Yea, I’m surprised how rarely domain comes up in these discussions. And, as you say, it’s quite obvious that knowledge and raw intelligence can give you far more than a 10x advantage in some domains.

For an extreme example, imagine that the year is 1993 and the task is to write a program that produces a proof of Fermat’s Last Theorem... My bet is that a certain “developer” named Andrew Wiles would have wastly outperformed a ”team” consisting of every other person on planet earth.

I'm with you on the domain topic; reading the extreme example I'd add repeatability of the performance (on a different problem of the same domain) as an equally important factor to distinguish the developer from the inventor.

oooh oooh oooh, what an interesting example! I recall an interesting comment on NH a while ago where the dev was describing their refactoring of some law code. The original code was if-then-else, and had the same 'shape' as the horrid law it was interpreting. Our intrepid dev redesigned it to use some kind of truth table instead, and was terribly proud. And then the next week the law was amended and a new clause added deep down in one subclause or something. It turned out that the clever algorithmic approach wasn't actually the most maintainable... So, using a sales tax calculator as an example, its possible for two different devs to take to completely different approaches, and for the betterness of one or the other to be in the eye of the beholder until the code has been maintained in production for a long time, by which time doubtless whichever approach was adopted is deemed to be the worse one as the other has no production experience so might be better ;)

Honestly, adding legislature into code is a losing game. Let the developer just deliver a flowchart engine so that even unpaid interns can get to modify the logic.

In my opinion, a whole part of being a productive developer is about empowering people without the technical chops. If someone getting paid 10% of your salary can complete your previous responsibilities, you're technically a 10x developer.

I learned a few years ago that there are laws in the UK, parts of which are written in a naturalish language dialect of some complicated proprietary Oracle product. There’s a productivity gain for you (if you’re an Oracle salesperson at least).

For anyone wanting to see the original comment: https://news.ycombinator.com/item?id=18996147

I remember this comment. I believe he discussed using a bloom filter.

I was just thinking about how at most jobs I've ever worked the main barrier to productivity is the inability to make minor (in my opinion) changes to organizational policies that have high business logic complexity relative to benefit.

Yeah. Particularly if you don't have the authority to implement changes that would improve things in the long run, there's really no way for a dev to be a "10x" in that situation, even if they may have the ability to be a "10x" dev in other contexts.

A lot of it is drive and people skills.

If senior people like/trust you and you're willing to stand your ground/fight for something you can normally achieve significantly more.

That and luck on how those above you are trying to play the corporate politics / personal career game.

Agreed on all counts. Drive and people skills matter a lot. Also, a lot of nefarious people skills are also rewarded.

    If senior people like/trust you and you're willing 
    to stand your ground/fight for something you can 
    normally achieve significantly more.

Yes, although in a lot of company cultures this is nearly impossible due to the lack of level playing field.

A "software architect" I once worked under liked to play the following game, and he played it awfully well. We'll call him "B."

1. "B." spends days, weeks, or months researching a solution all by his lonesome.

2. "B." presents this idea to the team. It is effectively an edict, but he presents it as a meritocratic chance for others to offer feedback or alternative ideas.

3. The rest of the team is caught flat-footed, because we... just heard about this problem minutes ago, and obviously aren't as prepared as... well, somebody who has prepared. We are also not allocated time away from our scrums in order to research and test alternatives, whereas that was essentially his entire job. In other words, we bore the burden of proof and yet had no time in which to produce said proof.

4. I don't recall one of the edicts-masquerading-as-RFCs from "B." ever being successfully challenged.

Management (who was non-technical) perceived this as evidence of B's superiority, because after all, it's not like folks had better ideas than "B." Since he structured things in a way such that folks really couldn't have better ideas.

Eventually many folks on the team (like me) tuned out and just stopped participating in the whole charade after a few trips around the block. This, of course, was seen by management as more evidence that "B." was the only one with legit ideas.

> If you are writing a sales tax calculator and need to implement 500 different local state sales tax regulations from a scanned Word document, your best developer is going to be maybe twice as fast as your average one.

Assuming "regulations" is actually functionally different and not just some sort of lookup table for a multiplication constant, then having been in a situation with only 10 rules, even such a simple example can easily have a difference on the order of 10x just from prior experience with organizing large amounts of lookups into code (+edge cases, +consistent interface, +etc) instead of into tabular data.

There are constants, marginal rates and limits. Furthermore there are questions about how to determine jurisdictions.

It's very not easy and some companies make bank fulfilling client tax obligations or licensing based on their domain experience.

A scanned Word document -> sales tax calculator sounds more like a death march and my first determination as an engineer would be to confirm if management has done the due diligence as to the buy vs. build cost.

That's a funny example because for me it's the opposite. I'm 100x in the tax domain for my country since I know the tax law by heart and can implement things from the top of my head while my colleagues struggle.

> If you are writing a sales tax calculator and need to implement 500 different local state sales tax regulations from a scanned Word document, your best developer is going to be maybe twice as fast as your average one.

This isn’t software development. It’s mostly translating a scanned document into data. If you leave that task to a non-developer (because it isn’t development-related), I’m convinced you can see a 10x difference in productivity for the remaining work (writing the actual software).