Is This An Intel Bigot?
John Dvorak has been a columnist for various PC magazines going back to at least the mid-eighties. He has a pronounced habit of being as provocative as possible, and to a large degree has been a computer gossip columnist.
This article makes some positive points about AMD, and a couple negative ones.
Here’s some of the positive statements:
“According to an interesting memo sent out by the analysts at the Intel Developers Forum, Intel has a huge hole in its upcoming road map — one that AMD can drive into to make hay.”
“AMD achieved nirvana with the Athlon line and is in many ways ahead of Intel . . .”
So we don’t exactly have an AMD-hater here.
He’s not exactly praising Intel to the skies, either:
“In his book, Inside Intel: Andy Grove and the Rise of the World’s Most Powerful Chip Company, Tim Jackson relates a creepy Grove quote. Grove tells an employee, who is innocently quitting the company: “You will leave no heritage for your children. Your name will be forgotten. You will fail. You will fail in everything you do!” What is the point of an executive at Grove’s level telling anyone something like this? It’s just plain mean. . . . [Y]ou have to wonder about him. The arguable long-term problem you have with a guy like Grove is that he creates a corporate culture that may not be a healthy corporate culture.”
What we have here is somebody edging away from the Intel-centric world he’s lived in most of his professional life.
Not entirely, though:
“Let AMD rule the roost for a little while. Besides, why worry? AMD has never shown any leadership. The company has always been a follower. . . . The company only toys with original architecture, instead of being original.”
“They make modifications on Intel ideas but nothing quite so unique as what you find at, say, Transmeta. Even with Intel stalled, I’m convinced that AMD will not blow by them but will instead wait for Intel to map the future. I think AMD is honestly shocked by the fact that it is now ahead of Intel in this game.”
During the 1964 election, a Barry Goldwater campaign aide once told a member of the press, “Don’t report what he says, report what he means.” I think you need to do that here; don’t let facts get in the way of the truth.
What’s AMD’s record?
Until a few years back, they spent their time copying Intel, no ifs-ands-or-buts about it.
For a variety of reasons, legal and otherwise, this wasn’t working out too well, so what did they do? They bought a company called NexGen, and built the K6 and K6-2 around it.
They needed something better than that, so what did they do? They imported a good sized-chunk of the Digital Alpha development team. The Athlon generation of AMD processors is basically an x86 Alpha.
Ahhh, but what about Clawhammer? Yes, what about Clawhammer? Clawhammer is merely a modification (and not a big one) of the old x86 platform to run on 64-bit, as opposed to Intel/HP’s decision to make a clean break from that antiquated standard.
The issue is not if we need to break away from the x86 standard, the issue is when and how.
Intel is saying with Itanium/McKinley “the time is now.” AMD is saying with Clawhammer, “Not yet.”
This is not exactly a glorious trail of innovation.
Copy, Improve, Innovate
This may hurt some feelings, but there’s considerable merit to the point that AMD is no great innovator. There’s nothing wrong with AMD doing what it’s doing; they just can’t keep doing it forever if they want to replace Intel.
It’s natural to first compete against a Goliath by copying him. It saves you a ton of development money, and gets you most of the way up the curve pretty quickly.
However, mere copying can never get you all the way up the curve. You’ll always trail the innovator, simply because you’re waiting to copy his next move.
The next step is usually modifying and improving the current designs, and that is what AMD is doing now. One can be very successful doing just that, look at Japan. Many consumer electronics technologies were invented in America, then came to fruition in Japan.
In all fairness, AMD would be somewhat constrained in any case by the x86 standard. People want x86 machines because that’s the standard; you innovate too much, you don’t have an x86 machine anymore.
Clawhammer vs. McKinley (Itanium is a dead duck) may look like a big change from business-as-usual, and in some ways it is. For the first time, AMD will offer an alternative to the Intel standard. However, what they’re offering is the old standard that Intel is trying to move away from.
This may well turn out to be a smart move on AMD’s part. It may prove to be brilliant. But it’s not innovative.
You can claw your way to the top by just coming up with a better version of what the innovator does, but once you get there, who are you going to copy from? Mind you, the Evil Empire is likely to stop thinking about Bunny People and start thinking more like “The Empire Strikes Back.”
If AMD manages to knock Intel off its pedestal with Clawhammer, they’re going to have to do the same sort of thing Intel is trying to do now, and sooner rather than later.
This is what Dvorak is talking about; can AMD mentally shift from “follow the leader” to leadership? He doesn’t think so, based on AMD’s track record. I think he’s far too pessimistic dismissing AMD out of hand, but there’s certainly reason to have doubts about it.
The issue is not “it can’t be done;” of course it can. Japan is a great example of the progression from copying to improving to innovating. The question is “Will it be done?” and there’s reason to wonder about that down the road.