I work in Intel’s Itanium microprocessor design group. Our most recently completed project was the Tukwila processor. Intel’s Itanium line has not had a lot of positive press in it’s 10+ year history, due to delays, underwhelming performance and slow market adoption. Tukwila looks much better so far.
These are enterprise-grade microprocessors, not like your standard Pentium or Core Duo chip. This baby packs 2.05 billion transistors, the little switches that control all the logic, store memory and such. For comparison:
Intel 4004 (1971): 2300
Intel 8088 (1979): 29,000
Pentium (1993): 3,100,00
Pentium 4 (2000): 42,000,000
Core 2 Quad (2006): 582,000,000
Tukwila (2008): 2,050,000,000
I’m a geek, I think that’s cool.
My job, specifically, involves writing software used to design and test the billions of tiny wires that connect all those switches together. At least, that’s one part of my contribution to Tukwila.
Each one of these projects lasts 4-5 years all together, though I’m only involved in the first half of that (design before it gets manufactured). After the first manufacturing samples — actual silicon parts we can plug in and test — my focus shifts to the next project, Poulson in this case. Poulson is my 5th major project in the 11 years I’ve been with the group.
I actually rather enjoy the work I do. I work with great people on challenging projects, doing a combination of long-term strategic development and daily problem solving. The challenges change frequently as we try to push the envelope.
Only problem is, I don’t see myself doing the same thing for another 20-30 years. So now I’m trying to figure out what’s next… Stay tuned for that!
This is really cool: a 360-degree 
