I enjoyed this interview. Marc says,
Substack isn’t a new form of communication; in fact it’s the original form of Internet communication — the written essay, the IETF Request for Comments, the newsgroup posting, the group email, the blog post. But until now you could never get paid for writing online, and now all of a sudden you can. I think it’s hard to imagine how transformational this is going to be.
Substack is causing enormous amounts of new quality writing to come into existence that would never have existed otherwise — raising the level of idea formation and discourse in a world that badly needs it. So much of legacy media, due to the technological limitations of distribution technologies like newspapers and television, makes you stupid. Substack is the profit engine for the stuff that makes you smart.
We may hope.
Later, he says,
My partner Alex Rampell says that competition between an incumbent and a software-driven startup is “a race, where the startup is trying to get distribution before the incumbent gets innovation”. The incumbent starts with a giant advantage, which is the existing customer base, the existing brand. But the software startup also starts with a giant advantage, which is a culture built to create software from the start, with no need to adapt an older culture designed to bend metal, shuffle paper, or answer phones.
As time passes, I am increasingly skeptical that most incumbents can adapt. The culture shift is just too hard. Great software people tend to not want to work at an incumbent where the culture is not optimized to them, where they are not in charge.
I was surprised that Amazon was able to develop a logistics system before Wal-Mart could figure out the web. But Marc’s point is that as software becomes more important, the incumbent firms lose their advantage.
Marc is bullish on crypto, which he says provides
distributed consensus — the ability for many untrusted participants in a network to establish consistency and trust. This is something the Internet has never had, but now it does, and I think it will take 30 years to work through all of the things we can do as a result. Money is the easiest application of this idea, but think more broadly — we can now, in theory, build Internet native contracts, loans, insurance, title to real world assets, unique digital goods (known as non-fungible tokens or NFTs), online corporate structures (such as digital autonomous organizations or DAOs), and on and on.
Maybe. But I wonder: in the process of moving from existing trust mechanisms to “distributed consensus,” how many Chesterton fences have to be torn down, and how well will that go? When bad guys do their thing, what will the “distributed consensus” do about it? Of course, “it will take 30 years to work through all the things” may make allowances for such problems.
“When bad guys do their thing, what will the “distributed consensus” do about it?”
This sounds like Stewart Baker. I am a big fan of his, and I 100% get where he’s coming from; indeed, I used to think that way myself. But this is still a blind spot for him, despite the recent publicity about his brief cancellation by Microsoft / LinkedIn and all the “Big Tech” / “Section 230” sturm und drang.
The trouble is that anything that can stop bad people can also stop good people it doesn’t like too. And all one has to do is redefine them as bad*, the oldest trick in the book.
To answer the question: the DC is going to let the bad guys use the internet to try and do their bad things. Networks should not be commandeered for policing purposes, precisely because they make policing easy – far too easy – and policing is just an instrument of power which isn’t always used for good ends. Duh.
If people want to stop bad guys, they are going to have to rely on non-internet ways of stopping them, and there are plenty of effective, traditional law enforcement methods and techniques that work in the non-digital world.
Sadly, we are just going to have to tolerate it. There’s no good alternative, and there’s no point if the DC doesn’t enable that.
The whole problem with needing new digital institutions is not that they are intrinsically superior to existing non-digital institutions in the abstract, but because the existing institutions are too controllable in theory and too controlled in practice, and in a stifling, corrupt way that is too politically entrenched and ripe for political abuses. People need competitive alternatives that provide an escape from those controls and abuses and shake up the bad, cemented equilibrium.
Anything that enables good people to escape bad controls enables bad people to escape good controls. The former is “the use case”, the latter is “the trade off”, and that’s it. One of the reasons the government keeps getting hacked and pwned is because there is no solution that achieves two essential objectives. You can’t have both (1) the rock-solid security of the most rigorous implementations of end-to-end encryption**, and (2) maintain a capability to record and eavesdrop on everything in your own (and other) networks and decipher it at will. Either it’s impossible to do at all by anybody, or it’s easy for you, and also possible for an adversary to do too, should they acquire a set of keys to your castle.
*You don’t even have to make your definitions of ‘bad’ legally specific or amenable to any kind of precise scrutiny useful for fair adjudicative discernment; you can just torture existing language and create semantically open-ended euphemisms for ‘bad’ like ‘extremist’ or ‘hostile’ or ‘hate’ or ‘racist’ or ‘offensive’ or ‘unsafety’ or ‘whiteness’ and get away with never having to define anything, even after you warn people that you will prosecute them on that basis (so much for ‘fair notice’ doctrine) and even when you are asked point blank about it in sworn testimony.
**While it can certainly ratchet up the pressure on all the major social media platforms into never offering the service, it may be infeasible for USG to outright ban well-implemented E2EE across the board even at small providers or for home-brew operations at this point, as the cat’s already too far out of the bag, and enforcement of such a ban in anything but a highly selective (i.e., corrupt) fashion would not be feasible. Nevertheless, one can certainly imagine other countries insisting on the ability to read absolutely anything, and it’s clear the US law enforcement and intelligence communities are going to use every incident and trendy excuse to constantly lobby in that direction. And it seems quite clear that the courts will not stand in anybody’s way when the targets are ‘bad’ guys and/or ‘extremists’.
There isn’t anything internet-specific to distributed consensus. It works in essentially the same way as a consensus that a publicly tradable company’s shares are worth anything. People have traded shares long before even telegraph existed, much less the internet. The mechanisms that make the trust work are social and political (of which “legal” is a subset), not technical. MM’s old post on the Ethereum DAO hack lays it out very well, and is well worth reading in full. Excerpt:
You’re underestimating WalMart. They’ve been IT leaders for decades. I didn’t notice WalMart’s web presence until around 2016 but it’s give-or-take “the same” as Amazon ( the main difference being Prime and Amazon aggregating small firms under their brand ). And in terms of total sales, WM is @ $.55T while Amazon is $.33T – and that includes things like AWS.