Avoiding jobs related to standard aviation or oil is one thing, but software is so diverse and complex, which means you have to inspect the clients of a company to evaluate how green it is.
I would also be quite interested in anything related to non-obsolescent, long term, lightweight software practices, to fight the good old Law of Wirth, but it seems like the industry is barely scratching the surface of that problem, and it's such a large problem that I don't know if it's actually a problem of just removing software entirely where it's possible.
I don't even know if there exists some manual or standard practice to write "green" software.
Examples:
- Reduce complexity in frontend apps. Run less Javascript and animations in the frontend to save your users' phone batteries.
- Reduce unnecessary use of AI that requires CPU intensive use. Some use this just to tick off a box to attract more VCs
- Running huge Kubernetes clusters that are barely utilized.
- Do you need to run a full QA/test environment while you sleep, or could you fire it up when you actually need it?
- Choose options that may require more work, but reduce the CPU cycles (like replacing Electron with native code)
- Blockchain/Bitcoin/NFT have been used for software that doesn't really need it.
- Do we really need that many micro services? One drawback of micro services is that you need to fire up more instances and also additional infrastructure (load balancers, network traffic, separate databases and message queues for inter-service communication)
But the word to fixate on is information. Is the org causing info explosions? Is the org increasing info asymmetry? Is the org producing info pollution? What kind of information is the org pumping into your head? Etc