It can take 5 minutes to explain a recommended solution to some technical problem.
But it can take anywhere from 5 hours to 5 weeks or even 5+ years (depending on the thing in question) to reach the point where you would know which solution to recommend and with confidence.
That differential is what should, in part, drive higher pay rates. In addition to the impact of the scarcity of raw talent, independent of education and experience levels.
The problem is when you get people who don't understand the issue, hiring people who do. And setting the price.
Also, the problem with these jobs boards is that you are competing with people in poorer countries, where wages are not as high in developing countries. There is no way you can do the same work at their price.
On top of all this, the hirer might be starting out at an incredibly low price, since it is not a huge priority, then increasing the price, although even for a low price this seems very low.
The job post referred to x.264, an open source video encoding/decoding library, while the commentary was directed at H.264, a proprietary video codec.
Yes, they are obviously related, but if you already know x.264, as the post asks, then it shouldn't be amazingly difficult to do. As in, the offer should have two more zeros on it, instead of four.
As someone who's on the business and technical side, I'm not sure this post shows much about the gap, except that some people are either cheapskates or ignorant. You can find folks who fit that description in just about any sufficiently large population.
The plural of anecdote is NOT statistic. One example of someone using a job bidding site to show their ignorance isn't indicative of anything. On the other hand, many non-programmers writing specs make the mistake of not asking/checking their assumptions. That mistake is not limited to any subset of humans, in my experience.
There are multiple gulfs of ignorance between geeks and business types, and they don't all go one way. Some of my coworkers and I spent ten minutes trying to decide how to answer a question in an RFI before it dawned on me (thanks to reading an article on HN) that the question was asking whether we would be willing to put source code in escrow. Until that realization, we thought they were asking how geographically distributed the backups of our CVS server were.
I think they're counting on someone looking in the code and the spec to satisfy their curiosity about what the job posting is about and then realizing that, 1) now that they've seen the bug they want to fix it, and 2) they've just spent all day working for free, so they might as well fix the bug and pick up $120.
I think he is pointing to an interesting phenomemon, which is the over-specialization of what we do:
Example:
CS degree --> compression/codec -> video or picture or sound (to me they are different worlds) -> H.264 -> specific standards of H.264 (no one knows everything)
The sweet revenge in this all is that the job poster will likely spend $120 ten times over churning through under-qualified programmers who think they can get the job done based on his description :)
It should not be horribly difficult to build minimally conversational AI's, arm them with names and e-mail addresses and let them address the problem by themselves.
It will be educational for the guy who wanted to pay $120 ;-)
But it can take anywhere from 5 hours to 5 weeks or even 5+ years (depending on the thing in question) to reach the point where you would know which solution to recommend and with confidence.
That differential is what should, in part, drive higher pay rates. In addition to the impact of the scarcity of raw talent, independent of education and experience levels.