Die hard inclusionism is as stupid as die hard deletionism. If you cannot understand the opposing viewpoint, you do not understand your own position.
My position is that we should work to improve the tools by which we organize information without ever having to destroy any. I firmly believe that "non-notable" pages should remain marked as such. By default, penalize them in search results and hide them behind a confirmation gate. But don't delete them. Decorate the hell out of them with boxes indicating they don't meet quality standards, but if just one person thought it was notable enough to write about, it deserves to be included. Inclusion need not lead to information dilution. The binary nature of "notability" is overly simplistic, so redirect the destructive energy spent on the deletion debate towards constructive classification of data.
I still wouldn't consider deleting stuff the best solution to that. It just is a matter of organizing the data better. A hard problem, no doubt. I suppose for the YC example, the list of YC Companies could be a separate article.
Maybe this is also a hint as to what the next search engine after Google could provide. I guess that would be more like what Wolfram Alpha had promised to be.
Giving every start-up including the failed ones their own page would be a waste of space, but to give them a single line in a list is pretty good use of the space available I would say.
Even if it is only a one sentence stub, it might be interesting to someone and their mother. No one else will ever know it is there, but a tiny piece of history, no matter how insignificant it may seem now, will be preserved. It's not hurting anybody.
Imagine if some YC startup did some interesting things, but ultimately failed and was promptly forgotten. Two years later, the same team builds something fantastic rooted firmly in the work they had done previously. This new startup becomes an astonishing success. Suddenly, that prior effort (especially why it failed) is noteable. Oh noez! We deleted it!
It is arrogant for any individual or even any group to believe they have any knowledge of what was, is, or will be noteable.
Which in my experience is not the case: if it was interesting or unique people will write about it - and then it is notable.
(wikipedia is not really for recording primary information; that makes it a little useless)