Many of the times I've gotten this it was issues with the GPU, a bad/low-quality PSU feeding unstable power to the GPU (uncommon), a faulty/bad PCIe riser cable (uncommon), or bad power cables/motherboard slot causing unstable power to the GPU (rare).
I want to emphasize the PCIe riser cable specifically because it's not mentioned in the article. In two instances I was chasing my tail on this error for a while until I removed the riser cable (Thermaltake) from the equation.
I've had this happen on multiple GTX 460s (defective parts), a GTX 660 (insufficient cooling + defective parts + insufficient power), and a GTX 970 (insufficient cooling + defective parts). For these they had issues with the GPU itself. The GTX 970, for example, would constantly throw this error as soon as the GPU was under load. Turns out it has an issue with a VRM MOSFET causing a cascade failure of the entire voltage regulation circuit. It requires I have the fans ramp up speed very aggressively above 55C GPU core. The GTX 660 actually had both a bad PSU and bad VRM. After replacing the PSU the issue was greatly reduced but still happened. Setting the GPU fan to aggressively ramp up speed and a reduced power target fixed it.
I have had a corrupt Windows install cause this (SFC is a waste of time).
So, technically most of these were covered by the article but I wanted to give more specific examples to possibly help others getting this issue.
I'm waiting for when we have a version of Windows or any OS for that matter that will just ignore all errors and crashes and just keep on going without any interruptions 😀 It would only log the errors or crashes then tell you later in a pop up message but it would not interrupt what your doing at moment until your done. 👍😉