Hacker News Reader

Comment by cxr | Hacker News Reader

Suppose Google dropped all pretense and offered full-time work to any qualified applicant willing to work for minimum living wage—a 21st century workfare program. What would those numbers look like? What are the costs of the overheads? Suppose Google is able to draw managers ("mentors") from the upper end of the same applicant pool. What's the upper bound on programmers after arriving at an equilibrium by iteration? Knuth estimates 2% of the world has what it takes to be a programmer. Is that a reasonable upper bound for such a program, considering that the existence of the program will by incentive probably double or quadruple the number of working programmers in the world, but also keeping in mind that the most of the world's existing and aspiring programmers won't be interested in minimum wage long-term?

View on HN

Knuth's 2% estimate is the number of people who "really “resonates” with programming the way I do" [1], it is not an estimate for the number of people who have "what it takes" to program for a living.

[1] https://www.red-gate.com/simple-talk/opinion/geek-of-the-wee...

> Knuth estimates 2% of the world has what it takes to be a programmer.

Without knowing Knuth's criteria it sounds arbitrary. Would it be Knuth caliber programer? Then 2 percent seems high. Or one could change webpage color / deploy some simple app with click of button or running couple line of script. I guess 80% of adult population might be able to do in 2030.

> but also keeping in mind that the most of the world's existing and aspiring programmers won't be interested in minimum wage long-term

Well minimum wage workers are also not interested in that for long term. However that's what people will do if it is the only thing available.

Just for context, the US adult literacy rate is 80%

In the UK, it's 99%, and I'd bet there's not a huge difference in talent

Are you sure? This source claims it is 99%

https://www.nationsencyclopedia.com/WorldStats/CIA-World-Fac...

My guess it required quite a bit of technical finesses not obvious to a person with some technical experience.

When Knuth made that prediction, there were probably tons of rough edges to programming. But now we have powerful abstractions that make it much simpler. I think his estimate makes more sense if you qualify it as someone who is able to investigate and devise a bugfix for anywhere in that great sea of abstractions.

I was interested in where the 2% number came from and found the full context:

> Educators of computer science have repeatedly observed that only about 2 out of every 100 students enrolling in introductory programming classes really resonate with the subject and seem to be natural-born computer scientists...I conclude that roughly 2% of all people ‘think algorithmically,’ in the sense that they can reason rapidly about algorithmic processes

Programmers does not equal computer scientists.

That sounds already pre-selected to people interested in programming enough to enroll in a class in some [higher] learning institution.

So out of general population, that percentage would be even lower.

I'm not really sure what you mean. Google already offers full-time software development work to any qualified applicant for more than a minimum living wage. If they wanted to employ significantly more developers I'm not sure how lowering the compensation offer would help them.

Its likely that the OP means that the 'qualified applicant' criteria would significantly drop in order to increase acceptance rate.

What do you understand the purpose of the GSOC program to be?

a cynical person might suspect it's a cheap way to increase the supply of labor (and thereby decrease its cost) while generating good PR at the same time.

Reply on news.ycombinator.com