The next tier of engineers won't manage people first - they'll manage fleets of agents. Learn to decompose work, set clear invariants, and verify output you didn't write. That's a leadership skill. Start practicing it now, on small tasks, before it's the whole job.
The fastest way to level up: rebuild something you already shipped, from scratch, with what you know now. You'll be embarrassed by the old version - that embarrassment is the growth. Do it once a year. It's the closest thing to a cheat code I've found.
Correctness is not a phase you get to later. The teams that "move fast and fix it in prod" are just borrowing time at a brutal interest rate. Build the boring guarantees in from day one. Speed without trust isn't velocity, it's debt with a countdown.
Your value isn't the language you know. It's the problems you can own end to end. Pick one system you only half understand - the database, the deploy pipeline, the payment flow - and go all the way down this month. Range compounds. Specialization alone plateaus.
Juniors ask "how do I build this?" Seniors ask "should this exist?" The most valuable skill isn't writing code faster — it's deleting work that never needed to happen. Audit your backlog this week. Half of it is probably solving problems you no longer have.
Every serious database can already tell you when someone else changed a row: SQL Server Change Tracking, Postgres logical replication, MySQL binlog, SQLite data_version. The capability was everywhere. The clean abstraction wasn't. So I built it: DbSignal (open source, on NuGet).
How I keep a large codebase honest: architecture enforced by lint/AST rules, not wiki pages. An import-graph test fails the build if a layer reaches where it shouldn't. Decimal for money, always. Treat the cause, not the symptom. Boring guarantees beat clever code.
I built DbSignal, a .NET library.
Your .NET app doesn't know when another program changes its database. So I built DbSignal, it auto-updates the app within seconds when the DB changes.
https://t.co/Z1dKNswuP7
I've rebuilt the same restaurant POS three times: WinForms, then .NET MAUI, now Rust + Svelte. Not because the last one failed — because every rewrite is a chance to earn the abstraction instead of inheriting it. Craft is deleting what you thought you needed.
@sickdotdev My occupation was really finding a field that really matches my vibes and my desired and finally AI launched and reach my life so happy for it