Then, simply put, you should go read the transcripts of what he’s saying and think about it until you’ve understood the points contained therein. His key point - the one that underlies everything - you don’t even mention. Although you do mention something which supports his argument, amusingly :).
Summarizing poorly: when you make a game, you have to compete. Competing means equivalent-to-your-competitors’: graphics, gameplay, quality, depth, etc. Doing this cost money proportionally. You have to provide reasonable chance of recouping that money and making more than it cost, otherwise you will not be in business for long. The PC games business is competing with the console games business. There are > 120 million playstations out there, every one of which has a high chance of buying any game you put out there. There are approximately 900 million PCs out there, which on average each have a much lower chance of buying a game you put out there. Intel graphics chips account for between 20% and 60% of the section of the PC market which has a high-ish chance of buying games (according to games industry reports and internal marketing-research depts). To compete against consoles, you literally cannot afford to not work on intel chipsets.
However, each year intel chipsets increase in performance by around 5%.
Each year (averaged), console chipsets increase in performance by around 20%.
Each year, non-intel gfx cards increase in performance by around 25%.
So, if Intel stopped selling crap, the PC games business would be slightly outstripping the console games business in performance (and hence geting the cost for writing PC games down a lot, making them a lot more easy to develop and sell profitably).
Instead, Intel’s sales are gradually destroying the PC games business.
I’d hope Mark would find that a not-too-bad summary, but errors and ommissions my own.