Before I opine, I want to note that the organizers did a great job running the conference. It was a terrific experience and I'm already looking forward to next year.
In sessions I attended and during side conversations, several themes reoccurred as challenges to the discipline, with no clear solutions. Among them were:
Fault lines within the field. Several well-know experts called for the pendulum to swing back from "quick and dirty" usability studies, to testing of statistically significant sample sizes. I was not surprised to find these folks all work for large software companies with budgets and long release cycles. Try working in an Agile development environment and you'll find little patience for these rigorous methodologies.
However, if you're a practitioner, you can't just dismiss these concerns. At first your business superiors may not question your methodology. But if you recommend changes based on a six-person usability test, and then find even worse results afterwards, it's too late to bring up statistical validity.
This divide goes deeper than just disagreement over best practices. Many of the newer generation of practitioners lack formal training in statistics, and either they don't know what they're missing, or don't consider it relevant. The more rigorous faction often considers their web-raised peers to be lightweights, not fit to work alongside them on "real" projects. I think it's best to be versed in the statistics, so even if you decide to run quick-and-dirty tests, you can present them as such and provide the appropriate caveats.
Good design is the elephant in the corner. If you want to make a usability expert twitchy and uncomfortable, ask her how companies like Apple can create acclaimed, wildly successful, and usable products using top-down design without usability testing. For extra discomfort, follow up by asking why there's no evidence that user-centered design works.
An objective look at the most successful products and companies suggests there's huge value in good design. Good design is intuitively attractive to people -- they'll like it and they'll buy it, and brands that produce it acquire a positive aura that turns into money. We should be spending much more time talking about design processes and enabling great design. Yes, you can define good design. It's measured in sales!
Meanwhile, while good design can change markets, usability is incremental. It's difficult to see how good usability truly can result in innovative products and new categories. In an entrepreneurial age, usability practitioners risk being left behind.
Productizing usability. Okay, this wasn't a theme for most people, but I see it cropping up often. Web analytics vendors continue to push software to replace usability. At the conference, a slick Keynote Systems presenter efficiently marketed his company's remote unmoderated user testing application. It was the most knowledgeable effort I've seen of products trying to make usability experts obsolete. Unless practitioners can clearly and consistently articulate the value of the many tools in an expert's personal usability toolbox, they'll start to get squeezed by these intuitive sales pitches.
Taking all these problems together, I believe usability has peaked. The future for practitioners is to keep pushing on the broadest possible definition of user experience, encompassing all end-user touchpoints. We must keep expanding our skills and learning related areas -- marketing, CRM, and so on. We have to keep aware of the cutting-edge technologies under development. Soon classic usability will be only a small piece of what practitioners need to master.