Electronics Engineering.
Technology changes so fast, the most important asset an engineer can have in the future will be a thorough enough foundation in the fundamentals of science in general to be able to bring intuition and insight into the particular work place.
Improvement in computers and applied computer science in the future seem to be dependent on improvements in the basic building blocks such as memory, and possibly in effective software concepts optimizing application of those building blocks. The first means challenge and opportunity for physicists; the second means challenge for computer scientists, but probably not a lot of opportunity.
When I made my choice of Electrical Engineering so many years ago, it was because of the insights MIT offered in that basic discipline, and when I chose Feedback Controls as my primary option it was because a major challenge at that time was improving precision and speed of controlling applied engineering concepts. Challenges like intercontinental missile delivery, space exploration and communication, precision tool controls, automatic tuning and adjustment of televisions, and ... computers. Today, I might come up with a different primary option but I would use the same decision process I did then.
Computer and software concepts will be a fundamental part of any engineering curriculum today, but the time when they were an optimal choice to maximize job opportunities and salary is past.
