How much in tokens/Claude subs would it cost to make a paint.net open source clone?

Seems like the kind of software that with enough tokens an agent would be able to create from scratch autonomously (because you have a clear goal and reference)

I'd guess with 5 max subs running fable 5.1 almost non stop it should be doable in a couple of weeks.

▲

jadar 6 hours ago | parent | next [-]

The author says it's about 700k LoC that he's worked on for 20 years.[1] I don't know how many tokens that arrives at, but that seems non-trivial.

[1] https://simonwillison.net/2026/Sep/2/rick-brewster/

▲

redox99 6 hours ago | parent [-]

700k LoC of human written code is probably like 2M LoC of LLM "slop". That's probably around 8M tokens. Assuming a single fable agent that one shots the new source code at 30 tok/s, that's 74 hours. Of course it needs wayyy more tokens because it wouldn't one shot it, plus the reasoning tokens. Generation speed would be faster because you'd have many simultaneous agents though.

Bun rewrite took 6 days of coding and was a similar number of LoC, so I think a couple of weeks is definitely reasonable, although you probably need more than 5 Claude subs.

▲

slowin 5 hours ago | parent | next [-]

It's way more than 8M tokens. That might be around the amount of the raw output, but there's going to be tons of tool calls, back and forth... it's probably more like 100M+ tokens.

▲

solid_fuel 6 hours ago | parent | prev [-]

> Bun rewrite took 6 days of coding and was a similar number of LoC, so I think a couple of weeks is definitely reasonable, although you probably need more than 5 Claude subs.

Bun had numerous standardized and extensive javascript test suites to validate the slop code against. I doubt paint.net has test coverage anywhere near that complete, and without it using LLMs to rewrite something is a much larger risk.

Besides which, Bun claims the rewrite took 6 days but they took months after that to finish hammering out the bugs and actually ship, so I don't really buy the original price tag or timeline on their rewrite.

▲

redox99 6 hours ago | parent [-]

Obviously the agent would need to build a very extensive test setup for Paint.NET, a lot of it with computer use or something similar.

Opening the same files, performing the same clicks, etc should output the same on paint.net and the clone.

▲

solid_fuel 5 hours ago | parent [-]

Obviously, yes. That's why it needs to be included in any estimate.

My point being "a few weeks and 8M tokens" is wildly unrealistic.

	▲	redox99 5 hours ago \| parent [-]
		Yeah for sure. Easily 10x that. And obviously we're talking just output tokens. I was mostly just pointing out a theoretical lower bound.

▲

keithnz 5 hours ago | parent | prev [-]

I think it's roughly possible for any piece of software. I think the problem would be how long it would take to polish it afterwards. I think it would be an interesting experiment to run an AI only open source project though. Not sure cloning another paint program is really worth the effort however.

	▲	redox99 5 hours ago \| parent [-]
		> I think it's roughly possible for any piece of software. Still not possible for games (many people are attempting it on X). It's definitely getting better with every new model and it's just a matter of time before they can actually do it. Fable 5.1 is quite strong https://x.com/loktar00/status/2095192155086057978 (that's just a main menu, other people are attempting entire games which is way more difficult of course and results much weaker)