For all intents and purposes, that’s what an email sent by Intel’s CEO Craig Barrett to Intel employees said.
I wonder if he sent a copy to himself. He should have (along with the rest of his executive officers).
The last few months, just about everything seems to have gone wrong with Intel.
For those of you around awhile, if you get the feeling you’ve seen this before, you have, about four years ago. Remember the partial-birth abortion of the Pentium III 1.13? The 820 chipset recall? The approaching fatal attraction to RDRAM?
This is deja vu all over again.
The Real Problem
Why is this? Why is history repeating itself?
Does it seem very likely to you that Intel employees just get stupid every leap year? I didn’t think so.
There’s a far better explanation than that.
Look at the two clusters of events. What do they have in common? In both cases, the culprit for failure most of the time pushing a generation of technology until it broke. Generally, it’s trying to stretch a particular technology set just one more generation, just one more year.
Then the rubber band snaps. The people holding the rubber band, or the bosses telling them to stretch it some more?
Usually, the breaking part is not a bolt out of the blue. There are usually warning signs, and frankly, if the people steering the ship don’t listen to the crew telling them that the boat really needs an overhaul before the next big voyage, who’s fault is that?
Take Prescott, for instance. The core problem with Prescott is its circuits leak electricity like a fractured pipe leaks water.
This is nothing new. Intel’s processors just didn’t start springing a electroleak. There was a dramatic jump in leakage when Intel went from the PIII to PIV generation, from about 10% of total power to about 40%. That’s no great secret, Intel even put that out in PR material.
They knew they had a problem. They knew they had to do something, and some big about it. But they tried to stretch the old stuff, just another generation, just another year or two.
The rubber band snapped.
Who’s fault is that?
Intel and 911
Recently, the National Commission on Terrorist Attacks Upon the United States (also known as the 9-11 Commission), released its report trying to answer the question, “How could this have happened?” It essentially says that those in charge the last number of years didn’t pay enough attention to the warning signs.
Intel needs its own 911 Commission, because it looks like the same thing happened here.
Telling the crew to shape up does absolutely no good when the captain keeps aiming the ship at shore and ignores the rocks.
There could be two problems here. The first is that those reporting to the captain are afraid to tell him he’s running the ship aground. The second is that the captain thinks that if he aims for the rocks, the crew will jump overboard, swim ahead, and remove all the rocks before the ship gets there. (Or perhaps the captain thinks the rocks jump out of the way all by themselves when they see who’s coming.)
Granted, it’s not quite as easy as that. The crew probably wants to play safe-rather-than-sorry to such a degree that the good ship Intel wouldn’t even get within sight of the shore. Left to their own devices, most are probably challenge-challenged.
Nor does the crew have the only say. Stretching a technology a little bit more saves Intel a lot of R&D money, and Wall Street types love penny-pinching.
So a captain of a ship like this has to push or he’ll get nowhere, but if he pushes too much, he’ll end up shipwrecked until the crew can cobble together a new boat.
That appears to be Intel’s real problem, and until Mr. Barrett and those around him stop saying, “You suck” to the employees and start saying, “We suck,” they’re not going to fix the root problem here. They’ll just keep crashing into the rocks, spend a year or two building a new ship, sail that awhile, then crash again.
Those who do not learn from the lessons of history are doomed to repeat them.
Be the first to comment