Different investors approach this differently; it depends on the velocity of investments they make and on how hands-on they are.
For instance, SV Angel (Ron Conway's fund) has an explicit policy much like YC's: they will invest in competing companies, but that they won't reveal one company's private information to another (and they don't really get involved in the management of their portfolio companies). Both SV and YC are making a lot of small, early-stage investments, so this makes sense for them.
A traditional VC, though, is very hands-on (usually a partner takes a board seat after an investment) and thus often has a policy of not investing in competing companies--it would just be too hard to not share information, at least implicitly, when you are sitting on the board and actively helping manage the company. Andreessen-Horowitz is a VC firm, and this seems to be their approach as well. It makes sense when you're making fewer, larger, later-stage investments.
The idea that you can mask learned knowledge of a company's strategy by simply not explicitly revealing sensitive information is a difficult concept to wrap my mind around.
It is not possible to unlearn the information. If you are acting as a strategic advisor, the information has to weigh in your mind and you implicitly will end up revealing information, practically subconsciously. A simple case would be where one company explains an experiment they ran (perhaps testing a feature with a small part of their customer group) and the result of the experiment. If the competitive company comes in and says they are thinking of running a similar experiment and explicitly asks the advisors what they think—what is the response given?
I have a hard time understanding how this type of accumulated information could not enter into future judgments the advisors are making. As a management consultant who faces this type of challenge frequently (and overcomes it by avoiding competitive clients and never sharing my work) I am honestly curious how one elevates themselves above this type of subconscious thinking.
For instance, SV Angel (Ron Conway's fund) has an explicit policy much like YC's: they will invest in competing companies, but that they won't reveal one company's private information to another (and they don't really get involved in the management of their portfolio companies). Both SV and YC are making a lot of small, early-stage investments, so this makes sense for them.
A traditional VC, though, is very hands-on (usually a partner takes a board seat after an investment) and thus often has a policy of not investing in competing companies--it would just be too hard to not share information, at least implicitly, when you are sitting on the board and actively helping manage the company. Andreessen-Horowitz is a VC firm, and this seems to be their approach as well. It makes sense when you're making fewer, larger, later-stage investments.