Dual Giupponi retweetledi
The best career advice I can give to SWEs looking at a bleak job market right now:
1. Get a Claude Code Max subscription, today, and start using it. Be intentional about trying to use it to make yourself better - there is skill involved in using LLMs well. IDE integration is not a thing you should be reaching for if you want to be Claude maxxxing.
2. Concurrently with working your job application pipeline, start working on a side project that has a prayer of making $. Making $ from it will be a nice side benefit, but that's not the primary goal.
The primary goal is to close your skill and experience gaps without bureaucratic or political headwinds and learning how to use AI to help you do it.
I'll give an example: one of the developers in my org is brilliant at debugging all sorts of brutal systems programming problems, but struggles mightily with greenfield development. Handing him a blank sheet of paper and saying "build something that does X" is the worst way to use him - in his words, it's "analysis paralysis" that kills him.
AI has been a massive help to him there, because now he can use it get options, evaluate their trade-offs, write proposals, and do all of that privately where he feels most comfortable before he presents plans + options to us.
You need to honestly assess your own skill gaps and ask "where can AI help me grow?" - the goal should not be to have the AI do 100% of the thinking for you, but to help you gain experience + confidence by actually doing things.
Want to learn Rust? Have LLMs help you learn the idioms, package ecosystem, and toolchain. This also means you personally having to step in, be critical of the LLM's output, and do things manually as part of your own growth process. Having the LLM work with you, not for you.
3. Ship your side project - this is not optional. If you can't ship, free from constraints of "other people" AND with the help of cheap LLM labor, then you are at best a mediocre candidate and deserve to struggle in the job market.
Don't skimp on this - actually ship; try to get users; monitor for errors; listen to them bitch at you about how bad your work product is; improve it; and continuously deploy updates. Learning things the hard way is something LLMs will never be able to do - but you can.
If you've been hiding in BigCo bureaucracy all these years and have never taken full ownership end-to-end for a product launch, it's time to step up for the sake of your own career. The era of JIRA jockeys making $400k pushing tickets around is over.
4. Write publicly about your learning and experiences as you do them - if you haven't landed a job after putting in significant effort, it's for one or more of the following reasons:
a. You don't have any remarkable or valuable experience
b. You don't have a network of people who would recommend you
c. You are bad at marketing yourself and what makes you special / valuable / worthy of belief
This step solves all 3 of those, provided that you are actually doing the work for real. And if you own a distribution channel - like even just your LinkedIn or X account - all you need is a small audience of people who get it to start finding jobs through the backdoor.
All of this, all of this - hinges on you running at full speed to ship things, being willing to get criticized online for your mistakes, and being willing to be uncomfortable doing all of it for the sake of growth.
If you can do that, you will be on a great trajectory for 2026.
English
