I don't buy this claim that innovation and engineering accomplishment was greater during the tech bubble:
"But all the past figures assumed something that is no longer true- that tech companies would innovate and release the same number of new products, with the same schedules as they did during the bubble. I think we all know that isn't happening. So if you don't release any new products or you cut down the schedule considerably as most software companies have done, then you don't need highly paid engineering talent, very true."
Here's why I don't buy this:
-- during the tech bubble, Intel was having the same kinds of problems absorbing vast numbers of new hires, and recruiting the best of those graduating, that all companies were having.
-- the tech bubble just happens to also be when Intel was having some very highly-publicized screwups and delays, a la the 1.15 GHz PIII, a la the Timna chip, a la the Rambus delays, a la the competition with Athlon, a la other delays and missteps.
-- counting specific chip releases is misleading for multiple reasons. In any case, various iterations of both P4 and IA64 have been appearing in the past 2.5 years of the tech wreck.
-- Intel is making fewer missteps now than it was making in 1997-2000, from everything I can see
-- innovations like hyperthreading are doing very well, and Intel leads in the highest-volume segment of the server/blade/rack/Beowulf market
-- in particular, the transition to 0.18, and then to 0.13 micron, and to 300 mm wafers in multiple fabs, and to other new process enhancements, seems to to be going very smoothly. I expect Intel to be the leader in volume production of the Whole Enchilada: 90 nm, SOI/strained silicon, copper/etc., 300 mm, automation, multiple fabs.
(By contrast, several large competitors are stumbling, delaying plans, seeking investment help from....Intel!)
" I know intel has a small number of new products coming down but this is nothing like the 90s when there were 2 chips released every 6 mos it seemed... I know, I had to deal with all the new SKUs at Dell, it was a nightmare. For a while we had a pentium pro, pentium and then the celeron line, all this required multiple teams. I think intel must have cut significant staff already. "
You seem to be making two wildly different claims here:
-- that you saw a "nightmare" with multiple SKUs of Pentium, Pentium Pro, Celeron, etc. at Dell.
-- but that now that Intel has fewer lines (a claim I don't necessarily agree with!) it must mean that Intel is innovating less!
I attribute the smoother product announcements of the past two years, especially the past year, to the learning curve: the vast numbers of new hires in the late 90s have now had a few years to get oriented and educated, the various divisions have had more experience working with each other, and Intel has had fewer staff losses to other companies (a big problem during the tech bubble, when it often seemed that Intel would be hard-put to recruit tech staff when so many of the dot coms were offering such substantial stock packages....this has now ended, and things have calmed).
In summary, I reject your confusing claims about how innovation at Intel is slowing. If you think the "nightmare" SKU situation at Dell was some evidence of greater innovation, you have a strange notion of what innovation really is.
--Tim May |
