AI NewsProductivityReported
Shahar Glazner essay on engineers in 2026 hits 111 Hacker News points
An essay by Shahar Glazner reached 111 Hacker News points and 158 comments. He asks what engineers still do once agents write and debug the code.
Why it mattersThe piece is a snapshot of how senior engineers now describe their work. If this frame matches your team, it is worth reading before the next planning meeting.
Shahar Glazner published "On Being an Engineer in 2026" on his personal blog on 8 November. The piece reached 111 points and 158 comments on Hacker News. Glazner founded Keep, an open-source platform for watching alerts. His about page says Keep was bought by Elastic. The essay is first-person and runs a few hundred words.
The frame is that engineers have given away one piece of their job at a time. Every other week, Glazner says, another part moves to an agent. He traces the sequence step by step. First writing code. Then writing Slack messages, emails and documents. Then debugging. Then high-level thinking. His own day, he says, is now mostly about making sure his agents are not blocking other people's agents.
The argument for what people still do
Glazner then argues against the conclusion that engineers are done. His test is proof by contradiction. If models did all the work, output from engineers using the same model would look the same. It does not. Great engineers still produce much better results than average ones, Glazner writes. Agents also get lost when the problem is not defined well enough.
He lists four kinds of human leverage. Asking the clarifying question and defining the problem. A sense of whether a problem is hard. Thinking about whether a decision still makes sense in two years. And a feeling, hard to name, that a piece of writing was produced by an AI. Glazner grounds this last point in Michael Polanyi's line that "we can know more than we can tell."
How the room reacted
The 158 comments on Hacker News do not agree. One thread argues the author is wrong about the top of the distribution: a strong engineer with agents will produce more code and more bugs, not better code. Another thread accepts the frame and asks who is still in the room to teach the next cohort. A third picks up the "we know AI writing" claim and tests it with examples. The disagreement is honest, not drive-by.
A reader does not do one thing differently after this. The piece is a snapshot of how a senior engineer describes their week. If your own team has started saying the same sentences out loud, the essay is worth sending to them. It names the pattern in one place and gives the other side a careful hearing.
Source
- Shahar Glazner, On Being an Engineer in 2026 on shahar.blog.
- Hacker News discussion: On Being an Engineer in 2026, 111 points, 158 comments.
This item was written by an AI system from the linked source. Reveneau is responsible for what it publishes.
Get AI News in your inbox
New developer tools, model and agent releases, and how teams are actually using them to release software. Short, and only when there is something worth reading.
