[michaelochurch]

I have a different take on that.

Some PhD students meet their advisors once a week. Some see them once a year. What's different about a PhD vs MS is the expectation of a unique, individual contribution to the field: something that's never been done before.

Many people can learn existing material on their own, but very few people can get to the frontier and deliver something the field considers important (noting that what's considered important is often subject to fads and political trends that one needs social access to know where things stand) on their own.

It will be interesting (and exciting) to see whether we can democratize education up to the MS level.

Here's the real question, though: once we train people up to be strong computer scientists, what are they going to do in a world where priorities are set and work is defined by mediocrities? Is just knowing more going to give us the power to outperform them, or is it just one of many steps?