The marginal costs of production are near zero effectively. The marginal costs of acquiring and retaining a customer are far from zero for a lot of overvalued enterprise saas companies.

Indeed. The economics that OP stipulate, do not factor in interest rate and cost of labour (engineers).

Given the Internet, the cost of labor for distribution of digital goods over that is cheap as all hell when compared to the cost of labor to distribute any sort of physical good because first you have make copies of this physical good, which starts off by requiring additional raw material input.

Say we've got this mp4 file that everyone wants. Let's use bittorrent for our digital distribution. So we need to: create a torrent, upload it, and then keep their computer on for a couple more hours? Let's pay $400 for the job. I think it's well within the realm for a clever teenager to do it, or a rando you find off of Upwork or Fiverr, that who knows what they're doing, to finish the first two steps in an hour, so I think that rate is plenty generous. But let's also pay for their Internet connection for the month ($100), and a cheapo laptop to do this work on ($500). This brings our digital distribution costs including the cost of labor, using bittorrent, to be able to make billions of copies to be... $1000?

Even if you pay an engineer to do that job, labor doesn't get residuals, so aren't paid for each copy made, so the cost of labor, whatever it is, is essentially flat. Compared to if you were trying to copy and distribute a physical good, the more copies you make, the higher your costs and $1000 just isn't going to get you far at all.

I'm more interested in what you mean by interest rate though, mind explaining how that fits in here a bit further?

Your cost analysis totally overlooks the cost of producing the “mp4 file that everyone wants”.

It also seems to overlook the typical expectation of non-zero revenue.

I'm not overlooking it, it's the brokenness of digital distribution that I'm trying to underscore. If I spend four hours baking a cake, I can sell that cake and make some money. But then I don't have that cake any more so I have to bake another one in order to make more money. If I spend four hours baking an mp4, I can sell infinity cakes. In the pre-Internet analog age, we invented the copyright system so someone else couldn't come along and steal money from the original baker, err writer. But back then, it still took quite a lot of effort to make and distribute those copies. Digital distribution over the Internet makes it so dramatically easier that copyright and its corner of capitalism is broken and we need a new system to properly handle this relatively new invention we have on our hands.

Because what is the point of the whole system? We have one system for two different types of goods. One type is limited in supply because it's physical in nature. The other type isn't really limited in supply because it's digital. Why then should the same kind of money be used to pay for them? How about having two parallel money systems, one of physical goods and one for digital goods. How about for digital goods we eliminate copyright and replace it with a system where we grant every person 864.00 digital dollars (DD) each day, and track what eve check person spends their time watching, then distribute the DDs accordingly. So if I spend 2 hours watching a Disney movie, they get DD 72.00 and they further redistribute those DDs to everyone who played a part in its creation.

Then, the government sets up a system where it will buy DDs for $s, so people will still want to make digital goods, but it becomes a distinct system. This lets us ease into a post-scarcity world - because we absolutely already are there for digital goods - but without having to solve some of the issues that UBI faces.

> I don't actually how much it costs to run a DHT node, or even who's paying for them today

By default all BitTorrent clients are acting as DHT nodes (if they can receive inbound UDP packets). You can disable it in the options, but most don't (unless for private trackers on per-torrent basis). I don't think anyone actually runs a DHT node like Tor exit nodes (except researchers and search engines like BTDigg).

The supply and demand curve model is either tautologically true (when applied in retrospect), or essentially entirely useless (if you try to use it to model a future product/price).

The cost of reprodcuing a pile of bits is quite unimportant compared to the cost of producing that pile of bits the first time around. This dynamic is poorly captured by supply/demand, and this is where copyright laws have always come in. While they are being abused with absurd expiry limits, they are a necessary part of any kind of money-based society.

Without copyright, for-profit companies would insist on ever more draconian DRM and software obfuscation technologies to replicate similar systems, while liberally taking any piece of code they can get their hands on - whether open-source or from another company with inadequate copy protections.

This basic problem would remain true even if all software and media companies were worker owned syndicalist co-ops.

I'd like to see you try to send a 10 GB file to every human on the planet through Cloudflare.

I mean, that's their claim, not mine. It terms of bandwidth, it might take a week or two, plus I don't have distribution rights to any 10 GiB files that the world wants, to actually try, so I've no way to verify their claims. I'd love to try, so hmu if you have such a 10 GiB file! Email's in my profile.