Just days after Apple officially stated that there was "no news to report" on the CEO search (despite a Heidrick & Struggles exec saying that an appointment was likely within 30-45 days), Mac OS Rumors comes through with the rumor that heading up Apple's short list is James A. Cannavino of IBM. Jim is a Senior Vice President at IBM and reportedly had his hands in the PowerPC's debut via the Apple-IBM-Motorola alliance, as well as in the stillborn CHRP standard. So he wouldn't be a newcomer to the Mac world, per se.

But what about the biggest obstacle in Apple's path to enlisting a permanent CEO-- namely, its interim CEO, Steve Jobs, known far and wide as a control freak with zero tolerance for differing opinions? After so many others have passed on the CEO position for precisely this reason, is Jim Cannavino willing to work under Steve's watchful (and critical) eye? Perhaps-- and perhaps it may not be necessary. Rumors is hinting that Jobs may be "planning to take a smaller role" in running the company when (or if) Jim signs on. Whether that "smaller role" includes a resignation from the Board of Directors is a matter of speculation at this point.

So let's say Jim does take the Job, and Steve quits the board. Will we ever know if Steve's disappearance was a requirement made by Jim? Or is Steve just getting tired of the Apple game?

The rumors of Sun's imminent buyout of Netscape continue to spread unabated, despite a certain CEO's recent foray into stand-up comedy. According to a CNET article, Scott "Just Call Me Letterman" McNealy had a top-ten list of his own, which he unveiled at a Churchill Club dinner a few days ago. In it, he takes "pot" shots at everyone from Microsoft to Compaq to Adobe to, uh, Microsoft. It's worth taking a look at if you're in the mood for a chuckle.

In his routine, McNealy jokingly attributed the rumors surrounding Sun and the troubled Netscape to "second-hand marijuana smoke," borrowing a page from the Olympic snowboarder who was disqualified last week following drug tests. Also on McNealy's List of the Drug-Induced: two-digit date fields, Compaq's buyout of Digital, and any number of Microsoft products, such as Windows 95 and the now-legendary Microsoft Bob, which Microsoft has tried desperately to suppress any and all trace of, in a revisionist attempt to regain its credibility. (If you search the Microsoft web site, for instance, you won't find a single reference to the product, even in the service areas.)

Not that any of this actually addresses the rumors of the Netscape buyout. When asked about it point-blank, McNealy simply said that he "couldn't comment on rumors, no matter how accurate or silly they may be." Not exactly a ringing denial, is it? But truth be told, after all the rumors of a Sun buyout of Apple, we're a little tired of Sun buyout rumors. Who wants to help us start a new rumor, with a more interesting player? How about we convince people that Netscape is about to be bought out by Ben & Jerry's? We could say that the takeover will be marked with the release of a new flavor: Navigator Nougat. Ehh, just a thought. Don't mind us, it's all this darn second-hand marijuana smoke...

Another day, another Pentium flaw. But this one's a little different-- it doesn't just affect the Pentium, but every x86-based chip in existence. A Newsbytes article reports that a company called "Intelligent Firmware" has uncovered a serious performance bug that could slow down an x86-based chip by as much as 70%. A full description of the problem, as well as a demonstration, is available at Intelligent Firmware's web site.

Apparently the bug is triggered when a large amount of data is read into the processor at once-- which is, unfortunately, exactly the way that Intel tells programmers to do it. A workaround appears to be to read in the data in nonsequential chunks. Programmers can rework their code with this method and avoid the performance hit. If the bug had been documented early on, then compilers for the x86 could have used the same workaround.

The interesting thing is, this flaw actually makes us at AtAT more impressed with Intel. I mean, we're certainly not electrical engineers or anything, but everything we've seen indicated that Intel's done a hell of a job squeezing every last teensy bit of performance possible out of its antiquated x86 architecture. (Remember how the RISC-based PowerPC was supposed to completely smoke the CISC-based Pentium? The fact that it's taken this long to really start happening is a testament to Intel's genius.) So now we find out that in many cases those Pentiums are actually running slower than they should be? Hmm. That's mighty impressive indeed. Of course, we suppose it'd be more impressive if the bug weren't there in the first place, but hey.