I'm not persuaded that the interview flub was the problem.
Your hypothesis seems to be that if only you had answered the one question right, then PG would have seen the beauty in your plan and let you in. But you followed up immediately with the answer. And then you sent them a link to an implemented feature, which they didn't bother to click on. Why would they ignore both followups if that was the one thing that kept out someone they otherwise saw as a promising candidate?
It would seem to me that the more plausible hypothesis is that PG still believes exactly what he told you: you haven't figured enough of the idea yet. And therefore, that it really wasn't about that particular question. I suspect you're fixating on that because it was the problem that you could see, and because it was a strong emotional experience.
When I've rejected job-seekers after interviews, it was never about them failing a particular question. It was about a pattern in the interview, generally a Dunning-Kruger failure in some key area. I think it speaks well of people when they follow up with, "Oh, I totally blanked on X, here's the correct answer." But that never makes a difference, because never blanking out for a moment is not one of the characteristics I'm hiring for, and the conscientiousness displayed in the followup was also visible in the interview.
Inspired, I hacked out the feature that was the answer to the flunked question and used it on the partners.
If it was such an easy feature to add and so important that you had several strong answers to the issue that required the feature, why wasn't it already part of your product?
edit: If the feature was already present in your product, it would have been almost impossible to have gone blank.