[krutulis]

As the number of conferences and speakers grows, are you concerned that the Law of Large Numbers will exact its toll and make the sample of speakers more typical of the population as a whole? Inviting the flashiest of SF and NYC to speak is quite a builtin selection bias, but there seems to be no apparent strategy for preserving the distinctiveness of speakers as TEDx scales.

[andr]

Yes, the average quality will probably fall, but the total amount of "good" talks will vastly increase. As with any popular content platform. A problem yet to be solved is how to highlight the better talks, because the current approach of featuring a talk a day on TED.com doesn't scale.

By the way, TED never expected TEDx to grow this big, so there is little strategy in what's been happening so far. There are I think 6 or 7 TED staff responsible for the whole TEDx program, with its 3000+ events.

[krutulis]

Agreed, andr, that the number of new good talks can continue to increase. This is an interesting story and relates to a variety of other organizations seeking to scale quickly, whether from necessity or ambition. Is there an ongoing discussion of it as it pertains to TED & TEDx anywhere?

[huherto]

I am not a mathematician, but I think the law of large numbers is not applied like that.

[Tyr42]

I am a mathematician, and the Law of Large Numbers just says that if you roll a die many times, then the average of your rolls will converge to the expected value, in some strong way.

I think he is just applying Sturgeon's Law (90% of everything is crap), and then and then the law of large numbers to conclude that the average video is crappy.

[yo-mf]

Probably got his knowledge from watching a TED talk on the subject...