Hacker News Reader

Comment by fataliss | Hacker News Reader

Somewhat tangential topic, but ... It's interesting to me how most of the companies with real transformative potential have very little need for (dedicated) software engineers.

I think half of that is that most other engineering professions now come with some non negligible coding skill and the other half is simply that "software can solve anything" is a plain and simple lie.

I can't help but feel a little left out of innovation with "just" software skills. Am I too sorry for myself or is that a shared feeling?

View on HN

Btw, check this out: https://usfusionenergy.org/opportunities

Software engineering is an amplifier in almost every domain.

There's a nuclear technology company that open-sourced a tool for measuring fuel efficiency. https://github.com/terrapower/armi

But it's by no means the most important tool. There's a lot of hard, physical engineering problems that can be solved.

I think of software engineering like motor oil. You can apply it across a wide variety of contexts, but it's by no means the engine.

Software doesn't solve anything by itself -- You have to apply it to something. I guess some developers are just pure code monkeys that implement routines as specified, but there are plenty of software engineers who use their knowledge of the field their working in to create software to help solve the problems in a good way. I suppose the main difference between the two groups is the level of experience and interest in the field.

I think it depends on your definition of "transformative." Have FAANG not transformed the way we live our lives? If you restrict "transformative" to mean physical technology instead of information technology, then yeah, it makes sense that software is somewhat peripheral in that effort. There are areas where software is helping (ex: AlphaFold). Outside of AI, though, having domain expertise is necessary in addition to having programming skills, so it's more effective to teach domain experts to program than vice versa.

They are doing basic research, why would they need software engineers? They will need SWeng when they scale up no doubt, but they are doing basic research right now. Most of the coding will be done by the scientists there. Most nuclear scientists have a pretty good grasp on numerical programming, even if it usually isn't the most robust code in the world.

Reply on news.ycombinator.com