s13e25: Universal Basic Compute
0.0 Context Setting
It's Wednesday, 21 December, 2022 and a cold morning in Portland, Oregon.
1.0 Some Things That Caught My Attention
1.1 Universal Basic Compute
I was talking with my friend Greg Borenstein on Monday this week and he injected something into my brain using, of all things, words.
I will attempt, in fifteen minutes, before my next meeting, to cover what caught my attention.
Any mischaracterizations are entirely my own fault.
First: competition is great until it isn't. There's enough to unpack in that sentence, so let's just assume that the competition is occurring in a sufficiently regulated environment (whether naturally or due to the wisdom and direction of humans) such that any negative externalities aren't treated as such.
It's great in that we get a diversity of approaches (a veritable Cambrian Explosion), some end up winning (for whatever reason, including no particular reason), but one thing that the particular combination of sufficiently regulated capitalism combined with competition is, well, spamming. Scale, as it were. And I know I've written about scale being harmful, but let's just put that to the side right now and agree that, as ever, too much is harmful.
So far, so Everyone Gets A Can of Coca-Cola, if we're able to agree that there are certain Coca-Cola-like things that make sense.
At that point, when ubiquity and scale have achieved, there's an argument that whatever that object or service is, it has achieved utility status. What's now valuable is what is built upon the utility.
It could follow, then, that if you want to keep creating "value", that you nationalize/socialize the utility layer, and you decide to make everyone compete a bit higher up the chain. Everyone gets clean water now, now figure out what to do now that you've all got clean water.
(It should go without saying that not everyone has clean water. And even that some people seem to think that you should have to pay, directly, for clean water. Idiots)
So far so universe of atoms.
It kind of goes upside when you've got a universe of bits. The downward pressure towards it costing less and less (citation needed, of course) to do things with bits instead of atoms, and the pressure to make things work at scale is on one level in opposition to the rest of the environment we find ourselves in, which is to keep growing and to make more money. There's a some ways you can deal with this if you're a make-money-with-bits-entity:
- Use the ubiquitous utility you have a significant market/stranglehold over as a leap/boost into the next market/stranglehold you want to develop, the idea that you're also delivering "value" in that next market
- Start creating scarcity, which again, super weird given the bits environment you're in
Number 1 is the option that many tech companies find themselves in, having pressed spacebar to interact with the monkey's paw. Take Spotify. It won, as it were, at music and made listening to music via streaming a commodity. A bunch of other entities realized they needed to join in, so now streaming music is a super commodity. It's just lying around. This is in some cases good (access to culture!) and in some cases bad (exploitation of artists!). In some interesting cases, as Spotify transitions from music-as-utility to next-thing-to-dominate, it ends up making the interface to music-as-utility worse, wanting to make all the numbers go up and to the right in its ambition to be not just the dominant podcaster, but also the dominant utility for "stuff that goes in your ears" and if we think even longer term than that "stuff that your brain perceives as sound".
This interface to music-as-utility degrading is not great! Instead of being able to listen to music, now you keep being recommended podcasts or other things! And now that music-is-a-utility, it's more valuable for Spotify to concentrate on developing the next market (more potential value, right?) than it is for Spotify to get better at whatever part of music-as-utility there might be. Oh, and it's gated and a rent-seeker.
The other option I outlined above was the creating-artificial-scarcity option, which let's just all agree is dumb, and I've only got about five minutes left anyway.
The provocative argument here would be: genuinely make music streaming a utility. Like, actually figure out how to nationalize it. Spotify or whomever wins, and you do an end-of-episode reset and everyone gets to start again, but building on top of music-as-a-utility. Congratulations, you checked off that box of your civilization tech tree, and no, we don't want just one entity to be able to hang on and have an effective monopoly over what would be more valuable and have the opportunity to create more value as a public good. There's a bookmark here for a discussion about what creating value looks like, and whether the arms race of SEO/spam/advertising is actually creating value or just moving the representations of bits of paper around.
Having a utility upon which you can build things is, I think, a bit, but not entirely like saying "well, APIs are a good thing". They are a good thing! They let people build on other things to make new things. Excellent job everyone. They're still gated, and you still pay for them, which means you get something like mapping data which should be a public good (some people might say: citation needed), and mapping data or "maps" have indeed been the subject of a lot of discussion and froth and activity about being something that others can build upon. Even recently! Again! (Take, for example, my characterization of the current OpenStreetMaps kerfuffle with the Amazon/Linux Foundation/Meta/Microsoft initiated Overture Maps Foundation).
So what internet services might be candidates for turning into public utilities, free at the point of service for anyone, and crucially not contributing to some of the ill effects of lingering things-that-should-be-utilities-but-aren't? I'm not talking about access to the internet, but the universal basic ability to put things on the internet.
One example might be some sort of universal basic compute/storage/publishing, which I've written about before, many times and at the moment cannot find any references to because I'm offline.
You might argue that one basic might be "publishing things": once you've got some data published, you can do a lot with it. You might, for example, publish some data (text! art! music!) that Stripe could turn into a store, or that Soundcloud might do something with or, even, Spotify. You might publish some news! It is, after all, just bits.
But you'd need to think this through quite carefully. What would make this different from existing ad-funded, free-at-the-point-of-consumption publishing services like Blogger, or Wordpress, or, well, Twitter? Or Mastodon? Wouldn't yet another free-at-the-point-of-consumption publishing service simply provide more fodder, grist, feedstock for the SEO/attention digital complex? If you make it even cheaper to publish, wouldn't it just contribute to an arms race? What are the different reasons you could use to put this into place?
In conversation with Borenstein, questions like those above were a good prompt to return to the initial concept: what's this for? What's it supposed to correct for? What's it supposed to enable? What's it supposed to be a moderating/feedback loop against?
(And again, this is all a thought experiment. Mostly.)
Here are some things we thought of:
Would you go for the charismatic institutional donor route to fund this at the beginning? Who might do it and why? Jimmy Wales might be an interesting choice if only because I think it'd be a feed for a certain kind of bottom-up journalism. And then how would go about delivering this through a philanthropic effort? Would it all be cashwashing?
If an individually charismatic megadonor who else might be interested or motivated to fund an arms-length universal-basic-computer/storage initiative? Some examples might be Stripe (more things to build commerce on top of), Amazon (not even AWS: just Kindle Direct Publishing, for example), and even, God forbid, telcos.
I'm of the opinion that you'd want, effectively, disallow: / for any indexing. Have this as some sort of reservation free of attention-based searching.
One thought I had was that if there were a built-in search engine for this reservation, it would not use something like Pagerank and links to evaluate the quality of a result. Links result in gaming. Instead what you might go for would be search results ranked by social trust. For some people, they might construct and curate [sic] their own web of trust: you might prefer results from Cory Doctorow and the people he trusts. You might decide to just trust whatever Oprah trusts or whatever The Economist or Logic Magazine trusts. If they happen to publish any trust web. What this gets more at, I think, is a sort of social reputational web, and I would rather that happened without invisible, inscrutable scoring. Just a very simple: I trust this person.
Anyway. Public utilities and the case for nationalizing emergent utilities that exist in the world. It'd be interesting.
That might be it for the year. It might not! I might have some time to write in the next week or so. But if not: I hope you may be happy, well, healthy and safe.
See you next time, and/or on the other side.
How are you doing?