One of the momentous things in dot-com history we are, uhh, "privileged" to witness in real time is Facebook's shift from Schumpeterian monopoly to Ricardian monopoly. In ten years, business school cases will be written about this stuff...
This isn't the first of course. Microsoft maneuvered itself into the same position ca. 1999, and the verdict back then was MS is irrelevant now that it missed the whole internet thing and is just flailing around, but MS never stopped being a hugely profitable key player.
There's others, but the key lesson learned should be: you can coast on cruise control for quite some time, 20 years plus apparently, when you can milk Ricardian rents. And that's a problem.
Economic rents (profits from exclusivity) catapulted into the cyber age:
Marxian: rents from smart capital investment
Schumpeterian: rents from access to smart labor
Ricardian: rents from best positioning (both physical and cyber).
Marxian: rents from smart capital investment
Schumpeterian: rents from access to smart labor
Ricardian: rents from best positioning (both physical and cyber).
The wisdom ca. 1999 was that dumb capital doesn't but anything and smart labor, read tech innovation, only buys ephemeral monopolies, so provides an incentive for innovation but without the long-term drawbacks of monopoly. Think patent = 20 year monopoly rights.
To summarize this semi-nostalgic tour through internet history with @GordonBrianR which twitter renders into a mess: Schumpeter and Ricardo rents are supposed to differ in kind (labor vs land), but they are related. Esp if land = mindspace...
If you think that's too far-fetched: DoJ v Microsoft, the "Browser Wars", was about upstart (Netscape) vs incumbent (Msft) jostling over who got the prime real estate on the Windows Desktop. The "product", the browser client was given away for free. Mindspace.
Also if you think this whole Schumpeter rents, exclusive access to smart labor stuff, too far-fetched: We just watched two SV heavies, Google/Waymo and Uber, slug it out and spending ten-digit $$$ over having exclusive access to one brain and his digital detritus.
The overarching SV business model is foolish money hunting foolish opportunities. For foolishness to become reality you need miracles. Money doesn't buy miracles, but it might buy miracle makers. [Btw that also explains the high concentration of super(un)pumped psychopaths in SV]
But speaking of #superunpumped and, well..., back to Facebook and its ebbing flow of Schumpeter rents. [Nevermind that it's an open question if Facebook was ever truly Schumpeterian, they basically copied and perfected the MySpace playbook. Hello Microsoft!]
But Facebook did a whole lot of things right from a very close, often very opportunistic understanding of how humans interact. I once coined the word "irrivation" for how they roll out changes to their UI: "irritating innovation". So how/when did things go wrong for them?
Maybe I should state my own relation to Facebook before I proceed. I never worked for them, never applied for a job there, but followed them from their very inception with a mix of fascination and disapproval, both as an industrial engineer and an economist.
A big reason for that is that I literally wrote the book on Facebook's MO before they came into being. And I mean literally literally. 
So yes, like any other SV startup Facebook had to go the gamut from some foolish "world changing, haha boy, go get a real job" idea to building something people want to scaling to monetizing to protecting their turf against all comers (Google+?).
And this gradual descent from plucky upstart that builds something people want to sinister middleman who milks its Ricardian advantage to further its own interest is what troubles @zeynep and a whole slate of social scientists. Along with the cryptoverse...
"Rent" is a tricky economic concept. By default, rents as monopoly profits are considered a negative, "rent-seeking", using privileged access to policymakers in order to receive royal patents has a long and infamous history.
Schumpeter rents was a way to reframe rents as a positive: a reward for socially useful economic activity (aka "innovation") under two premises: they must be ephemeral and they must create positive spillovers (externalities) for the society at large.
Facebook learned a lot from the dot-com bust and was able to parameterize their indirect business model (trivia for attention/info on personal preferences, prefs for ad revenue) much better than the first gen of internet startups, but this model was always flawed.
[Little timeout for an exercise in trivial self-exploration: In your FB/Twitter/Instagram career, how often did you follow through on a "customized" ad? To investigate further? To actually buy something?]
Facebook is currently running into two self-inflicted Malthusian checks. One is generational (the "pesky parent problem"), the other one is endemic to the way they design social interaction. (A third one, the political bots, fake news, filter bubbles problem, is well-documented.)
Public Facebook is twelve years old, that's not "generational" in itself, but it is subject to an inverted "pesky little brother" problem: to the extent it attracted parents it lost the users seeking refuge from parental supervision. But that's the smaller problem...
Facebook's huge problem is how to properly allocate their scarcest resource: attention. This is where they're constantly retooling and recalibrating, and this is where they, with a hard jerk, pulled the rug from under a lot of ventures that built their businesses on Facebook.
Every social media platform requires its participants to put in free labor in the form of acknowledgement of attention. Facebook et al have perfected this self-sustaining economy by creating the minimal unit of attention token, a.k.a. the satoshi of social media: the "like".
In return they promise to offer "interesting content", predicted by their proprietary algorithm. (Google, with a few tweaks, works very similarly). This algorithm is their "Aumann machine": the device that controls the flow of information, trivia, fake content and clickbait.
The conflict at the core of this Aumann machine is that it has to choose between "interestingness" and "reciprocity". Getting this wrong kills social media sites (Myspace, Flickr, Tumblr...) and getting it right is what I call a "drunk unicyclist problem".
Social media (aka P2P attention token economies) run on two implied contracts: "In return for your attention token we show you interesting content" and "...we steer attention your way". Most social media sites gravitate towards the first contract, and this is where they die.
They are turning a reciprocal P2P economy into an I2N (influencer-to-nobody) economy: online television. There are of course many clickbait sites that work successfully on that model, but I have yet to see a P2P site survive the switch. It's death by interestingness.
The interesting thing for me is to watch struggling social media sites try to halt their impending death-by-interestingness by cranking up the interestingness quotient (e.g. switch from chronological sorting to "most interesting" or "icymi" sorting). Drunk unicyclist indeed.
The interesting thing for the cryptoverse should be to pay attention to how Facebook quite abruptly retreated from the ditch and is now trying to revert to a reciprocal P2P model, disrupting the business models of the players who banked on the I2N model.
