Seth Walsh
Iconoclast
Contributor
- Joined
- Jan 12, 2020
- Posts
- 10,292
- Reputation
- 20,984
The CS grad market is brutal right now.
For years, the implicit deal was simple:
Do the degree.
Learn to code.
Get the junior dev job.
Build from there.
That ladder is breaking.
Not because software is dead. Software is eating more of the world than ever.
The problem is that entry-level software work is being compressed from both sides.
At the top, senior engineers with AI tools can now do more of the work that used to justify junior headcount.
At the bottom, ambitious interns, offshore teams and AI-assisted non-engineers can produce basic scripts, dashboards, automations and internal tools cheaply.
So the average CS grad is stuck in the middle.
Too inexperienced to own production systems.
Too expensive to justify as pure learning headcount.
Too replaceable if all they can do is CRUD apps, basic tickets, tutorials and generic LeetCode.
The old advice was: “Learn to code.”
The new advice is: “Learn to own a valuable system.”
That means infra, security, data engineering, trading systems, AI tooling, cloud cost, reliability, latency, product engineering, automation tied to revenue, or domain-specific software that actually matters to the business.
Generic junior coding is no longer enough.
The market is not saying software is worthless.
It is saying undifferentiated software labour is getting repriced.