Just seems like not a very smart long-term move to build on it.
Tough to get changes upstream when the majority of engineers working on Chromium are Google devs.
Lots of features you'd hope would be available in Chromium aren't there and have to be implemented manually, but then you need to keep your fork interoperating with a massive, moving target. Safe Browsing, Translate, Spellcheck, Autofill, Password Manager aren't available in Chromium and Google cut unauthorized Chromium browsers off from using Google Sync in 2021 (https://www.zdnet.com/article/google-cuts-off-other-chromium...)
I'm not really seeing the alternative here, though. If your main gripe with Chromium is getting patches upstream through the vendor, that's an issue with all the browser engines available to them. Gecko would be an even riskier bet than Chromium, and WebKit verges on being chronically undermaintained on Windows and Linux.
But I thought AI was supposed to be writing 90% of code by now, you're telling me building a new browser engine is still a multi-year, difficult effort?
