Comment by perlpimp | Hacker News Reader

Comment by perlpimp | Hacker News Reader

Andresen has mentioned that government's funding of science is more crucial because it takes a while for a breakthrough into a new area to happen, like 20 years long - companies can grow on results achieved in science. However well intentioned VCs are they are more goal oriented.

Bell being huge and not unlike government organizations had bell labs - i am sure you all know that transistor came from there. Idea for bell labs was to put many great scientists together and wait for things to happen. Internet is the other thing - one cannot imagine private money investing that much in such nebulous goal - where everything from microcircuits to software stack had to be developed to birth the Internet.

VC doing science is not a bad thing but in the long run government can provide a steadier hand to people that do science than market driven entities. Financial market cycles are detrimental to projects that run on 20-30 year span or clusters of research projects.

If VCs can do a hybrid model where government can step in during lean times and help science projects survive with minimal bearacratic interference I think that could the best IMO.

I have mixed feelings about an impending Flight to Substance.

It's a good thing overall, but it can be bad for talented generalists. If there's enough demand for smart people that science and "real technology" companies will train talented generalists up into the specialties they need, then it's good. On the other hand, genuine research and technology have been so under-funded that the culture has become parochial and unreasonably elitist (against talented generalists). I deal with a ton of PhD bigotry in machine learning. (I don't have a PhD, but I read papers and I'm more current and, almost certainly, more effective than someone who stopped reading papers 5 years ago.)

For example, there used to be BA-level science jobs. There still are-- but they're filled by PhDs. Real science has been under-invested for years and we're paying a price because a lot of our brightest people don't have any experience with real work, because they can't get jobs on projects that actually matter.

In the 1970s to '90s, Silicon Valley wasn't pedigree-obsessed and talented generalists could "just fucking build it" without having to acquire permissions and credentials. What happened later, unfortunately, is that smooth-talkers moved in, outcompeted us on social polish, and are now raising huge sums of venture capital on lame "social media" ideas and build-to-flip bullshit. Right now, the genuinely talented generalist engineer loses in both theatres. He can't get a proper role in the Substance world (real science and technology) because he's not a PhD with at least one paper per year since graduation, but he can't get a proper role in the generalist's promised land of Silicon Valley because the smooth-talking douchebags with lame ideas have taken it over. Between "ultra-hard" science that requires either 5+ year job tenures (to make it worth it for employers to train people up) or advanced degrees, and that typically involves a decade between innovation and profit; and the "easy" social media stuff that any idiot can do (scaling such services is genuinely hard, but tech talent can be back-filled later) there is a "moderately hard" space that is under-explored.

The value of a Ph.D isn't in the knowledge you learn or papers you read, it's in demonstrating the ability to take on a vaguely-specified original research task where you can't see the eventual outcome at the beginning, and drive it forward until you come up with something useful. All without having the support of other people who actually know what they're doing to guide you. If you're measuring knowledge by what you've learned from papers, you're missing the point.

There are other ways to develop this skill. I've never once felt hampered by the lack of a Ph.D, but that's because fairly early on in my career, I took on tasks that were scary, that nobody knew how to do, and drove them forwards until they were useful to other people. But it does require following through on original work, and pushing your projects forwards even when other people think they're a bad idea, and accepting responsibility for results rather than effort or ability.

The value of a Ph.D isn't in the knowledge you learn or papers you read, it's in demonstrating the ability to take on a vaguely-specified original research task where you can't see the eventual outcome at the beginning, and drive it forward until you come up with something useful. All without having the support of other people who actually know what they're doing to guide you.

Fair, and I think it's pretty obvious that the education of a PhD (including the meta-education of having to teach yourself how to solve a new problem) is quite valuable, and that it's relatively rare that a person gains that outside of a degree program.

I don't like the hypocrisy and inconsistency. There are a lot of jobs that (even if the work doesn't require it) expect an advanced degree in the better jobs, but refuse to ante up and pay for employees to get them. If they really thought that the education was important, they'd hire people without the degree on condition that they pursue one, and pay for it.

they'd hire people without the degree on condition that they pursue one, and pay for it

It's my understanding that some companies will still fund your education, although it's not as common as it used to be. Here's a sampling: http://www.businessinsider.com/companies-that-will-pay-for-y...

Maybe that's not really what you mean, though...

I think the kind of deal that you're thinking about between an employer and an employee may have made sense at one point, when companies thought of employees as long-term assets, and employees had a reason to be loyal. The only way this might work now would be for an employee to agree to reimburse the education expenses if they leave before x years of employment are complete. That sounds like a loathsome form of indentured servitude, to me. I guess if you're desperate then you'll take it, and that's why some people sign up to the armed services in the US. But someone who is capable of PhD quality work probably has more flexible options.

"If they really thought that the education was important, they'd hire people without the degree on condition that they pursue one, and pay for it."

The EV of that would seem to depend a lot on the completion rate of PhD's.

Where were you working in Valley tech in the 1990s? Because I was there --- I moved there in ~1997, raised a round of funding in San Francisco in 1999, and started my career in 1994, and the Valley you describe where "talented generalists" could "just fucking build it" without acquiring "credentials" was not in evidence. The Valley was much, much more status-obsessed in the 1990s than it is now.

Did you not have the experience of needed a rabbi with grey hairs and a business degree to get a second meeting with any source of funding? Because that was my experience.

"Just fucking build it" might have worked if you happened to have a rack of Sun servers, some Oracle licenses, and a partial DS3 piped into your Cisco 7500. That's how things got just fucking built in the 1990s.

I understand that you have a problem with Valley tech culture as it exists too. That's unsurprising, because there's a lot not to like. But as a rhetorical strategy, attempting to rehabilitate the 1990s is weak.

Reply on news.ycombinator.com