Apr 1st is an opportunity

I like giving talks: they are an opportunity to share insights that go beyond the formalese of a paper with its proofs and figures, and show pictures, the paths you took, and the ones you wish had worked out and yet might, in others' hands. On the whole though, the act of giving talks is a slippery opportunity. It is like a banana peel left on the sidewalk, and depending on how you approach it, you step over the peel or land on the sidewalk.

I gave a talk on April 1st, and difficult as it was to follow the momentous announcement that day, I made it more difficult by choosing to give a brand new talk about recent results on sponsored search auctions, some just having left the oven. I made it even more difficult by choosing to show that the ((a) we can be creative about modeling and posing problems, and indeed we should do so because radically new methods may actually get tested and have an impact, and (b) we can bring our CS type thinking to an area worked over by Economists, and generate surprises.

I did not quite succeed in doing the above with any clarity, and as Mihai (who I have always found to be a careful thinker when analyzing situations) said, I ended up marginalizing algorithms. True. Somehow all the algorithmic hard work ended up being ultimately summarized as dynamic programming and stable matching, ~50 year old algorithmic concepts. That banana peel was slippery!

ps: Piotr was pleased that I showed new pictures. Ever clever, he also coined the word "googleheim" in developing a concept over dinner of how search companies can help art museums.