yes i'd been thinking about how to make it even since computers have no problems playing in their "heads"

You have 3 GM's in a room, discussing what to play,

This is really amazing! :-) I think of Kasparov, Ivanchuk and Salov.

PS: I´m sure, you are a democrat. Majority is always right! This is true, but only in the long run. Meanwhile you have lost hundred of games (or wars or fellows).

     I'm sure we could find three 2600-2650 level GMs who would do this for the amount of money you could bring in the way you are describing, but they would have no chance. To get Anand, Kramnik, and Carlsen together to play a six game match this way would probably take one million dollars. As for allowing one (or more) GMs to use a board or screen as an aide while playing, I have no problem at all with this, but I think if you are talking about Elite players it would only raise their ratings by something like 20-25 Elo. You mention tests on this, but you don't mention whether the tests were on Elite GMs or just ordinary amateurs (or something in between). Personally it would help me considerably, but I'm not even 2400 FIDE. Can you point to those tests or give more information about them?
     As for giving the human access to the same book and TBs as Rybka, I could accept that as well, as long as the book is just a database without the move choices of Rybka highlighted. In other words, the Rybka 3 book will have some moves highlighted in green, others not; the GM should just get the book without coloring. He should have equal access to the games comprising the book, but it's not fair for him to know which choices Rybka will make against each move he might choose. Of course Rybka could easily give this handicap, perhaps by playing with an uncolored database herself, but you are trying to propose advantages that are not clearly a handicap. As for TBs, equal access is okay there, it's probably only worth a few Elo points, although some CCRL data on two programs indicated an average value of 30 Elo, much higher than other estimates.
     So if we give the GM a board to analyze on, an identical but uncolored opening book, and equal TB access, I think we are giving him something like fifty Elo points of advantages that might not be considered a handicap. This value is debatable, but it's very unlikely to be outside the 20-100 range. Anyway, it's nowhere near enough to give the 2800 superstars any more than drawing chances. To make it at all equal without departing from standard chess rules, we would have to give all White pieces, no ponder, and huge time odds, maybe 30 moves in three hours vs. 5' blitz.

This idea certainly appeals to me! I'd gladly pay some ducats to see any top-grandmaster figure out his moves that way live!

It would not only be more aesthetic, but also more instructive for the spectator than any form of handicap-play.

The only thing I'm afraid of, is that the difference in strength would still be very large. What boost in playing strength could you expect, if the human has access to an openings book, endgame tablebases and a second board to work his moves out? Say 100 elo for the book, 50 at the most for the tablebases and perhaps another 50 for the visual aid? That's 200, on top of say 2800 for the strongest human player these days, and there's still a huge gap with Rybka 3's projected 3150 elo ...

But nevertheless, I for one would like to see this idea in operation!

humans tend to lose to engines due to blunders. with a team effort these blunders would occur significantly less. therefore i think the GM team would emerge the victor. (assuming they were super GM's)

Chessbase posted an interesting article about the match today (written by Larry):

The Dzindzi – Rybka 3 Handicap Match