Stop writing code
The best code is code that doesn't need to be written.
If you spend more than 10% of your time as a software engineer writing code, you're doing it wrong. (I mean the time you spend doing your job, not the time you spend slacking your coworker that your PM unironically said "let's double click on that" in a meeting.)
The code you write isn't going to be around in 5 to 10 years. When your product is successful, old assumptions won't work anymore. Requirements evolve. You need to think about how your code will get replaced. Will it be easy or hard?
Don't leave behind a half-assed project, finished in one quarter and abandoned after you got promoted. If you can't commit to maintenance and support, you shouldn't write a single line of code. Use a popular open-source solution instead of writing your own. With zero effort on your part, there's documentation and a community that will keep working even when you stop. If there's no community solution, reflect on why that might be.
Stop writing code, and start thinking about not writing code. Only use 10% of your time writing code (or telling AI to write code) and the remaining 90% of your time figuring out what code not to write. Find off-the-shelf solutions. They age better. They come with better support. AI can auto-complete boto3 calls, it can't auto-complete your bespoke DynamoDB API wrapper written in Lua.
In the business world, they sometimes talk about core competencies. Very few companies have a core competency writing databases. Or build pipelines. Or CI/CD automation. Or container orchestration. You would never write those things by hand. Just like you probably shouldn't be writing whatever you're about to write by hand.
If you're the best company in the world to solve that problem, then do it. Otherwise, figure out how to avoid it.
Another Dockerfile? You sure?
A Helm chart? Are you serious right now?
Authentication??? The door is right there.
