At first I was like "i want to use ai but dont have the money to burn for api tokens" cool.
But then I realized the backers are essentially saying "i have money and could support developers but i choose to give the money directly to a mega corp and skip the human".
I recommend you remove the policy of "Whatever the run didn't spend goes straight back to your backers' wallets." and make sure the human behind the wheel gets to eat. Somehow
While we're in the token-equivalent of ZIRP, tokens don't cost what they cost, so there's sort of arbitrage to be had. I have tokens I've been given than I'm not using, but that's not the same as me having been given cash in the first place.
Better yet: give them cold hard cash instead of what is arguably monopoly money for many OSS devs. Ironically this is something GitHub made "easy" with sponsorships several years ago.
It does seem like a worse version of a FOSS donation platform that uses regular old money. I guess one advantage is that you can ensure your money goes directly to using AI to solve specific problems on the codebase, but what does that solve? Are people genuinely worried that if they donate to some FOSS platform that their dollars would go to something else? It seems to me like this removes agency from the FOSS maintainer and gives donators more control over their donation, even though it's explicitly designed not to.
Its efficiency also relies on being better than whatever other platform/harness the maintainer is already using. It's limited to whatever the harness the platform provides, and they're taking a 20% platform fee on top. So I have to, instead of taking $10 from a donor, i take $10 worth of tokens, which may or not be spent more efficiently than me just going in with my claude subscription and fixing it, and I get $8 of those to run in a platform I don't control? In what world as a FOSS maintainer would I sign up for this? It just seems strictly worse than just having a platform that can back resolution of issues with real money... which already exists.
Howdy Solar_Fields, these are great questions and I can give you my thoughts on how I feel it's different than cash donations (which are great if you can get them!). I look at this less like patronage and more an exchange of a resource to meet the needs of both parties. I want to support the projects I care about, some I'll give carte blanche, but some I have no connection to and really just want a bug fixed. Rather than fire up my own Claude Code and throw a PR at that maintainer, instead I'm saying, "hey, you know this codebase and can use this resource (tokens) better than me, please fix it with 'free' tokens." The platform fee is really just for AWS costs and is based on modeling, but I'm sure that's not the final form. Does that make sense?
Your reasoning is logical, but fails to pass the bar of "better than or even equal to just literally using some existing platform to attach a $X bounty to some issue I want resolved". There are several popular solutions that exist already to do that, your solution doesn't materially improve there, so what is the value add? It certainly gives the donor more confidence that the issue is being resolved in the way the donor wants it to, but if your problem is to make FOSS maintainer's life easier it doesn't move the needle in that direction, because it gives more power to already demanding FOSS users and less agency to the FOSS maintainer. And even if you solve that problem, does that value add cover a very-steep 20% platform fee?
I think I've read from a few different sources that the Claude Code $100-$200/mo plans are subsidized so hard that it's basically $2k-$8k/mo in "would-be" equivalent API token usages.
This kind of makes sense in that space while the subsidies (if true) last?
Unrelated, "tokens" feels very like... back-then blockchain to me. All the craze.
Yeah, rising token costs definitely played into my thinking too. I want builders to be able to use the best models and it seems like they are getting more expensive. But maybe local models will get there?
Its efficiency also relies on being better than whatever other platform/harness the maintainer is already using. It's limited to whatever the harness the platform provides, and they're taking a 20% platform fee on top. So I have to, instead of taking $10 from a donor, i take $10 worth of tokens, which may or not be spent more efficiently than me just going in with my claude subscription and fixing it, and I get $8 of those to run in a platform I don't control? In what world as a FOSS maintainer would I sign up for this? It just seems strictly worse than just having a platform that can back resolution of issues with real money... which already exists.
This kind of makes sense in that space while the subsidies (if true) last?
Unrelated, "tokens" feels very like... back-then blockchain to me. All the craze.