The more affordable parallel computing platforms there are the better, so I wouldn't see Jetson TK1 as
competition as such
Also, the two platforms are quite different and while the Jetson TK1 may have more cores and a higher FLOP rating, it's a SIMD architecture and not MIMD, and so for many applications it's going to be an awful lot harder keeping those cores doing useful work. And then CUDA is proprietary and while you get schematics for the board, it's not an open source hardware design. You also appear to have to register to get access to the tools and I'm not sure how open they are either. In making comparisons you need to look a little bit deeper than headline specifications.
That said, there will clearly be applications where the Jetson TK1 is the obvious choice, and diversity is key for a healthy ecosystem. Finally, it's great that Nvidia see the value in providing a platform such as this and ultimately we all stand to benefit!
Cheers,
Andrew