3 ms·
10X engineers deliver tech debt 10X faster than your average engineer. If there are some out there who truly are of unparalleled execution ability, focus on te
by arcticbull 7y ago
10X engineers deliver tech debt 10X faster than your average engineer.
If there are some out there who truly are of unparalleled execution ability, focus on testing, testability, reusability, modularization, componentization and simultaneously lifting up the team, I've never met one in my years in the industry. I don't think even if you spent 100 hours a week, you could achieve all of these things at once, so you have to compromise somewhere, and I know where people usually compromise. And that's before factoring in work-life balance haha.
In the early days of your company sometimes you do need people to just bang shit out, and eventually, you'll pay 10 1X engineers to clean it up. That period in a company's lifecycle is limited and the 10X'ers are outgrown very fast. Often they don't understand why but to their peers it's painfully obvious.
- Hermitian909 7y agoHaving worked with someone I'd consider a 10x engineer who did not create much technical debt, I think the defining characteristic of the group is the ability to choose good abstractions. This of course means that they are better able to shine in different contexts (he was not 10x faster at creating webforms in React, for example). A story I heard from him recently came from a consultation with a team at unicorn that had spent months modeling a particular problem using a set of graph algorithms. For the past 3 months they'd failed to get the performance they wanted despite a great deal of effort. They brought him in and explained the situation, he realized quickly that the results they needed could be achieved with a simulated annealing style job. Three days after walking in the door he successfully implemented a solution to a problem that had stumped the entire team, in their area of expertise, for 3 months. His career is littered with such anecdotes. Definitely a 10x engineer IMO.
- Nasrudith 7y agoThat isn't neccessarily true for technical debt - at least not in a straightforward way. But that applies to all technical debt really. If they say go through various iterations quickly and discard ones which don't work as well and leave quick notes as to how the failed approaches worked or didn't they could generate less debt than an equivalent team of ten. Working fast can lead to underdocumentation certainly.
- pvorb 7y agoThis is something that came up very often in the discussion on Twitter, too, and I think while there are these kinds of persons, there are also others that are more productive and are able to keep technical dept at a minimum. I think one can only determine the productivity of an individual by looking at their performance over multiple years on the same piece of software. Otherwise it’s hard to distinguish engineers that are truly productive from others that are only fast at coding.
- je42 7y agoyou describe net negative engineers not 10x engineers.
- arcticbull 7y agoI'm describing the engineers I've encountered who have been dubbed by others in the business as "10X".
- qdog 7y agoThat's pretty much my observation, I've seen many high-output types, but seen maybe one or two people who "might" be far above average output/quality. Generally I see people who become 10x by pushing down the effectiveness of those around them to 1x. If I write something that works, but am then told to pretty much re-do it all to fit a different design, my productivity becomes pretty low. Yes, I could start with a different design, but my impression has been that there is basically no initial design I can do that won't be re-designed by the '10x' individual, because they feel it's their job to 'improve' everyone else's design. Now, as another commentor pointed out, some people can be brought in to solve a problem others cannot solve (simulated annealing solution, which I'll just say "oooook" to). Sure, there are individuals who can solve actual problems others can't, but those actual problems are very rare. Most of the problems I have are 'ship the product' types, not 'solve this one very hard problem' types.