Good talks are rare gems. Good talks about interesting topics even rarer. Good talks that make you want to change fields and design E. Coli which smell like bananas are the best. I saw a good one earlier this week, and its now online: Learning to Program DNA by Drew Endy. If you get a chance, check out the picture of Drew going off a waterfall in a kayake on the Lower McCloud river. That’s very close to where I grew up (and don’t you city folk come up there and ruin that beautiful neck of the woods. Stay way slicker!)
I agree that this an exciting field, especially in the long term. However let me play devil’s advocate. First, it seems that he might be spending too much time setting up too much formalism too early in the game. This runs the risk of being made obsolete quickly, or of being made redundant as the field advances past that point. It reminds me of people designing programming languages for quantum computers, an area that has no theoretical consequences and won’t have practical consequences for twenty years, at which time it will be reinvented with the contemporary perspective. Discussing what the boundary between “devices” and “parts” should be seems a little like this, at the current state of technology. However, I am not in a good position to judge, and the real judges are the biologists and other bioengineers. Do they find it useful or not? A second quibble is that if I had fifty minutes to introduce a new area of research to a technically literate audience, I wouldn’t spend even five minutes talking about social consequences. Yes it is important, but there is nothing there that any technically illiterate person couldn’t read in a newspaper editorial.
Sam, I like your point.
I’m fully convinced, on the other hand, that abstractions which are being invented now will be useful in some way. The problem with quantum programming languages (so far) has been that they were taking an old concept and mapping it onto a model where we don’t even understand why it gets the speedups it gets. But it seems that useful abstractions in synthetic biology, which will most probably not be like our current abstractions of computing devices, if they hold up, will be useful for constructing a basis for the theoretical models where solving problems now will have impact later.
I thought the social consequences part was brave. If this were a “real” technical talk and not a colloquium, I’d agree with you, probably. But the audience were mainstream computer scientists, and not biologists, so I thought it actually added value (or at least added value to the conversations I could have afterwards 🙂 )