More importantly, I don't think NVidia is far enough ahead of AMD to make this a smart move. Their Linux drivers still have a better reputation compared to AMD, but it seems to me this advantage is eroding. Pissed customers might simply get an AMD next time.

By comparison, Intel can get away with similar crap in the CPU world (no ECC RAM support except on Xeon processors and "workstation" chipsets). Their lead over AMD in CPUs is big enough that most people will swallow the bitter pill and buy Intel anyway.

If you compare benchmarks where multiple generations of GPUs compete against each other, such as the Passmark benchmark, later AMD GPUs seem to have a better ratio of benchmark scores to theoretical computing power (as given on http://en.wikipedia.org/wiki/Comparison_of_AMD_graphics_processing_units )

Examples:
Radeon HD 3850: 427.5 GFLOPS, Passmark score of 532
Radeon HD 3870: 497.3 GFLOPS, Passmark score of 744
Radeon HD 4850: 1000 GFLOPS, Passmark score of 1043
Radeon HD 4870: 1200 GFLOPS, Passmark score of 1361
Radeon HD 5750: 1008 GFLOPS, Passmark score of 1399
Radeon HD 7750: 922 GFLOPS (at 900 MHz), Passmark score of 1624

The 38xx surprise by bucking the trend - maybe some AMD developer had a bright moment there? But in general, drivers for current cards seem more efficient. In the 7750, the change in architecture may have helped.

For Linux in particular, the open source drivers are gradually getting closer (at least to the AMD Catalyst driver). For some older and presumably simpler games, the reviews on http://www.phoronix.com/ already show 80% of the performance of Catalyst. In other, more demanding, tests they still suck but the long term trend is encouraging.

Ahem. Not even close.

They have been porting Steam and some games to Linux, and now they are doing a Linux-based gaming PC. That are some of their activities, not all. And most of those are still closed source. If I am mistaken here and there is an Open Source Steam client, feel free to prove me wrong by pointing out the repository ;-)

This said, the whole Steam Box project might bring some very welcome improvements in Linux driver support. Actually that is the one field in which there are some reports about Valve contributing to everyone's benefit:
It seems they worked with Intel to improve their graphics drivers :-)

Not according to Michael Larabel on phoronix.com.

The guy frequently tests the latest Open Source drivers for AMD's chips, and sometimes nouveau too.
While AMD Open Source drivers are still outperformed by Catalyst, they are getting closer. Comparisons of nouveau to the nVidia binary drivers show a much greater performance gap.

As an occasional linux user I tried Mint Debian Edition with xfce some months ago, but was not entirely happy. My general impression was that maintanance had stopped shortly after the release. Latest kernel version from the Mint repository was 3.2.0.4 when kernel.org was around 3.2.0.38, and configuring the system was not entirely painless.

My next attempt was with xUbuntu, and it seems to be much better maintained while it also got rid of the Unity GUI (which I dislike). So that one is certainly worth a try :-)

I think the low wattage trend is most important in mobile, because there it directly translates into battery life. And yes, ARM is making inroads there. Intel also tends to offer more performance per watt.

For desktops that don't need huge performance, AMD has a fairly good product in its APUs. Wattage is not quite as important there. But as Intel's integrated graphics get better, AMD is coming under pressure there as well.
Now reviews of APUs vs. discrete cards and CPUs show that the APUs tend to be bottlenecked by memory bandwidth, so I wonder if AMD might do better with a PC version of the chip they built for Sony's PS4. I'm thinking of a small form factor mainboard (Mini ATX? ITX?) with soldered in APU and 8 GByte of GDDR5 RAM. Which would obviously not be upgradeable, but should be quite sufficient for the sort of PC their current APUs are typically used in. Compared to the PS4, that board would have to offer more USB and SATA connectors, plus maybe 1-2 PCIx slots for a bit of extendability.

In the gamer market, it seems that most people don't care that much about power consumption. I do, but I'm the exception among my friends. For most it is about framerates and higher framerates. AMD has a problem there with their CPUs, because Intel has a huge advantage in performance. But in the GPU market they are fine.

I hope AMD can survive on the console business for now and close the performance gap to Intel with the next version of their Bulldozer architecture. Because if they don't, they will disappear from the x86 CPU market eventually.

I think a proprietary Amiga would still have had better chances than a proprietary IBM PC. If both IBM and Commodore had kept their systems proprietary, maybe a third party would have appeared and displaced both.

In the late 80s, the Commodore Amiga 500 was quite popular in my circle of friends. For a while it seemed uncertain if the "PC" or the Amiga would become the next dominating platform for private users. The "PC" won, but I doubt that would have happened if there had been only original IBM PCs available. The price would have driven most people away.

I got a "PC" in 1990 myself, but it was one of the cheap knockoffs. An original IBM would have been waaay beyond my budget ;-)

After using Win7 for almost two years, I find it a marginal improvement.

Good:
-UAC is a good feature for those who understand the concept of "administrator" vs. "normal user".
-A lot of minor things (like "ejecting" USB sticks) that I found unreliable in XP work well now.
-64 bit support that deserves the name (Windows XP 64 never really took off).

Bad:
-Low level system settings are hidden deeper than in any previous version. You can still find them with a bit of Google help, but for people who are need to manage stuff like IP addresses and subnet masks directly I find Win7 actually harder to use. BTW that trend started with XP if not earlier...

Overall, I see the technology improving but Microsoft trying harder to keep the user from tweaking the system, often to the effect of annoying people who need special configurations. On the bottom line I consider Windows 7 an improvement, but one that could have been greater.

End of XP support is announced for April 2014. And yes, it has been extended before. If Microsoft is serious this time or if they will give in with another support extension is anyone's guess.

First, TFA is about a relatively small giveaway by the city of Munich, Germany. Nobody offered to provide free Ubuntu DVDs to large parts of Asia.

Second, Canonical is a British company, owned by Mark Shuttleworth who is a South African.

Third, locking people in with a version of Linux would be a lot harder than with Windows. Because you can get most, if not all, of the system from another distributor. Legally too (if you don't care about that, Windows is also available from "other distributors", at a very low price ;-)

By April 2014, Ubuntu 14.04 should arrive as the next LTS version.

I would certainly want that (or a derived distribution) rather than Ubuntu 12.04, especially in a PC with AMD graphics. The open source drivers for ATI/AMD are still catching up in features and performance, and 14.04 vs. 12.04 should make a significant difference.

I tend to agree with you, and those are only the faults of the project management that is directly working on a given product.

But if you think past the job of the project manager, management also includes hiring the right people for the job. There will be honest mistakes of course, but if a large company like EA repeatedly releases crap, the problem is not just one or two bad developers that were hired by accident.

In this case, the hiring process seems to suck at some point. Either they are unable to find competent developers, or they tend to hire bad managers.

How do you implement an exchange of encryption keys that is
a) convenient for the users and
b) resistant to man-in-the-middle attacks? Attempts by the NSA to install those at your ISP can be expected.

Sincere question, as I'm not an expert on the subject. But I believe both a) and b) need to be solved to gain widespread acceptance for the new protocol.

Makes sense.

The right time to switch would be when a non-trivial change to the applications on the VMs becomes necessary anyway. At that point, you might as well go all the way to port the applications to some other system (Linux?).