Since there are more and more unbearable TED talks, it becomes useful to recommend the one or the other, as I am sure many have given up on watching them all, which in the beginning of TED was still possible.

The talk by Nicholas Christakis on social networks actually taught me something new and potentially useful in many scientific or other endeavors:

1) You want to target the most connected (say for advertising, or vaccination to ensure population immunity).

2) However, you do not know who the most connected are.

3) Just take random people and let them refer one of their friends to you. Point being, the friendship paradox [Scott L. Feld: “Why Your Friends Have More Friends Than You Do.” Am J of Soc 96(6), 1464–1477 (1991)] ensures that on average a friend of a random person has more friends (are better connected) than the random person.

None of it new if isolated, but putting the pieces together seems more interesting than I would have expected.

Apart from other interesting issues (some questionable, especially the claimed 16 days of early warning, but I will not start nagging here), the most impressive for me was: Using this strategy, Christakis claims to get population immunity against flu for example, which usually maybe requires 96% of the population to be vaccinated (if I understood right), with only 30% of vaccination rate instead. Not sure how he wants to do it practically (lets bring your friend in for vaccination - right) but nevertheless: Where this kind of strategy works, the gain in efficiency is immense.