It's about the idea that a "skeptic" leaves the door open to all possibilities, but "deniers" have already closed the door.
From what I have seen of the people who question that AGW is 100% proven and undeniably going to cause the end of all mammalian life in less than a century or so at least on slashdot you are beating on a strawman.
I do not consider AGW theory to be even remotely proven and yes I have looked at the so called "evidence". However I would never argue that AGW is impossible. That would be silly. The basic mechanism is sound, but it might take 10,000 years or 100,000 years or the small effect may be overwhelmed by other factors.
The climate of an entire planet is extremely complex and cannot be accurately 'modeled' with a naive computer program. Just like we cannot build a model of the universe in a computer and use it as a substitute for testing our hypothesis in meatspace to see what happens. Science is about testing hypothesis by actually trying stuff out to see what happens. Testing ideas against nature itself. Testing ideas against naive assumptions inherent to computer simulations is not science. When it comes to climate "science" this would mean waiting to see what happens.
At least so far the effect does not seem extraordinarily great. In a few hundred years we can again take measurements to see how things are going. Or however long it takes to see a noticeable and arguably dangerous warming effect. So far the rise has not been dangerous. There is nothing inherently bad about a rise of 1 degree over a century. Just as there is nothing inherently good about a drop of 1 degree over a century. Nor is a 100% perfectly stable temperature inherently good.
The theory is plausible. The effect is plausible. Now AGW just needs to be actually demonstrated in meatspace. It's just a matter of being patient. There is no shortcut when trying to test long term theories. You have to wait. Writing a computer program to accurately mimic a natural process is not a shortcut. At the very least first you would have to demonstrate that your model is 100% accurate and again that requires waiting many years to compare the model's predictions against observations.