Why AI Changes the Math on Working Two Engineering Jobs
Most of what’s been written about overemployment assumes a 2021-era remote job: async-first, light on meetings, low scrutiny. That world is mostly gone. Return-to-office pushes, tighter monitoring, and leaner teams after multiple layoff cycles have made the original playbook a lot harder to run.
What’s changed in the meantime is that the tools available to a working engineer are dramatically better — and that’s arguably a bigger lever than any operational-security trick.
Where AI actually helps
Compressing routine work. Boilerplate, test scaffolding, first-pass code review, documentation — the parts of a ticket that used to eat hours can now eat minutes with a competent coding assistant in the loop. That time doesn’t disappear; it becomes headroom for a second job.
Covering meetings you can’t attend twice. Async standups and meeting summarizers mean you can be “present” in spirit for a meeting that overlaps with your other job’s calendar — reviewing a summary and posting a considered async update is a real substitute for live attendance in a lot of engineering cultures, though not all.
Triage and prioritization. Two jobs means two backlogs. Using AI to help you triage which tickets are lowest-risk to move quickly on, and which need your full attention, is how you protect your reputation at both jobs simultaneously instead of doing mediocre work at either.
Where it doesn’t help — and can hurt you
None of this is a substitute for genuinely understanding the code you ship. An engineer who’s clearly using AI to paper over a lack of understanding gets found out fast in code review, regardless of how many jobs they’re holding. If anything, running two jobs raises the bar on knowing when to lean on AI assistance and when a task needs your undivided, undelegated attention.
There’s also a detection angle worth naming honestly: some of the same monitoring tooling that watches for anomalous activity patterns can also flag unusual AI-tool usage patterns. Using AI well here means using it the way any single-job engineer increasingly does — visibly, as a productivity tool — not as a way to seem present when you aren’t.
The actual shift
The old OE advice was almost entirely about operational security — hiding calendars, managing two Slack presences, avoiding detection. That’s still relevant, but it’s not the whole game anymore. The more durable skill in 2026 is using AI tooling to make two jobs’ worth of real output fit into something close to one job’s worth of hours — which is a different, more sustainable strategy than just not getting caught.
We’ll go deeper on the specific tools and workflows in future posts.