He became well-known via his writing in the early-mid 00s; first his book about Lisp then his essays that became popular on Slashdot. His investing happened as a consequence of his writing.
Paul Graham had an audience before he was an investor, which is the point Tom was making. He didn't get famous because of Viaweb; he got famous as an eloquent smug lisp weenie.
YC grew out of a talk I gave to the Harvard Computer Society (the undergrad computer club) about how to start a startup. Everyone else in the audience was probably local, but Steve and Alexis came up on the train from the University of Virginia, where they were seniors. Since they'd come so far I agreed to meet them for coffee.
The talk he gave was How To Start A Startup [2]. The reason he was asked to give that talk was not because he had money, but because he was a Harvard CS alum who had built/sold a successful startup then spent the subsequent few years sharing his knowledge/ideas via books and essays.
The reason Steve knew about pg was that he had read/liked his Lisp book and read/liked his essays on Slashdot.
Money was a necessary but not sufficient condition for him to start YC. Nobody would have applied to YC if not for his books and essays.
Yes, PG is influential primarily because of his writing. He wrote On Lisp, which is still well-regarded today, before founding Viaweb. Viaweb itself did not make him well-known; there were many newly wealthy dot-com millionaires at the time, and he was just one of many, except that he wrote far more persuasive essays. At the time that I wrote https://paulgraham.com/redund.html we were discussing programming language design on a mailing list, and there was not yet any suggestion of a VC fund.
The reason he got into venture capital was that some of those essays urged ambitious young people to start startups, drawing on his own experience, and specifically to write SaaS software in Lisp, as he had. The Y Combinator fund came afterwards, and its dealflow came from the people who had been persuaded by those essays. That's why the first version of Reddit was written in Lisp.
If Y Combinator had not been successful, he would have remained well known for his essays.
Though some of his ideas were not correct, looking back, it is hard to name any other writer of persuasive essays of the past quarter century whose work has been similarly impactful, except perhaps Mencius Moldbug. Moldbug's work may turn out to be less impactful—I certainly hope it does—but it's too early to tell.
Sorry but 'On Lisp' is not well regarded amongst common lispers at least. I don't fraternise with that group for over a decade, but they really disliked his general content.
