What If You Gave A Party, And No One Showed Up?

The Inquirer reports that AMD gave a party to unveil the FX-60 that was so secretive that almost no one showed up.

They were so much into playing Secret Agent Man that they even used a code word for the event.

Perhaps a little etymology lesson is in order here. “Publicity” comes from the word “public,” and inherent to the concept of public is the idea of people.

If you need more help, the word “secret” is considered an antonym of the word “public.”

In the great scheme of things, this is a soon-to-be-forgotten trifle, but it does illustrate a big weakness of AMD: their obsession with secrecy.

Want to see a current tech document? AMD used to update when a new product showed up, but these days, from what I hear, you have to sign an NDA for it. Sometimes they get around to eventually updating the ones available to the public, sometimes they don’t.

Let’s say you have the misfortune of being your brokerage firm’s analyst for AMD, and you’d like to know how many processors AMD sold and for how much so you can do your spreadsheet. AMD used to tell people items like that, but now it’s a state secret. Furthermore, it’s a stupid state secret because anyone with mild math skills can approximate the figure from third-party sources.

Some may say, “Well, Intel doesn’t say so.” Well, AMD doesn’t earn a steady two billion or so profit every quarter from its CPUs, either, nor is the default CPU manufacturer. For Intel, such numbers are less important than they are for AMD, whose numbers can and do fluctuate a lot more.

Some may say, “AMD is fighting for its life against Intel; it can’t afford to be open.” Well, I’m not suggesting AMD post “Tips on how to make a 65nm SOI chip” on its wehsite, but do you really think Intel can’t figure out how many watts an Athlon 64 available for public sale chews up just by buying one? We know Intel dissects and reverse-engineers AMD chips, they’ve said so publicly in the past.

Might I suggest that anything Intel would really want to know about AMD’s chips isn’t something that would be found in any public documents, anyway?

So who are they hiding from, and what are they hiding?

The answer to the first question is easy: us, meaning the public.

What are they hiding? Well, unless somebody writes a tell-all book someday, we won’t know 100%, but if you have an attention span/memory of a few years rather than a few minutes: the most likely suspect isn’t any great bolt-out-of-the-blue achievement, but quite the opposite.

Not being able to make chips on time (Palomino, Hammer). Not being able to make them fast (Thoroughbred A, Hammer). Not being able to make many of them long after product introduction (again, Hammer). Sudden 180-degree turns in strategy (socket 754/939). There’s been plenty to hide.

Not that Intel has exactly been Honest Abe all the time the last five years, they’ve run two CPU designs into the ground, and before then, there was the 820 fiasco, but in all these cases, Intel ‘fessed up sooner or later.

Maybe the best way to put it is after years of observation, when faced with a big problem, I’ve found that Intel handles it straight and does the right thing right away sometimes, while AMD seems to always look for the slick way out.

It’s “trust but verify,” versus “OK, where’s the loophole this time?”

Yes, people have short memories, and AMD has labored in relative obscurity insofar as the mainstream media is concerned, but should AMD become a lot more successful the next few years, and become Pepsi to Intel’s Coke, the culture of secrecy isn’t going to cut it when the first big problem hits.

(And that first big problem may not be a faulty chip design or manufacturing problems, but the collateral damage from that lawsuit AMD has against Intel. Believe me, Intel will be trying to shed light on all of Green’s secret stuff then. 🙂


Be the first to comment

Leave a Reply