Is the job talk portion of interviews a waste of time?

Via Daniel Drezner, an interesting attack by political science professor Dan Nexon on the job talk, which appears to be a central focus of major university hiring for political science assistant professor candidates:

The worst part is the job talk, which is an extremely artificial exercise bearing little resemblance to most of the professional activity of political scientists... 

...In fact, the job talk is most useful for… assessing the ability of a candidate to give a job talk. The reason we place so much weight on it * is that most academics (and I include myself in this category) are too damn lazy pressed for time to skim carefully read candidates’ portfolios. 

Professor Drezner demurs and says the following:

Dan's claim that it serves no purpose other than giving a job talk seems short-sighted to me.  In part, a job talk is an act of editing... A job talk lets me see whether this candidate will be able to talk to anyone outside of the five other people on the planet who know this specific topic cold.  

...If I've read the paper, I'm always curious to see how a candidate crafts his or her presentation.  And if the presenter can't hold my attention, that's a bad sign, because if they can't make their own work compelling, good luck keeping the attention of less interested students with work that's not their own.  

...Truthfully, however, the most important part of a job talk to me is not the talk, it's the question and answer session afterwards.  How well can a candidate respond to tough questions?  Stupid questions?  What are the reservoirs of expertise that lie below the surface?  In my professional experience, I can only think of a handful of candidates that blew their chances with the actual job talk.  I can think of a LOT of them, however, that deep-sixed their chances because they couldn't handle good questions.

Readers of this blog will recognize that Dan Nexon's negative comments about job talks could be made about both academic chemistry and industrial chemistry hiring interviews. The job talk is indeed artificial -- in the number of years since I left my postdoc, I've very rarely found myself in front of a group of strangers, wearing a suit and holding a laser pointer, while giving a artfully crafted 45 minute presentation while sounding relaxed, but energetic.*

I think that there are a number of things to be said for the job talk (or "interview presentation") for chemists. Like Drezner says, a job talk is an act of editing; it shows your ability to boil down the last 5 to N years of research into 45 minutes, your ability to ask interesting questions and to find out the answers. I think it's great, because, for the most part, you should have tremendous home field advantage over your audience -- after all, it was your own work.

I do think that the questions portion of industrial job interviews can be most illuminating of a candidates' ability to think on their feet and their instant recall of the fundamentals of their field (I can remember the moment, I think, where I sunk myself at the small company job I really wanted.) They can also be a time where the audience decides that they want to pick nits with your projects or play "you shoulda done this experiment!"

Are there interview tactics that are less artificial and more applicable to life in industrial chemistry?

• Synthesis problems can be worthwhile, i.e. this is a molecule I would like to make -- can you talk about how you might make it? 
• Scenario questions are fun, i.e. how might you scale up your 10-step total synthesis? Which steps might you focus your research on? 
• I think there are probably legal strictures that disallow some sort of examination at the bench; it'd be interesting to see what people could come up with.

Readers, what do you think? Are job talks stupid? Can you come up with a better idea? 

*Maybe in sales?