Note: I'm writing every day in November, see my blog for disclaimers.

Crypto could have been really good, before it became a synonym for scams and money laundering schemes. Having a native currency for the internet could have (and I guess, still might) enable many interesting innovations, such as allowing for significantly cheaper and smaller transactions. The high friction and financial cost of doing small (<$5) transactions is, I argue, a reason why subscriptions are the dominant form of payment on the internet nowadays, instead of once-off payments. Many internet-based products have very small unit-prices (e.g. one YouTube video, one essay, one search result), and because it’s infeasible to charge for these individually, we end up with subscriptions and adverts.

I’m aware that it’s much nicer for a business’s cash flow to have monthly subscribers that semi-reliably make payments. It’s terribly inconvenient to pay for fixed operating expenses (payroll, rent, etc) with unpredictable revenue. But what’s the cost to the business of these niceties? I’d certainly pay for more internet services if it were easier to do once-off payments. The existence (and success) of micro-transactions lends evidence to how many small payments can still be profitable as a business model. I’m not 100% convinced that many small once-off payments would be more profitable than subscriptions, but I’m fairly certain.

Imagine this scenario:

• A friend sends you an article by the New York Times
• You click the link, and get bombarded with a subscription pop-up

Today, you’d either already have a subscription and log in, or you bounce from the article. Occasionally, you might sign up for a subscription. But in Boyd’s alternate timeline with easy & cheap transactions:

• the NYT website sends an HTTP 402: payment required response to your browser
• This response tells your browser that viewing this article (and this article alone) will cost you $1.
• You like the NYT, and have allowed your browser to auto-pay for any single webpage from the nytimes.com domain, so long as it’s less than $1.50
• Your browser makes a small payment to the NYT, and replies to the HTTP402 response.
• The NYT website, seeing that you’ve paid for the article, responds with the actual content. Because you’ve paid for it, there are no adverts on the page.
• Critically, all this can be done automatically, without you having to do anything besides set a limit for the domains you trust.

This works out quite well for all involved. The readers don’t get bombarded with popups asking for subscriptions, and also get to read the thing they wanted to read. The journalists get paid. The NYT gets incredibly detailed insight into how much people are willing to pay for individual articles. More precisely, this mechanism has created a very precise market for individual articles, ideally leading to a more efficient allocation of resources.

This of course has security risks, since an attacker can now transfer funds via a well-placed HTTP response or browser exploit. I’m not sure about a good solution to this. But web servers are incentivised by their reputation to be secure (not that this is fool-proof). Web browsers could ship with very low cash-transfer limits by default, and require a password before any payment is made.

Beyond these significant problems, I would personally love to pay some once-off price for ~every article that I get sent, but subscribing to every news platform, substack, streaming service, etc is too much for me.