcc action) but levels the playing field for the next generation of companies. Thats the kind of goal antitrust action should be looking to accomplish. Not to punish facebook, but to make a fair competitive marketplace for social media and messaging clients. Facebooks privacy and other scandals should be handled by other government bodies, they arent antitrust imho. FCC, AG, DoJ (for other reasons like criminal negligence.)

Reply on news.ycombinator.com