The speculation about the 8600M failures is interesting, but it is probably not related to this problem.
The failure we are seeing is related to intermittent operation of the LED BACKLIGHT, not the LCD display itself. These are two logically separate components (even though they function together as a unit). I admit that I haven't taken my MBP apart, but I would speculate that the backlight is *not controlled by the GPU*. Even if the nVidia chips are bad, this might explain some of the scrambled screens and weird scrolling that are reported in other threads...but I find it unlikely that it is the problem here.
Apple uses a bunch of different types of backlighting in its MBP screens (if you count the 17" models). The GPU also needs to be able to handle a wide variety of displays and screens (since it's not just used in Apple products). The GPU should internalize the common subset of tasks that ALL displays need (figuring out how to render graphics instructions into pixels, 3D operations, BITBLTs, etc.), but the DISPLAY-SPECIFIC components (how to initialize display model ABC, how to enable backlight, and so forth) are all likely different from display to display, which makes it unlikely that they are handled directly by the GPU. I would be willing to bet that the backlight is actually controlled by dedicated circuitry (perhaps part of the Macbook's power management hardware and/or firmware).
Furthermore, if the backlight control were somehow handled by the GPU and if it were really failing, we would likely be seeing these failures in many other non-Mac machines in high quantities. So far, we've only seen a tiny handful of reports of supposedly-similar issues on PC hardware, and it's hard to say if it is even the same problem. This points to the source not being a bad batch of GPUs.
In short, while this article is cause for concern, I don't think this means that we're screwed, nor does it necessarily mean that a firmware fix is impossible.
