I don't see how you can argue that each generation is equally ambitious, while also arguing that the previous generation was seeking out "safe jobs".

I agree that people are probably as ambitious as they were decades ago on an individual basis, but that doesn't mean that their goals have stayed the same over time. Even in the last decade, there's been a massive shift of technical talent toward entertainment technology, whereas it's getting harder and harder to find R&D jobs in physics, chemistry, biology and computer science.

There are always a few companies doing interesting technical work, but if the number of under-employed PhDs I know is any indication, the fraction of companies doing R&D is pretty darned small, and getting smaller every year. It's much easier for a PhD in Physics to get a job as a sysadmin at a web startup than it is for her to get a job in her field. I think that's sad.

I find it frightening that in the U.S. at least, basic research seems to have seriously declined at the government and large company level. Bell Labs is gone, and most companies now are focusing on applied rather than fundamental advances.

And before people start piling on large companies and the government with the usual rant, try to figure out how you could invent the laser, fiber optics, tcp/ip, lunar landing, orthgonal frequency division multiplexing , etc, with 2 college dropouts well-versed in rails and javascript.

I think there are cycles where technology get's to a state where basic research has huge value until you have enough basic research to to build a wave of new technology's. Once you have fully exploited the last wave of technology you are ready to build the next level of new technology. But, you rarely change technology's while there is still room to grow the old tech. Example: Modems starting at 300bps to 56k used the same basic technology, then DSL, then Fiber.

IMO, there is value to understanding the limitations to existing technology before you start doing basic research.

Yes. New technology is invented in response to problems with the older technology, and if you haven't capitalized on the old technology's potential, you'll likely just be reinventing the old solutions instead of coming up with new, better ones.

Interesting. A counter anecdote would be the advances in physics in the early 20th century, when we were still far from finished innovating with Newtonian mechanics. I've always doubted the whole "necessity is the mother of invention" thing; I don't think relativity was discovered because it became necessary to discover it. On the other hand, when it comes to technological innovations (semiconductors, transistors, lasers, etc.,) I just don't know enough scientific history to assess your hypothesis.

I think the limitations of Newtonian physics where only discovered as people built devices that broke them. There is a great time line for that period, but when you consider the instruments required to make the discovery's you see they are based on specific assumptions about how the world operates. It's only when their design reaches their limits that meaningful discovery's can be made.

http://timeline.aps.org/APS/Timeline/

"The Dutch physicist Heike Kamerlingh Onnes finds that mercury loses its electrical resistance at temperatures near absolute zero. This low temperature effect is observed in other materials as well."

That takes advanced equipment to reduce things to that temperature, an expectation that resistance is effected by temperature, and an formula that works at higher temperatures to notice the discontinuity.

PS: Consider all the benefits high speed computing has provided when designing aircraft. You need a lot of wind tunnel / real world tests to build a model, but with that model and lot's of computing power you can design aircraft out to the limits of your simulation. At which point you need to collect more data.

My point was that if you can make as good a case for x > y as for y > x, then x = y.

Do you have any evidence for this massive shift toward entertainment?

Surely there are more research jobs in biology than there were 30 years ago, considering biotech was mostly invented in the intervening period.

Are there more biotech startups than 30 years ago? Sure. But probably not more than 20 years ago, and certainly not more than a decade ago.

There's plenty of evidence of a shift toward entertainment. Look at the big startup successes of the last decade: Google, Myspace, Facebook, Friendster, YouTube, Blogger, Flickr, Digg, Bebo, etc. Save Google, nearly every major success story has been an entertainment/media play. In fact, I don't see this argument as particularly surprising or controversial, given that the web has been the dominant technology of the last two decades, and that the web has always been about media.

I think we both agree that personal ambition is pretty much constant over time. My argument is that if that's the case, then it's not reasonable to argue that people preferentially sought out "safe" jobs three decades ago. Smart, ambitious people have always sought out the opportunities that make them most successful; today those opportunities are clustered in industries that reward ephemera.

A quick side note. The number of biotech startups is mostly a function of the legal environment. The industry is more regulated that almost any other sector of the economy.

Your evidence is that most of the startups you've heard a lot about are mass market consumer startups? Would you expect to have heard as much about the ones working on infrastructure?

If I didn't know anything else about the industry, then probably not. But that's kind of the point -- that's the list of successes that everyone knows about; the list that every CS undergrad wants to found.

Their privacy policy mentions: "patent-pending edit synchronization algorithm". I'd much appreciate hearing about their experience with a software patent (I'm in a similar position).

Unless they want to not publicize it, to avoid anti-patent sentiment? e.g. two TechCrunch comments: http://www.techcrunch.com/2008/11/19/etherpad-shows-google-d...

>The Etherpads, for example, had to literally prove theorems to get their real-time collaboration sw to work, and they hope eventually to use it in a whole range of applications.

Could you tell them to publish the papers or the notes please?

The Etherpads finished college, though.