It wouldn't be the first time Atwood made a generalization based on one data point.
Perhaps one could generalize badly with 2 data points, but isn't it just impossible with 1, and therefore he must actually be doing something else?
"This wouldn't be the first time that Atwood came up with a theory that only had to fit one data point."
He could just as easily have said that all open source developers are Italian or that Italians can't spend money, and he would have been just as wrong.