Papadimitriou was alluding to an earlier time, the 1980s, when indeed STOC and FOCS were dominated by papers (and program committee members) from MIT and Berkeley and those with strong connections with these institutions. When I attended grad school at MIT I grew to believe there was MIT theory and there was bad theory. Once I left MIT and moved off axis to Chicago I discovered whole beautiful areas of CS theory completely ignored during my MIT days.
The MIT-Berkeley axis still exists to some extent but has far less influence than two decades ago. The vagaries of the job market have forced MIT and Berkeley Ph.D.'s to spread out over a large variety of colleges and universities diluting the axis. Also the field has grown too large to be dominated by two institutions or two conferences.