One of the major fears every company faces today is patent trolls. Regardless of whether you are in the right or not, a patent infringement lawsuit will be an immense cost to fight. Opening up internal source code or having public record of which tools/libraries you use by contributing back to them significantly increases the attack surface for patent trolls. Many companies, like Netflix, are willing to take on that risk. Many other companies are not willing to take on that risk.
I don't know anything about AOL specifically, but I've worked for a handful of BigCo's. All have required signing some sort of agreement at the start of employment which explicitly forbids leaking confidential information. So it's very likely that this person was under an NDA. What may be surprising to people who haven't worked at a BigCo is that due to legal implications, often times companies actually have guidelines about what should be discussed in recorded formats versus in person/phone.
Windows internals are very different than Linux and 32bit DLL files are not SO files. 32bit DLLs are typically relocatable (as opposed to .so files being position independent), which means they are compiled with a with a preferred load address they must be loaded at. If that address is not available at link time, the code must be moved to a different address and jumps must be fixed up to compensate. Because of that, in practice, loaded libraries are often not able to be shared between processes in memory. The reason for this decision is that pic code requires an indirect jump through an offset table which adds extra processing overhead. 64bit Windows is implemented closer to Linux style .so files due to the addition of new pointer relative addressing.
Before we get out the pitchforks, Google is still supporting the current Microsoft formats (.docx, .xlxs, .pptx). If you're still running Office 2003 or older, you can install a compatibility pack from Microsoft to support the current file formats. (http://support.microsoft.com/kb/924074)
I read the article. Out of 4 images shown, one was of the live demo and 3 were suspected doctored photos from Nokia's promotional materials. The article doesn't clarify whether the live image taken with the Nokia phone was taken in front of the reporter or if it was taken ahead of time. Without that knowing that, it's difficult to gauge whether the conditions were truly similar and if this was a fair apples to apples comparison.
I like the way you've phrased this, so let me give a contrived but quasi-realistic example: I have a network of nodes which calculates and returns some floating point value. I then accumulate those in the arbitrary order in which they come back from in the network. (According to Google) 1111.0000000000001 * 2 * 1111.00000000000001 = 2468642 BUT 1111.0000000000001 * 1111.00000000000001 * .2 = 246864.2. By using floating point we've already decided it's OK to give up some precision, however by applying the commutative property of multiplication now we've also given up determinism.
What's likely going on is fairly complicated to delve in to in a comment thread like this. However one possibility is that Linux just has a terribly hard time scheduling with that many cores. Processes attempt to maintain locality up to a point, but tend to move around between cpus when another has more free time. Moving results in a cold cache needing refreshing, and that significantly slows down work.
FTA: "It's actually a hybrid that is largely built on Harmattan, the legacy Maemo 6 code base that Nokia shuttered when it committed to MeeGo. It seems sort of dubious on the surface to call the software MeeGo when it's really still Maemo, but the hybrid is apparently designed in such a way that it has full API compatibility with MeeGo 1.2."
(Also not a lawyer)