Back in 1974, when a lot of people thought that trade secret law couldn't survive alongside a patent system that encouraged public disclosure, the Supreme Court in the Kewanee case patted us on the head and said, "don't worry," assuring us that anyone with a patentable invention would be crazy to elect secrecy instead. Patents were exclusionary and "strong" while secrets were "weak." And for a number of years after the Federal Circuit was formed it seemed that patents kept getting stronger all the time, while the risks of secrecy (what if my competitor gets a patent on this?) were pretty obvious.
How times have changed. The courts have been shrinking the universe of what can be patented (business methods, software, therapies), the bars to patenting (obviousness, indefiniteness), and the enforceability of patents (injunctions, damages, fee-shifting). And Congress, through the America Invents Act, has made it easier and cheaper to challenge patents without going to court, establishing the Patent Trial and Appeal Board, which some have referred to as "patent death squads." While patents seem under attack, trade secrets are basking in a new level of attention from industry, reinforced by provisions of the AIA that virtually eliminate the old risks of protecting innovation by secrecy.
So does this mean that we should abandon patenting as a strategy? Not at all. Good patents remain strong, not only in protecting novel inventions from theft, but also in building recognized value, enabling financing and collaborations. Yes, our calculus needs to change, particularly in some technologies. But it was never an either/or situation anyway. The question of patenting or secrecy is less like arriving at an intersection than it is like eating at a buffet: you get to have something of everything that you like.