This article is full of it, but for more reasons than have been articulated so far.

One problem is that the article mistakes light for heat. Just because they get less press than Facebook doesn't mean that old-school high-tech engineering startups don't exist. Facebook et al are focused like a laser on getting their names out. That is what social networking is: word-of-mouth viral publicity. Whereas most hardware startups are born below the radar, and the majority die below the radar, and many of the successful ones get bought out by bigger companies and never rise above the radar under their own name. But that doesn't mean they don't exist. Some statistics would have been nice.

The balance of companies has shifted a lot over time, but that's because computing has grown and matured. Back in the 1970s Silicon Valley's only customers were nerds. That meant a lot of emphasis on nerdly products. H-P calculators. Oscilloscopes. Chipsets. After thirty years of work on mobile phones, broadband networking, and UX design, we've improved computing to the point where there are now many Valley firms whose customers are non-nerdly iPhone owners and Web users. That's a huge number of people. And it turns out that you can make as much gross revenue by selling a $2.99 app directly to 2 billion people as you can by selling a $2.99 part to a firm that will make 2 billion mobile phones, but the former requires much more marketing and makes a lot more noise.

But the big problem with this article is that it is trying to blame a forest-management problem on the individual leaves. If the USA is not investing enough resources in "alternative energy, better batteries, and nanotechnology", don't blame Silicon Valley startup culture. That's putting the cart before the horse. Entrepreneurs go where the money is, and who controls how much money goes into cutting-edge hardware research? The government, mostly. Either directly (by giving out research grants) or indirectly (by providing an artificial "market" with a single customer -- often the Department of Defense, but sometimes an entity like NASA -- that happily commits to spending money on the very latest high-tech toys; or by creating and enforcing the incentives that drive the market for, e.g., alternative energy).

The Internet got developed because the government commissioned it. Rocket technology got developed because the government commissioned and bought it. H-P and Intel became big companies by selling equipment and parts to the high-tech companies that sold to the government. The nanotech researchers I know are paid by the Department of Defense, the National Institutes of Health, or the National Science Foundation. Biotech research? Funded by the NIH or the DoD. Alternative energy? The size of our research efforts have tended to wax and wane with the size of the DoE budget; the economics of deploying alternative energy have everything to do with the cost of oil and gas (determined by forces outside of an entrepreneur's control, like taxes, and tax-funded highway projects, and tax-funded transportation projects, and zoning policy, and tax-funded military maneuvers, and government-mediated international agreements) and the extent of environmental regulation (determined by government and politics).

If you want to address these problems you must first analyze them correctly. If you're a rag like Newsweek it's easy and fun to pick on Mark Zuckerberg, but the real problem is economic and research policy: We've got an economy with flat or declining demand and lots of excess capacity (I, for one, could certainly be talked into applying my Ph.D. in EE to alternative energy problems instead of to PHP applications... if the price was right) yet our government officials are perversely engaged in an effort to further depress demand and raise our excess capacity by cutting spending. I wonder if there are any magazines that are ostensibly devoted to discussing problems in society, government and policy, instead of trading in mindless gossip, cheap shots, and PR hit pieces?

If you're not aware, most of the old guard companies pulled 15-20 hour days starting out too. And making a physical product is generally MUCH more frustrating than anything in web app land. Hurrying up and waiting, for a flawed product messed up by the machinists. Also much higher startup capital costs for anything nontrivial like semiconductors. Expensive equipment, office space, manufacturing space, all this stuff you need to start. Not just two or three guys with open source software working in a dorm.

Studying the physical sciences is also more mentally taxing than coding PHP. I've gotten paid for both. You can't compare designing aircraft in CAD and Excel (what I do) that could kill someone if done improperly, with setting up email verification for your social networking site. Not even close.

Yeah, it would be easier innovate on hard problems, Mr. Myhrvold, if you you weren't trying to strangle startups and small tech firms every chance you get with your Intellectual Ventures.