Support for other browsers are coming. I am neither working on this nor a Google engineer, but I saw a tweet from relavent developers working on this.

In 2002, it was much more difficult to support multiple browsers, and not doing so was still (eventually) frowned upon.

In 2017, with the combination of pretty close parity between VM's technology support (even Chakra) and the range of powerful preprocessors, build tools and shims, if you're not supporting multiple major user agents you're an objectively bad frontend developer. This isn't about the cost benefit of committing time to smaller market shares because the cost is negligible. The only explanation can be incompetence.

... or that you're the maker of said browser and stand to gain from actively deprioritising your competitors. I don't know that this is policy at Google, but it's either one or the other.

You're right. I've built pretty complex offline first web applications (using Indexeddb, Service Workers etc) initially on Chrome but when I tested them on Firefox and Edge there was just one minor bug on Edge (Indexeddb expected undefined but was given null in one case, Chrome and Firefox did not complain). So it definitely is possible to build sites that work on all browsers.

Maybe "use Chrome" means "not tested on Firefox and Edge?"

I wonder if the Allo team would have thought of doing this if Chrome's browser share was half of what its current today.