Thinking about Moore's Law

I was reading Randy Mears at the EDS blog.

"I have often thought that focusing on Moore’s Law may be too single minded. When it comes to barriers in IT, it may turn out that the issues won’t be with hardware advancement but rather with fixing software technologies."

Actually he is missing out on 2 of the biggest spend categories in IT - services and telecomm. Our industry's productivity is hugely driven by performance in  those 2 major areas.

As I have asked before "what if Intel was a services vendor?"

Can you also imagine if we had an Intel like player in place of a Verizon or a T-Mobile how much faster and cheaper broadband, cell service etc would be?

Our software, services and telecom sectors may never deliver to the high Moore's Law benchmarks of doubling performance every two years, but is 25% to 50% improvements every two years too much to expect?

Update: Moore's Law is not dead yet. Scoble on Intel's latest breakthrough.