You are no doubt trying to present an individual to individual truth (that one can't know from looking at "a guy"), and then use that as a springboard to suddenly present as a general truth ("flaccid tells you very little"). You start out correct, but then make sure to paint it into the corner you want it to be in. What can you base this 'general truth' on, other than the 'individual examples' that can be presented? This using of individual analysis as a 'general overall proof' that correlations don't exist is an old worn out tactic that never loses steam. People can't seem to realise the difference, or else refuse to allow themselves to see it is not a general proof. Not picking on you, it abounds all over the place.
Superlarge,
I will agree with you in large part, but in the context of this thread, I think there is a little more to the story. The claim really isn't that flaccid size tells you all that little, but that it can be very deceptive to try to base judgments of average off of your own flaccid size and your perceived growth to erection.
Just an example, suppose you are a grower, 3 inches flaccid and 6 inches erect. You double in length upon getting an erection. Now, you see other people flaccid and observe an average flaccid length of something like 3.75 inches. If you project your growth onto them, you get 7.5 inches. However, studies indicate larger flaccid penises do not grow as much on a percentage basis, so you've overestimated the average.
He's right that the only penis you will definitely know to be bigger than yours is the one that is bigger flaccid than yours is hard. You're definitely right that if you see a flaccid penis larger than yours, the good money is on it being larger than yours erect as well. Since this thread is about perceptions of the average, however, he is right that trying to estimate the average based on flaccid penises is tough work (i.e. flaccid penises tell us pretty little). This is because you have the conflicting forces of flaccid length and growth to erection, both of which vary from study to study.