It is difficult, I agree. But that's not really the point Kay is trying to make. Why should we implement these things in the web browser? Further in the interview, Kay remarks that
that is exactly the problem - that we are reinventing technologies that were around for decades (reinventing them poorly to boot, as he puts it "reinventing a flat tire").
I happen to agree with him in theory: I think it's a huge missed opportunity to force the browser to be an application delivery mechanism instead of a content-delivery mechanism (or "object"-delivery as he puts it).
But of course in practice, there are historical and political reasons as to why web-as-an-app-delivery-mechanism happened, not least of which - if we took Kay's example of MS Word - is Microsoft's unwillingness to open up the platform (or proprietary software, in general). So we had to re-solve these problems in non-proprietary ways, while forcing proprietary app providers to conform to web standards that would ultimately and perhaps knowingly cannibalize their software sales.