What's a surprisingly useful tech skill that's underrated today?

1. Giving a damn. The majority of people don’t care at all about the work being performed. You can easily distinguish yourself by simply caring about whatever it is you’re working on.

2. A natural consequence of the above: fully reading the documentation, be it man pages or the official technical reference for a given technology. It’s incredible how many people don’t read documentation. You can become one of the foremost experts in your company simply by reading the documentation from front to back. You will seem like a genius.

Being able to use the features of a debugger to understand a problem in a single run, as opposed to editing the code adding/removing print statements and (recompiling+)rerunning the program over and over again.

There's been some kind of macho attitude dispersed wherein it's uncool to use good tooling, so a lot of people don't even learn how to use debuggers properly.

* Clarity by which you communicate in writing, such as emails. If you cannot do this you will never be put in charge in any kind of supervisory capacity.

* Task management, which is the ability to efficiently accomplish multiple small tasks without notes, reminders, or JIRA. It’s the notion of getting a bunch of shit done.

* Perceiving things in terms of facts. That means not guessing at measurements and not making assumptions. Most developers cannot do this. The inability to do this on any level is the most identifiable trait of Asperger’s.

* An ability to question the assumptions of others or even the modes of existing common practice. Most people generally cannot do this and it is this, not programming mastery, that makes for 10x developers.

* As for more technical things the ability to dive deeper and lower without leaving the current language or platform is what typically separates the masters from the commoners. These are the people who can solve problems others cannot, because the non masters are always more restricted by current conventions.

- writing thoroughly tested code that works and has minimal defects - being someone who is pleasant to work with and who shows interest in others - digging down into requirements and rooting out the real problem that needs a solution, actually solving that problem - individual 'tech skills' are just things you can look up on Google or read about in a book. I suspect the above are more important.

Server Management might be one. People seem to waste a lot of money for fancy cloud solution they don't actually need. And you said a little bash can solve do many problems

Have a deep understanding of your tools. And for a programmer, the top three are: the editor, the os shell, and your programming language tooling.

Knowing keyboard shortcuts of your chosen tools. Makes you so much faster and effective.

1. Understanding the fundamentals deeply, so that it's "muscle memory". Everyone is trying to race and learn as much as they can. I think they'd get much farther by focusing more on the fundamentals of the technology they use and the whys of that technology.

2. Caring. A lot of people just plain don't care. This might not be their fault. It may be a compensation issue, I don't know, but it's pretty rare, in my experience.

The question is what you mean by highly valuable. The sad truth is that in most workplaces many just coast along doing only what's necessary to not get fired. In which case doing your job more efficiently might draw negative attention (tall poppy syndrome).

Having said that, I think that having a good (both broad and deep) knowledge of the database schemas is invaluable for almost any information related role. Also, being able to manipulate large text very quickly (with vi/sed/awk, let's say), is useful for production emergency debugging on servers.

Simplicity/clarity, whether in code, conversation, or otherwise. Too many technical people tend to obfuscate their speech, their code, and so on.

Lazy Optimization

Before solving a problem, learn to apply lazy optimization. Don't just follow your intuition or the standard process. First, observe from the upstream or downstream of the problem to see if there's a better solution. Sometimes, just making a few simple changes upstream or downstream can solve the problem cleanly and perfectly.

I'd say carefully reading and understanding an error message before turning to google/ai

Honestly? Social skills to deal with customers. It’s not glamorous, but knowing how to talk to users, ask the right questions, and translate their vague complaints into actionable feedback is massively underrated in tech. I’ve seen engineers spend days debugging the wrong thing simply because no one thought to clarify what the user meant. On the flip side, a quick empathetic conversation can uncover the real problem in five minutes. It’s a skill that’s rarely taught, but it turns support calls, bug reports, and even roadmap debates into something way more productive. Hugely valuable—especially if you’re building or supporting a product.

Knowing when to demand proper tooling (such as IAR for embedded ARM)

Structural editing and editor-REPL integration

Using comm(1).

Vimium