Now, the thing about science conferences, and this is certainly true for physics conferences, is that they are more than just a place where you present your work, and you listed to presentation by others of their work. This is the place where you try to establish, for lack of a better word, a network among people doing the same thing as you do. This is the place where you try to know who's who, and where hopefully you make a name recognition of yourself to others. If you are just starting out, this is a very important step because this is where you "introduce" yourself to others. This is why, in my So You Want To Be A Physicist essay, I strongly recommend graduate students to consider going to various physics conferences towards the end of their graduate years when they have something to report. To me, this is an important aspect of the process of being a physicist that isn't covered in a school's academic bulletin.
But one can argue that doing the above can be accomplished, although not as easily, via teleconferencing/video conferencing. Still, there is another important aspect on why one has to actually be at a conference and where it really can't be done (at least, I don't see how) via electronics means. In fact, Charles Day actually quoted someone that hinted at this aspect of it, although I think he missed it:
One of the respondents to the Science poll, John Burke Burnett, left this comment on the poll's website:
Until we come up with holographic teleconferencing with the ability to eat virtual lunch together in smaller groups, there will always be a need for large gatherings from time to time.
Now, the important thing here isn't to eat lunch together in smaller group. What Day missed is the fact that, in many instances, a lot of important science discussion occurs over such informal discussion. I know for a fact that new experiments and new ideas often emerge during a conversation over a cup of coffee and some pastries, or during a quick lunch at a restaurant near the conference area. Many such back-of-the-envelope calculation are done during such encounters, and these often lead to very productive results. I definitely know that one of my PRL publication came as a result of a discussion during a coffee break at a workshop that I attended.
The point here is that, while technology can certainly do many of the things that we used to have to do in person, science is still a human endeavor, and human interaction at a one-to-one level have many elements that cannot (yet) be duplicated or substituted. We use our sensory perception often to judge the "atmosphere", and can often detect a shift or a sense of direction in a particular subject area. We often gain immeasurable information simply by talking to someone or having a discussion with a group of people. I still don't see how that can be done via video conferencing or any other means.
Bottom line: there is no substitute to attending a science conference.