Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

After Credentials

paulgraham.com

Post reply on HN
Print view View original

101–110 of 209 posts

Re: After Credentials

#101

Post by kingkongrevenge »

Credentialism and Mandaranism are artifacts of state dominated societies. It's really that simple. Free market economies feature so much creative destruction that power has little chance of persisting across generations. Only in a system lacking Darwinian market selection mechanisms do credentials become so critical. Credentialism emerged in China because the Chinese state was always much more powerful than the European crowns.

Credentialism has declined in the US to the extent that certain segments of the economy have adapted and grown so fast as to escape the clutches of government control. The segments still under firm influence of regulation are heavily Mandarin and Credentialist. Think medicine or law.

Re: After Credentials

#102

Post by MaysonL »

aswanson wrote: post #70
michael_nielsen wrote: post #40

Earlier quoted context omitted.

Nobel Prizes per undergraduate is an interesting measure. More than one in every 1000 Caltech undergrads wins a Nobel Prize, and that number rises substantially if one restricts just to science students, not engineering (which doesn't have a Nobel). This suggests that the admissions office there is doing something right.

Are you controlling for those not admitted that went to similar caliber schools, or have an example of how that statistic compares to those schools with a similar quality applicant pool?

I know that Swarthmore, my alma mater, has had 5 Nobel laureates (if you count one Economics prize), and is comparably selective, and only slightly larger than Caltech. Note that it only has 20-30% as many science grads as Caltech [rough estimate].

Re: After Credentials

#103

Post by alexandros »

netcan wrote: post #84
alexandros wrote: post #15

Paul seems to be hitting on the optimization by proxy issues seen in many areas, even in pagerank. Google looks not for quality of a page but for indicators of that quality, namely inbound links by quality websites. Similarly exams and credentials do the same. If you can write about these books, you know literature. If you got a degree there, you must be very capable. And in the first generation it works. However, as…

How would you run a cram school for YC? Say you were given 19 yr olds at the end of Year 1 of Uni. You have 2-3 years part time. How would you get them in? Activities? Interview coaching? Resume building? Portfolio building? Specialised learning?

Intresting question, but the feasibility of it is not dependent on my ability to give a good answer. Regardless, I'll give it a shot for argument's sake. Of course everything you say is valid. you would need to do these things.

I would also mine all the successes, details of interview process and characteristics of successful applicants and their applications for commonalities and then see how many of these can be easily added to an application for 'bonus points'. Of course it depends on how black-hat you want to get about this.

The first thing that came to my mind though, having read pg's essays, would be 'teach them LISP' :)

Re: After Credentials

#104

Post by derefr »

netcan wrote: post #84
alexandros wrote: post #15

Paul seems to be hitting on the optimization by proxy issues seen in many areas, even in pagerank. Google looks not for quality of a page but for indicators of that quality, namely inbound links by quality websites. Similarly exams and credentials do the same. If you can write about these books, you know literature. If you got a degree there, you must be very capable. And in the first generation it works. However, as…

How would you run a cram school for YC? Say you were given 19 yr olds at the end of Year 1 of Uni. You have 2-3 years part time. How would you get them in? Activities? Interview coaching? Resume building? Portfolio building? Specialised learning?

I'd probably build their social network: connect them to hundreds of influential sources of free marketing. If you're judging performance, the initial performance of a connected founder will greatly exceed an unconnected one. If you don't wait long enough to watch market effects take over before judging which startups to keep, you'll only end up with the networked ones.

Re: After Credentials

#105

Post by tlb »

cperciva wrote: post #45
pg wrote: post #10

Earlier quoted context omitted.

Some of what you say is true, but it's not true that the principal value of YC is the brand name. The YC brand has little weight to the customers of a startup, and without those it quickly dies. YC's principal value is the advice we (and the other YC startups) offer. The part you're right about is that "admission" to YC is structurally a lot like college or grad school admissions. The difference, though, is that our…

College admissions officers do not, as far as I know, follow up on the careers of the people they accept and see how they did. In Canada, one of the strongest selling points of universities is what fraction of their students got a job in a related field within X years of graduation -- so at least here, institutions do track this sort of data. In addition, there's a lot of work done to validate admissions processes ba…

The fraction of people getting a job in a related field isn't a good measure except for trade schools. Often, the most successful and interesting people do something unrelated to their undergraduate degree. Or they do something other than a "job".

Re: After Credentials

#106

Post by kwamenum86 »

"History suggests that, all other things being equal, a society prospers in proportion to its ability to prevent parents from influencing their children's success directly."

I would love to see some evidence for this claim.

Re: After Credentials

#107

Post by netcan »

alexandros wrote: post #103
netcan wrote: post #84

Earlier quoted context omitted.

How would you run a cram school for YC? Say you were given 19 yr olds at the end of Year 1 of Uni. You have 2-3 years part time. How would you get them in? Activities? Interview coaching? Resume building? Portfolio building? Specialised learning?

Intresting question, but the feasibility of it is not dependent on my ability to give a good answer. Regardless, I'll give it a shot for argument's sake. Of course everything you say is valid. you would need to do these things. I would also mine all the successes, details of interview process and characteristics of successful applicants and their applications for commonalities and then see how many of these can be ea…

My intention wasn't to corner you, just to start a discussion. For arguments sake, let's get about as black hat as a Chinese mother with a shot at Oxford.

I don't know much about YC or what influences them apart from fragments from pg's essays.

Insisting that LISP or some other language no one uses is absolutely the best tool seems like it might be a bonus point. But I think that is as a bonus. What would be the core? They say smartness & an attitude that they are going to do something big with or without the program. The latter can be gamed. The former, I don't know much about the criteria.

Re: After Credentials

#108

Post by Retric »

mhartl wrote: post #98
michael_nielsen wrote: post #40

Earlier quoted context omitted.

Nobel Prizes per undergraduate is an interesting measure. More than one in every 1000 Caltech undergrads wins a Nobel Prize, and that number rises substantially if one restricts just to science students, not engineering (which doesn't have a Nobel). This suggests that the admissions office there is doing something right.

I think it's actually Nobel Prizes per alumnus, i.e., it includes advanced degree-holders as well. (See http://www.admissions.caltech.edu/about/notable , which states that "Caltech has over 20,000 living alumni—of whom 17 are Nobel Prize-winning scientists", and goes on to list lots of notable Caltech Ph.D. alumni.) So the graduate admissions committees are apparently doing a good job, too (though clearly they screwe…

You can get a Nobel Prize science prizes for "physics, chemistry, physiology or medicine" but they often give them to 2 or 3 people so let's say 17 / (~8 per year over ~40 years) which gives you ~5%. That sounds great but I think most Nobel Ph.D. have 2 or 3 schools (if not more) in their background so it's really more like 2% which sounds about right for one of the top 25 science schools in the world.

Re: After Credentials

#109

Post by rams »

"All it takes is a few beachheads in your economy that pay for performance. Measurement spreads like heat. If one part of a society is better at measurement than others, it tends to push the others to do better. "

There is little evidence of that - Car companies, investment banks, etc, etc

Re: After Credentials

#110

Post by flashgordon »

All credentials are just another way to reduce the "blameable" risk on a party. People fork out hundreds to do a whole bunch of certifications which are happily accepted by large organisations as they are somehow proven... This in turn is just a manifestation of an aspect of the Principal Agent Problem - Adverse Selection!!

(by the way I hate doing exams and I absolutely suck at them because when I study I am interested in learning more than I am in getting a great mark at an exam...)

Post reply on HN