标签: scientist

I recently read an analysis column " Watch Out Intel, Here Comes Facebook " in Barron's which offered the premise that Big Web operators like Facebook and Google are demanding radically different chip designs rather than the traditional architecture typified by Intel's CPUs. Instead of the classic von Neumann architecture, graphics processing units (GPUs) are increasingly favored as they are better suited to sort and match, rather than perform numerically intensive operations in a calculation chain. As a result, IC providers such as Micron Technology and Nvidia are doing well, while Intel is facing fundamental challenges in the future, the column claims. What's my take on this? I don't have one, frankly, as it is well out of my area of expertise. There are plenty of desk-bound experts and analysts out there to advise Intel on the future, and Intel doesn't need my opinion. Still, "so far, so good," I thought as I read this piece: I was learning something, or at least getting something to ponder. But one statement in the column really, truly bothered me. The columnist noted that Intel and other processor and memory vendors had benefited from Moore’s “law” over the past decades (really not a law in the strict physics sense; it’s really an insightful and surprisingly accurate conjecture), but the end might be in sight for this remarkably prescient conjecture. Then the columnist let loose with one of those casual, offhand, totally erroneous statements that drive me crazy. He wrote, “…the prior generation of chip makers got lazy from making easy gains with Moore’s Law.”   Wait a minute: they "got lazy" and made “easy gains"? Didn't we have countless engineers, scientists, technicians, physicists, mathematicians, chemists, programmers, and more working on each new step of the road map, at an RD and factory (aka fab) cost of billions? I guess it's easy to say "lazy" and "easy" when you are analyzing breakthroughs while sitting at your keyboard. The reality is that getting to each next step along the road map requires significant work, countless large and small breakthroughs, dedication, insights, trial (and errors), and frustration to make it happen. If this columnist thinks it has been easy, I'd like to know what he thinks is hard. In many ways, though, I don’t blame him for this attitude. Instead, I blame “us”—meaning the IC and electronics industry. No other technology segment lays out its goals, objectives, and timelines so clearly and specifically in advance via the "road map." It's not done by genetic scientists, cancer researchers, or pharmaceutical vendors. Instead, they all speak in broad terms about their goals and timelines, and with many caveats and "ifs."  By only doing this, I think they are actually a lot smarter than the IC industry is.   While it is necessary for our industry to internally plan and work with partners to try to make it happen, that’s as far as it should have gone.  What the public pronouncements and hyping of the IC-process road map did was make all this progress — which takes so much work and money — look oh-so-very easy and straightforward. All that has to be done to succeed in getting to the next step on that road map was for our industry to spend the money, do our homework, step the many pieces up to the next notch, and poof it would happen, just like that, with no surprises, no major issues to resolve, no real breakthroughs to be made. The many needed elements, including modeling tools, materials, fabs, and packaging, silicon or other substrate details would just happen with a little bit of hard work. It's my view that we in the electronics and IC industry have largely brought this result on ourselves, with the loud advance notice of what we are planning to do and then doing it. It’s the old story: when you create miracles on a regular basis – and the last few decades have produced countless amazing components and products – it is actually counterproductive to getting or maintaining respect from the public. Do amazing things over and over again, and pretty soon not only do they expect them to happen, they no longer appreciate them when they do. What's your view on how engineers and scientists, and our industry, ended up being "dissed" with statements that say they “got lazy from making easy gains”?