ICC Prior Art

2022-12-28  (Updated 2022-11-10)    20 minute read

Continued from: ICC

bounties

Bountysource is a platform that allows bounties to be assigned to feature requests in an open-source code repository like GitHub (it currently does not support sr.ht, GitLab, Codeberg, or Bitbucket). Bountysource differs from incremental converse crowdfunding by:

  • allowing both users and developers to make bounties for a project, rather than just the project owners
  • acting as an escrow agent between feature requesters and devlopers (ICC explicitly does not hold any money, and allows funders to withhold funding after the deliverable is published)

Other examples:

There have been quite a few arguments over the years against using bounties:

patronage

I am personally a fan of the feudal patron system where individuals with large amounts of money (like the Medicis) fund artists (like Mozart or DaVinci) to produce art. These days, it doesn't really make sense in the modern world where we have the labor specialization to fill important jobs like food production efficiently with just a few people, leaving pretty much everyone else to do whatever they want, and wealth consolidation is generally frowned upon. But given the situation at the time, I think it's cool that rich people (and churches) had an appreciation for the value and utility of the fine arts. Rich people these days have no sense of noblesse oblige.

That being said, it is neither democratic nor sustainable for the power of patronage to be so densely concentrated to so few people.

There are some small-scale examples of classical patronage, such as art commissions and feature sponsorships. http://ikiwiki.info/consultants/ describes their model as a way to "fund the development of a specific feature", which is frankly very similar to the proposal of ICC, but without being incremental, converse, or crowdfunded.

crowdfunded patronage

To address the issues with feudal patronage, platforms like Comradery, Ko-fi, and Patreon take this model and combine it with crowdfunding. These platforms are a pretty good way for creators with a decent following to get money from the their audiences. In general, I am in support of these platforms (Comradery and Ko-fi in particular), and I believe these types of crowdfunding will still have a place in the future, but a qualm I have with them is the issue of cross-subsidization.

To focus on a particular example, let's look at the Patreon of insaneintherainmusic, who makes Jazz (and Jazz-adjacent) covers of video game music. Funders pay a set price per cover video, which suggests that that is the product they are paying for. However, the patron tiers tell a different story; depending on how much a funder pays per video, they receive perks like access to special patron-only communication channels, early previews, and free lossless music downloads. In terms of incentive, it is clear to me that these perks, and not the actual videos, are what these patrons are paying for. After all, the videos will come out regardless of whether I, as an individual, choose to give $5, $1, or $0. For producers who don't have this kind of tier, Patreon is essentially a donation platform.

UPDATE: Patreon is phasing out per-creation billing, and will discontinue it by November 2025: https://news.patreon.com/articles/understanding-apple-requirements-for-patreon

Ko-fi is similar, but only allows monthly videos (rather than per-deliverable), and also has infrastructure for shops, commissions, and one-time donations. While their language is more clear about the fact that you're paying for the perks, some users don't have perks, or the perks are very low in value compared to the value of their creative or technical output, which, again, means that this is essentially a donation platform (albeit a very successful one).

case study: Philosophy Tube reading list

Philosophy Tube is a YouTube channel that aims to "give out a philosophy degree for free" (paraphrased). There used to be a book wishlist where people could buy books for the channel, and it is explicitly advertised that buying books from this list could have a big influence on the topic of future videos. There are some similarities to ICC: the artist proposes topics for future works, and people can use donations to steer the direction of the channel. There are, however, a few critical differences:

  • not converse: payment comes before the deliverable
  • there is no guarantee that an episode will be published that uses the book
  • funding is one-to-one; rather than being selected by a crowd, it is selected by an individual
  • the money goes to Amazon
  • it does not scale very well (limited by reading speed)

I believe that this wishlist has been abandoned, and nowadays, the channel is mainly funded through Nebula and Patreon, which is understandable given the last bullet point and the size of the channel's audience.

commissions

It's great that technology has made it so easy for me to commission a work of art from an artist I like. However, commissions are pretty expensive, and are generally heavily restricted by copyright. This leads to a less extreme version of the issue with patronage, where only people who can afford high prices for a luxury product can influence the creative output of artists. In my opinion, the critical downside of this is that commissions are a pretty inefficient use of artist time; since a digital product can be copied infinitely, there's a lot of missed revenue from only being able to sell the product to one person.

Commissions can be done fairly easily even without a dedicated platform, but Fiverr is a site that specializes in commissions. I've used Fiverr to sell work in the past, and it's fine, but it's not really creator-focused, and the quality is a bit shoddy.

As a sidenote, it may be useful going forward to think of ICC as a combination of commissions and crowdfunded patronage.

donation

Donation is a great way for a funder to show their appreciation for the creator in a tangible and beneficial way. However, the thing that makes this method of funding so meaningful, namely the lack of direct incentive for the funder, is exactly what makes it an unreliable and disproportionate form of funding. Unreliable because whether a person does or doesn't donate is difficult to predict and influence, and disproportionate because better quality or quantity doesn't have a direct influence on the donation total.

That being said, soliciting donations is often sufficient and sustainable for maintaining a project, even large projects like Wikipedia. Or rather, especially large projects like Wikipedia.

Quick sidenote: donations to the Wikimedia Foundation do not go to the volunteer editors of Wikipedia, many of whom believe that it is important that Wikipedia editors don't get compensated in order to prevent conflicts of interest and perverse incentives. A small fraction of the donations go to server infrastructure, but most of the money goes to employees of the foundation, whose job is, as far as I can tell, to solicit more donations for the Wikimedia Foundation. A "Bullshit Job" if I've ever heard one.

Examples:

  • OpenCollective - a donation platform with a focus on transparent spending
  • LiberaPay - a donation platform where funders can provide periodic donations to creators who contribute to the commons, whether they "make free art, spread free knowledge, [or] write free software"

advertising/sponsorships

I'm not a fan of advertising.

I don't like that I, as a producer, have to choose between ethics and money. Sure, I have the option to just ignore sponsors if I value ethics over money, but say I were a 3D printing video producer, and other people in my space are getting paid to shill some 3D printer filament while I have to pay for my own, plus they can use that extra funding to make their product better. Similarly to the previous case, I'm at a disadvantage as a result of my stubbornness.

I don't like that I, as a consumer, can be so easily influenced by a third party to buy their product. This ability to influence me also gives them a financial incentive to gather data on my behavior, ideally without me knowing. Knowledge is power, and through advertising, that power can be leveraged to move money from my hands to theirs. Also I think advertisements are annoying and I have yet to hear of a single person who believes otherwise.

I don't like that I, as a third party, can get a producer to leverage the trust they have with their audience to make me money. If I don't advertise and my competitors do, and my product is not good enough to compensate for the difference in brand recognition, then I lose sales. I'm not sure what the ideal method of informing consumers is, but I think the current state of advertising is definitely not it.

To be clear, "advertising sponsorships" refers to a third party giving a producer money in exchange for promotion of the third-party's good or service to the producer's audience. Affiliate links are often used to track the effectiveness of these sponsorships. Systems like GitHub Sponsors are more akin to crowdfunded patronage or donation.

There are some benefits to advertising though, and these benefits make it a very enticing (and difficult to replace) system for the people that use it. The main benefit (for the advertiser) is that participation is extremely passive. Advertising requires no login, no payment information, no calls to action. Sometimes can even happen without the viewer noticing with techniques like product placement (and that lack of awareness is, in my opinion, a big part of the problem).

microtransactions

In a sense, it is the opposite of advertising, in that participation is much more deliberate, and rather than targeting a wide range of people, it targets a very small subset of people who pay large amounts (known as whales), and constantly re-affirm their committment. Whales often make up the large majority of income for free-to-play games, and subsidize development, allowing even huge online games like Fortnite to be widely accessible at no cost or burden to the player. "From each according to ability, to each according to need", if you will.

I also put things like TikTok's live gifts and Twitch bits in this category, but they are a notably distinct subset that encourages its own behaviors.

collectibles

There is a class of item which has near-zero utility and cost to manufacture despite having a comparatively high price. Items in this category include signed portraits, art books, and vinyl records (way more people buy them than people who actually listen to them; myself included).

The primary issue with this class of product is the fact that they only provide funds by cross-subsidizing other offerings. With this useless token whose value is practically only influenced by how much other people are willing to pay for it, you can take a token like a vinyl record that costs $6 to produce, and sell it for $100 (or more). Its value comes primarily from its rarity, which limits the amount of these that you can produce before you start losing money, but for most people, it's worth making a few of these tokens to cross-subsidize their actual product, even if the collectors make far more money than the producers.

NFTs

Though they were a passing fad that pretty much disappeared as soon as they started being regulated, NFTs were advertised as a solution for this type of problem, and I think they're at least worth acknowledging. Non-fungible tokens were a means of abstracting a concept or non-tangible item into a tradeable digital token, allowing people to effectively trade it with each other as they would a unique physical item. There was a period of time when people brought these up whenever I mentioned ICC, but it seems those days have passed.

Open source pledge

Aside from the fact that it primarily targets maintainers of infrastructural FOSS projects (which ICC is not very well-equipped to serve), the Open Source Pledge (OSP) seems in some ways similar to ICC. Notably, it follows the same converse principle of "if you produce it, I will fund it", and is also nominally "pledge"-based.

However, because of the perspective that the leadership is approaching it from, and the shame-based nature of the campaign, I am skeptical in the OSP's ability to effectively solicit donations. My criticisms of the OSP can be summed up with two observations:

  • the only difference between the OSP and a voluntary periodic donation is the branding
  • the money is coming from large companies

Difference to voluntary periodic donation (or rather, the lack thereof)

The progenitor of the Open Source Pledge (OSP) is Chad Whitacre's article "Open Source Is a Restaurant", where they spend most of the article giving a long-winded and, in my opinion, ultimately irrelevant analogy about different models for charging for food.

They criticize grants because there are "hoops" that companies have to jump through to get them, but running a whole marketing campaign to shame corporations seems to me like a lot more work than navigating the grantwriting bureaucracy. Plus, I imagine this campaign will have to be repeated indefinitely, with much more momentum and support behind it than there is now, because, as the OSP campaign's own advertisements point out, tech companies are extremely stingy, and they never seem to feel shame in that fact. Heck, there are tech companies have been directly targeted for their active role in genocide, and I haven't seen any meaningfully change their behavior on that front.

Not only that, but I don't see any recommended mechanism for choosing which projects to donate to in the OSP literature, which indicates to me that it's up to the companies to decide? In which case what you've essentially done is replace a grant writing team with an entire marketing departement that is trying to convince the companies depnding on your project to donate to you instead of one of their other dependencies. Not only is this a more expensive hoop, but it is arguably more annoying than simple grant writing because the target audience is every dependent company and their developers, rather than just the entity distributing the grant. In contrast, ICC, rather than pitting different projects against each other, its more that different tasks under the same project are competing with each other, and assuming all tasks share a marketing budget, it doesn't really make sense for them to advertise against each other.

TODO: clarify why that's the case, stick in the anecdote about how
tobacco companies agreed to an advertising ban because something
something tragedy of the commons, and more profit when they didnt have
to compete with each other in advertising

Perhaps I'm approaching this with too much naïvité, but for the type of FOSS project that this targets, why not just threaten a strike? The maintainers could just say "this is how much money we need to sustain the project, and if we don't get it, then we will stop maintaining it until we do". Essentially, that would put the burden on the software companies to figure out a good way to fund it.

As it stands, the only companies that will choose to accept the OSP are those that feel honor, or ethical obligation, creating an artificial financial disadvantage for those companies (compared to those with no such honor or ethics). One could argue that this downside is worth being able to fund FOSS projects, but I'm skeptical whether that's the case.

The money comes from large companies

This might be unintuitive, but I think money should come from individual users more than large companies. Or at the very least, the amount that large companies give to a project should be a small proportion of the total. Taking the US government as an example, sources of money have a huge influence on the decisionmaking of an organization, even without an explicit quid pro quo, so collecting money from many smaller donors is, from my observation, important for making sure a software project remains beholden and accountable to its actual users instead of just one giant user.

thanks.dev

Sentry, one of the main participants of the OSP, is also one of the main donors on the platform https://thanks.dev, another donation platform. However, unlike the OSP, thanks.dev at least tries to implement a mechanism for determining how money should go to dependencies. The main flaw in this is that this creates a metric that can be gamed. While I'm hopeful that this will generally not be the case, consider this imaginary example:

Assuming the metric is based on number of calls into a API. A developer has a choice to implement something as either one call or two, and by making it two calls, they can increase the calls from each of their dependent libraries by an average of 3. Assuming each dependent has roughly 1000 API calls to various libraries before the new changes, then using two calls represents about a 0.3% increase in the metric. If they have 60 dependents with an average donation pool of $100,000 per year: $${60 * 100,000 * 0.3% = 18,000}$$ That's a roughly $18,000 per year increase in that one choice. Of course, these numbers are complete spitballed fabrications, but hopefully this shows that just one small change could potentially lead to huge changes in income, making it very enticing to make your library subtly more cumbersome and verbose to use than it otherwise would be.

With careful decisions on how the metric is calculated, I think it's possible for this to be mitigated enough that it's a net positive, but it does end up being a bit of an arms race.

However, unlike many other algorithms, the ICC algorithm works under the assumption that all involved understand its inner workings, and in fact, only works if people understand how it works. This hopefully makes it more stable and robust, and less prone to change.

Some other attempts at similar systems:

  • https://docs.drips.network/
    • https://www.drips.network/blog/posts/radworks-gives-1m-to-foss-dependencies-with-drips
  • https://tea.xyz/
    • had a scandal a while back where people tried to claim project income by adding .yml files to FOSS repositories.

Notably, both of these are cryptocurrency-centric projects, and for the same reason that I think large companies should not be the primary funding source for FOSS projects, I also think that speculative cryptocurrency bubble money is also not healthy, with the added downside of not being economically sustainable.

TODO:

  • TikTok yum yum ice cream girl
    • she's not doing it because she loves doing it and people want to see it.
    • she's doing it because tiktok has created a system and she's figured out a way to game it.
    • my tinfoil hat hypothesis is that the thing people like about it is the sense of control and connection (the instant gratification helps).
    • how that concept be reframed into something that actually works for the common good?
  • https://artfight.net/donate
    • not sure how to categorize this is because its a donation, but there are goal tiers that collectively paywall features (i.e.
  • bitwig's update model
    • explicitly not a subscription
    • i havent bought a bitwig update in like 5 years because i dont use it as much, but im glad i still can
    • how common is it for people to just chain updates back to back instead of just waiting?
    • how do the users feel about it? how do they perceive it?
      • do they consider it "free updates that come with the update they bought"?
      • or just "buying a year of updates"? (probably this one)
  • group buys / presale
    • massdrop