Source Code got cheaper, the rest of the job didn't

DenisManig

In early October, AI agents turned out look-alikes of Photoshop, Premiere, Lightroom and Autodesk Fusion within days. If you're a Software Maker, whether you run a SaaS, an app or an open-source project, that lands close to home: if people can rebuild what you built, what do they still come to you for?

What actually happened

Phoronix reported on 7 October that PhotoCraft, FilmCraft and LightCraft were built with Claude Code and other agents. A Fusion clone called SolidCraft followed. PhotoCraft calls itself "early alpha", and the reports on how complete these projects are differ a lot.

What interests me is the motive. Somebody decided it was worth their time and a lot of tokens to rebuild a tool they already use.

Why would anyone do that

Two things come together. Building got cheap: a developer on mrkaran.dev describes replacing a shut-down TV tracker in two days with an agent, with a result that does exactly what they want and nothing else [1]. Robin Sloan called such apps "home-cooked" back in 2020 [2], and people have built their own tools for much longer than that. Eric von Hippel found in the 1970s and 80s that users developed about 80 % of the important innovations in scientific instruments [3]. What changed is the price of building.

The other half is what people get from vendors. According to OMG Ubuntu, which read the Reddit posts of the developer behind ArtCraft, the project started with the fees, hurdles and hostility they ran into when cancelling an Adobe subscription [4]. In June 2024 the US FTC sued Adobe and alleged that it hid early termination fees and made cancelling hard. Adobe says its cancellation process is simple and wants to contest the claims [5].

That's one example, and the pattern is familiar. Software companies decide for their users. Rules get imposed that users never wanted, features arrive that nobody asked for, and prices stay where they are because there's hardly an alternative. These are the reasons I see, and I'm not passing a verdict on any single company. Cory Doctorow describes high switching costs as what allows platforms to get worse for their users [6]. An agent that rebuilds your tool in a weekend lowers the cost of leaving.

So my reading is that people mostly do this because they want something other than what they get. I haven't found a study that measures how many users really rebuild what they used before, so this rests on examples and essays.

Coding is one part of many

I've been developing software for 20 years, and writing code was always one part of a lot of other work. It doesn't drop to zero either. Someone still describes what should be built, reviews what comes out, tests it and ships it.

IDC's survey for 2024 shows how developer time splits [7]:

  • Application development: 16 %

  • Requirements and test cases: 14 %

  • Security: 13 %

  • Deploying code: 12 %

  • Application performance monitoring: 12 %

Atlassian's 2025 survey of 3,500 developers puts coding at 16 % of their time [8]. The definitions differ between surveys, so read these as a range.

A back-of-envelope example, my own arithmetic: if coding is a sixth of the work and gets ten times faster, the team ends up about 17 % faster overall. The rest of the work stays, and its share grows.

If you maintain a project alone, all of that lands on you. Deciding what to build next, security fixes, releases, support, upkeep. Flavio Copes argues that the first version is cheap, and that keeping software reliable for years is what people still pay for [9].

Where Swoin fits

Nobody spends their time and burns tokens on a copy without a need. If your users already have software that fits them, and new extensions arrive quickly and the way they need them, rebuilding has little to offer beyond curiosity. Raising that bar means closing the gap between what users want and what developers build. That's what Swoin.io is for: communication between users and developers, as good as it can get.

our Swoin supporters spend Innovation Tokens (InTo) on the roadmap items they care about. Their subscription pays for the tokens, so every push carries weight. In the last two weeks the roadmap got a calmer look, with items grouped by status and the funding progress shown per item, and backing an item got easier. Your supporters can back an item with a long press, several items in one go, and undo a push right after booking. A card view walks them through open ideas one at a time, and a timeline shows what you delivered and a rough projection for the rest, never an exact date.

For you, that means you see how much your supporters put behind each item, and the recurring support pays for the work that stays expensive when code gets cheap. You can mark an item as "In progress", and the timeline shows your actual pace.

Albert Hirschman wrote in 1970 that people who are unhappy with an organization leave or speak up [10]. Rebuilding your tool is a cheap form of leaving. A say with real weight gives them a reason to stay. That's my reading, and nothing guarantees it. Swoin doesn't stop anyone from copying your software, and I don't think anything does. Satisfied users just have little reason to.

If someone could rebuild your project over a weekend, what would make them stay? I'd like to hear what you think.

Sources

Diskussion