yeah, looking back i definitely made the wrong call trying to save some marginal costs as it ended up costing us many times over in engineering effort trying to work around toolchain problems. spotty platform support was a hazard we failed to anticipate.
that is an interesting decision on their part, and i am curious what their rationale for building their stack in Swift on Windows was and if it had any overlaps with our original thought process in using Swift on EC2. i have never seriously used Swift on Windows, but if it is anything like Swift on EC2, it strikes me as a path of most resistance. incredibly, they seem to be eating the cost of self-distributing custom toolchains just like we are.
They are working on a desktop web browser, a market Windows still dominates, so it makes a lot of sense for them. I suspect they are a bit better-capitalized than you are, "Miller’s blandly named startup, the Browser Co. of New York Inc., has raised about $25 million from strategic investors such as Salesforce and big-name founder-CEOs," so they can probably get away with it. OTOH, they don't appear to have a business model yet, so that may not last long.