In other words, YC got lucky on Airbnb. Because YC's first impression of hundreds (or thousands) of other founders has been wrong. Basing investment decisions on first impressions may be worse than random chance but, as long as it works some of the time, they can convince themselves they're geniuses at picking founders.
On the basis of the idea and progress, the Airbnb founders were judged as bad founders.
But on the basis of personality, estimated in a few minutes, they were judged as great founders.
It explains a lot about YC's success rate that they use this fuzzy "culture fit" test to filter founders. Anyone able to seem genuine, earnest, and flattering ("oh, look, they took notes of my ideas!") for 10 minutes can game the system, as the Airbnb founders did.
Of course, it's lucky for YC that the Airbnb founders were able to manipulate them into investing. It's just kind of pitiable that YC can't do better than this.
A huge improvement for a YC-like investment firm would be to do something more objective. Competent people do not hire employees on the basis of a first impression, they spend many hours assessing their skills, abilities, personality, etc.
A better YC would spend a lot more time interviewing founders and really delve into their abilities and personalities. YC has excuses for why they don't do this, but they're bad ones. But, again, they've had enough success to not feel the need to improve.