Your comments make for interesting reading and I'm sure the majority of people would agree with you that we are moving more and more towards the era of complete remote observing. In the last generation we've already seen the move away from the astro-engineer types towards the astro-computing people and this trend will obviously continue.
It's also a truism that there are many enthusiastic and motivated observational astronomers out there who have never been near and observatory or an operations centre. Clearly with your telescope of choice orbitting the Earth this is not an option! Also, it's evident that fruitful careers can be had out of mining data archives, and as much good science has come from this as from the original science objectives. I completely agree that making the data public is a huge plus for astronomy.
Your make a useful point about training students at the Universities and of course we all hope this will continue. I am somewhat less sanguine than you though about the viability of University instruments in Australia in the era of SKA/ELT. You say it's unlikely Parkes will be operating in 2040 - but even less likely that MOST will be!
The future large radio telescopes, LOFAR, ASKAP, SKA etc will clearly operate in "satellite" type mode where you just submit your proposal, you get time and the data arrives in your in-box at some point thereafter. This should be, and is, the model towards which we are working. I would say that very few people are against the concept of remote observing per se.
One could argue the philosophical point about how the great discoveries are made by intimate knowledge of the instrument (cf Jocelyn Bell) and the millions of modes made possible in a large radio interferometer have to be fully understood. But this is a debate for another day.
Another point which has often been made is that we are at the end of the world here, and it's not the same as being in Europe or the US. Interaction with the user base is extremely important and something which differentiates us from the rest of the world. Pulsar astronomy, to take one example, is hugely improved here because of the international interactions we have with the observers.
There's the more concrete point about how the proposed ATNF Science Operations Centre (SOC) will work. I have several reservations about that. One is that the user-operator model under which we currently work, does not really seem appropriate for an SOC. The problem then is that many of the "disadvantages" of observing (long hours, no sleep, travel costs, time away from home etc) are still present in the SOC whereas the "advantages" of observing (the vicarious thrill of being on-site, interaction with local staff, ability to fix things quickly, being fed) have been lost. This inevitably leads to the "why do I need to come to Sydney from Melbourne/Manchester/Masssachusetts to sit in a dark room when I could just as easily do this from home?". Furthermore, the SOC as envisaged would seem to need to cope with 3 different types of observing. The ATCA/Mopra observations can be done remotely, for Parkes it is difficult to see how completely unattended operations can work, and for ASKAP the user-operator model is invalid. This is going to be very tricky to get right but I'm sure Phil and his team will be on the job soon! But perhaps further serious thought should be given to going away from the user-operator model towards the more common remote model of eg the GBT and VLA whilst retaining the ability to go to site if necessary/desired.