AGW is about as solid as estimating the results of a chemical reaction in an uncontrolled environment where you know only some of the reactants involved.
What you give implies a warming trend, but trying to account for everything which affects surface temperature to the point where you can give a meaningful number leads to error bars which include both deep freeze and boiling water (although I admittedly recalculated those over ten years ago, I am not aware of any major developments which would warrant redoing that.)
I understand it is the best we can currently do, but our best does not yield an answer most people would consider meaningful.
An honest answer would be something like "What we know implies a warming trend, but we are incapable of putting a number to it at this time." Claiming the kind of certainty the IPCC does is very dishonest, which annoys me enough to be willing to argue it.
And to answer your question:
The energy goes back to space, but this is delayed more than it normally would be if it is absorbed by a CO2 molecule.
The energy absorbed by a CO2 molecule will be emitted as a photon after an average of about ten microseconds (collision rate and therefore pressure will affect this.) The wavelength of the photon depends upon temperature, but due to the very limited absorption spectrum of CO2 it is unlikely to be absorbed by another CO2 molecule. H2O is however far more likely to absorb the photon, and we have a lot more of it in the atmosphere. The potential problem comes about from the interaction of both CO2 and H2O (a very small increase from CO2 amplifying a much larger effect from H2O.)
The reason increasing CO2 has a noticeable effect is due to its small concentration in the atmosphere, as it will not yet absorb the wavelengths it can to extinction.
As concentration increases it becomes "less bad" to increase it further. It should be kept in mind that increasing CO2 concentration will give a logarithmic falloff in absorbed energy. If you increase the concentration of a gas by 1000x, you will get about a 7x increase in absorbed energy.
In short:
We probably will see some warming, but are very unlikely to see a runaway greenhouse effect. A potential future problem which bears some watching is being pitched as a doomsday scenario, and I would consider this is a good example of the phrase "making a mountain out of a molehill."