A quick story: Twitter.

 Had grand dreams of building their product, Ruby on Rails was all the rage.
 Built on a relational backend through activerecord.
 It was supposed to be easy, most of us should realize that the reality is usually the opposite.

 Development time was longer, due to having to find staff who knew Ruby and the Rails framework. The reality is that it was no where near as sustainable as promised. Their tech platform was so bad the company almost went out of business (see. Charlie Rose episode interviewing the founder).

 Little by little Twitter had to fully rebuild itself. Replacing relational technology with Casandra. Replacing Rails and activerecord with Scala. Replacing front end with Ajax.

 Today the company rarely mentions its rails disaster (they still run some of it I think), thank god they survived it.

 Catalyst is little better, I would be very nervous better big on it. And yes the learning curve is high (and for what?).

 In short, the future will be probably some mix of Javascript with services.
 Something like facebook, the front end dynamic type for basics, processing MOSTLY done through Javascript, the backend being high performance mix of dynamic and static type (compiled) code using a mix of RDBMS and No-SQL.

 ORM and templates will survive, but they will survive independent of a larger frameworks. Javascript frameworks (JQuery, prototype) will do just fine on the frontend.

 Therefore what I am saying is, DBIx::Class and template toolkit will probably be fine.
 Catalyst on the other hand is DOA IMHO.