I think software was vastly underrated in 2007.
Google also more than 10xed since then.
Amazon is up 50x or so.
Not sure what the other popular software companies were in 2007. Would be interesting to look at their performance.
The type of "dead" Paul Graham referred to in this essay is "not dangerous anymore".
I think Microsoft is very dangerous to OpenAI now. As it seems they are building an inhouse competitor. While at the same time having a lot of control over and insight into OpenAI.