Hacker News Reader

Comment by sokoloff | Hacker News Reader

One problem in these discussions is the curve of value delivered by an individual developer in a particular situation is short-and-wide and while I agree the senior developers are on a curve shifted to the right, there’s a massive amount of overlap.

Most company’s best 5% 25 year-old developers are probably creating more value than their median 50 year-old developer, even though that median senior dev is creating more value than the median 25 year-old.

View on HN

> One problem in these discussions is the curve of value delivered by an individual developer in a particular situation is short-and-wide

I kind of disagree with this, although I don't think it's age that separates developers. But developers who can work independently are a lot more valuable than those that require close supervision, and those developers are considerably more value than those who are actively destroying value.

Yes, when you look at those differences any performance drop due to age is going to look tiny in comparison.

What I have found though is a lack of willingness to put up with bullshit. Stuff I accepted as normal in my twenties I'd just walk away from now. This is probably good for the employer but maybe not for your immediate manager.

“ But developers who can work independently are a lot more valuable than those that require close supervision, and those developers are considerably more value than those who are actively destroying value.”

True. But most people reach that point after a few years or never.

If I understand your comment correctly, I think we're agreeing (that there is a wide range [high variance] of value delivered across the population: think of a bell curve that's "smooshed" down to be short and wide).

Quite possibly. I realised after I posted that I wasn't quite sure what the axis of your plot were (I'm still not).

X axis is value delivered per engineer, increasing with X.

Y axis is a histogram of “number of engineers producing that value”, increasing with Y.

This, but for value instead of IQ: https://en.m.wikipedia.org/wiki/File:IQ_curve.svg

Yes, you hit the plateau of that curve pretty quickly. You can't just be an individual contributor and justify 3% raises every year for 30 years, when you hit that plateau on raw skill you need to start finding ways to player-coach the generation(s) of more junior developers behind you. I don't think that necessarily means you have to move into management, but it does mean broadening your definition of impact beyond LoC/commits/tickets.

1.5 to 3.0% of those 3% raises are inflation/cost-of-living adjustment, which seem like they should be able to continue forever.

> broadening your definition of impact beyond LoC/commits/tickets.

Is this the 'classic' expectation for a dev? Where I work the least technical person is the PM and everyone else comes together to determine what gets done and how it gets done.

Everyone knows the tech side and decisions gets filtered down to the rank and file. Taking initiatives are expected - I know this is likely the exception not the norm, but I hope future workplaces are more like this and less like that.

Agree about the value curve. Most developers do stuff that can be learned in a few years and then they just repeat the same thing with slight variation. There are really not that many jobs where you get challenged to grow throughout your career. It’s just repetition (like most other jobs)

Anyone in the trades is doing such a job. If you're a plumber, or electrician, or cabinet maker, or framing carpenter, or a mechanic, even if those 12 years (or 20 years or more) seem similar to people outside the field, you're learning all the time. It never stops, until you do.

It’s not like learning stops but the curve flattens out a lot. Another problem is that in tech 20 year old knowledge is worth way less than 20 year old knowledge in for example cabinet making.

> Another problem is that in tech 20 year old knowledge is worth way less

This is a widely held position, which sort of makes it a self-fulfilling prophecy. But I'd argue that it's not true. Most of what I know now that is really valuable isn't related to the technology of today or 1986. It's related to how to work, how to design systems, how to decompose problems, how to scale, etc. etc.

You might be surprised at how many companies are running 10+ year old technologies that were designed around the constraints of the 10+ year old systems that they replaced. If that's your thing, you can build a pretty strong career around being the person to keep that system humming along.

I refer to my resume as "a graveyard of obsolete technologies"

I could do the same. But really, it's also my value proposition. "What's that you say? Your org has a dozen crappy systems of questionable lineage and you need to migrate to SomethingFromThisDecade? Let me pick it apart for you and we can build a plan together." It's not much, but it's honest work. (and I happen to enjoy it...so win-win)

But this happens in tech management too. Last 20 years: CMMI? RUP? PMI? XP? Scrum? Design Thinking? Project to product management?

Also known as "1 year of experience, 12 times" (or "2 years, 6 times")

I frequently hear that, but in practically every other profession doing the exact same thing for 20 years is still valued. And the underlying reality is you can’t do the exact same thing for 12 years in the computer field. Experience is all the little edge cases that cropped up over time.

"Experience is all the little edge cases that cropped up over time."

Absolutely. I will be using this one.

This is not exclusive to dev work. Same can be said of managers.

I’ve worked with a lot of young developers and never met a 25 year old developer that’s creating more value than the average 50 year old developer with 20+ years experience. In theory sure it’s possible, but in practice it simply doesn’t happen even close to 5% of the time.

The thing is most people quit the field long before putting in 20+ years. So, I suspect it’s less about the experience than the filtering process.

So senior devs are better than 95% of 25 year olds.

Since finding a "best 5%" is almost impossible, because who knows until you hire them and they work there for a while, the best bet is to hire people who are older. Also, for a bone fide best 5%, they are going to work at the cool companies, and never will work at a sewer processing government utility company in their IT department, or work at for a local chain of tire stores with 50 locations and an IT staff of 3 people.

So, one might as well not even worry about the top 5%.

>One problem in these discussions is the curve of value delivered by an individual developer in a particular situation is short-and-wide and while I agree the senior developers are on a curve shifted to the right, there’s a massive amount of overlap.

This can also be said of managers. Then why aren´t they discriminated by age too?

Because they are smarter and work the system. Same for sales guys, lawyers and so on. They work the system to their advantage. Developers are one of the few groups that believes in a just meritocracy.

Reply on news.ycombinator.com