Yes, nothing is more depressing than listening to the a16z podcasts where they're discussing the minutiae of well-established b2b saas strategies for capturing long tail revenue. Shit, let's build some robots instead.
7 years ago I left silicon valley and moved to Raleigh, NC. It was part of our grand plan to move all the extended family back to the same area.
As part of this move I became involved with a couple manufacturing companies. I couldn't help but start helping with their IT set ups. I built some web apps to provide access to information locked up in their horrible ERP/MRP systems. 7 years later and these folks still call me a hero. Some simple web apps made a significant improvement to their productivity.
What I did was not rocket science. But it required spending time to understand the business, their workflow, and the bottlenecks. The solutions were simple after all that time was spent. The work ended up being very rewarding. There were no project managers, so I was full stack from the metal to the end user. It was also not bogged down by scaleability / uptime concerns. If something crashes, my user base is something like 30 people. I deploy with an ssh and git pull.
Anywho, I hope you get my point. I'd love to help Elon throw someone at Mars and he's my hero on many levels. But your simple IT skills could make you something of a small pond Elon for lots of non silicon valley people.
I started last year and inherited a horrific ERP system (and ancillary code for things like production scanners).
They realised they needed an ERP but they where unlucky in who they got to build it, things I take for granted as simple things to do (an interactive timeline showing job start stop, per user/department in d3) has had a massive impact on how they run things (it surfaced issues around job tear down/turnover).
Efficiency against estimate tracking (it's a table that's red/green if over under) in real time again made a huge difference since it surfaced issues with optimistic estimates from some operators and pessimistic estimates from others).
That matters because it throws off capacity planning and pricing on jobs (we go in too high, too low, too high we might lose the job, too low we take the hit on profits).
All vanilla stuff really.
None of this stuff is particularly difficult (refactoring parts of the existing shitpile are but that's because it's one gigantic pile of mud).
The fundamental problem is that there was a disconnect between the business and the specification of what they wanted, they didn't know enough to know what they could have asked for and the original devs didn't think about how they could improve on the existing business flow at all (I'm not even convinced they even considered the business in a lot of cases..).
In terms of users, about the same 50 on a typical day and I have a similar level of control, from the hardware we buy to the OS to the framework to the programming languages to well everything, it's an awesome feeling of freedom to be able to pick the correct tool for the job.
Technology applied properly can still have a multiplicative effect for a lot of SME's, there is still a vast amount of low-hanging fruit out there.
Exactly. You don’t need a 4-10 year degree in CS to solve them.
However, actual rocket science and other hard engineering disciplines is where educated talent is needed. Yet, the salary and startup opportunities are so much worse, that many mechanical and electrical engineers end up working in IT, consulting etc. You don’t build rockets, airplanes, cars or energy plants for the money
I'm an industrial engineer who deals with a fair bit of the IT side of manufacturing. I will echo the bit about it being very rewarding. It'd be cool to deploy something to millions of people, but there's something about making a physical thing work better that's just really cool. It's also way easier to explain to friends and family what you do.