You need to agree with your boss about what you need to get done that week, what are the metrics of success. Sometimes you need more hours, sometimes you need fewer hours.

I've seen people spend days, if not months, researching and gathering data, but only at the end did they finally figure out what they were really looking for; then they have to redo a lot of stuff. If after a day or so you force yourself to put together your tentative conclusions, then you'll have guidance for the rest of your research.

The most important thing is to talk to people that have jobs that you might like and to see what they say the job involves.

You have various institutions like law firms and accounting firms which bill by the hour. I'm really against that. You have an incentive to go slowly, be there as long as possible, to over-research things and over-staff.

America is the only major country that tries to ascertain who was the first applicant to invent the product or procedure. This may seem fair, but long proceedings to determine precisely when each party conceived an idea result mostly in keeping innovations from hitting the market.

The quality of American patents has been deteriorating for years; they are increasingly issued for products and processes that are not truly innovative - things like the queuing system for Netflix, which was patented in 2003. Yes, it makes renting movies a snap, but was it really a breakthrough deserving patent protection?

The night before, go over your schedule and see what you're going to do and what the purpose of what you're doing is. I advocate having a two-column schedule. On the left, put down all your appointments and phone calls. On the right, put down what the purpose is.