Engineering foundations
Engineering foundations matter more when agents write the code, because the work that remains is judging it: framing the problem, seeing what the agent missed, and verifying output that a model inferred rather than compiled.
Written by
-
Benedikt Stemmildt Founder & Co-CEO
Updated:
Getting Started 2
Our position. We see the skills gap as one of engineering fundamentals, not of AI skills: the fundamentals were always important, and AI made them urgent (opens in a new tab).
Agents absorb the well-scoped, low-risk tasks juniors once learned on, which threatens the pipeline that turns juniors into seniors. The trap behind it is the seventy-percent problem: the agent gets roughly seventy percent of the way, a junior cannot see the missing part, and a senior absorbs it in review (The AI-Native Software Engineer, Addy Osmani (opens in a new tab)).
Hiring has moved to the skills above the automated ones. Asked what they now prize, twelve CTOs named critical thinking, systems thinking and translating business context, not prompt engineering. Their interviews hand candidates slightly flawed AI-generated code and watch whether they pause, ask and audit it (CTOs Reveal How AI Changed Software Developer Hiring in 2025 (opens in a new tab)).
Verification becomes a foundation of its own. LLMs are inferrers, not compilers, so their output never becomes a settled layer you can stop reading. Every use weighs the impact of a failure, its probability and your odds of catching it before it ships (I still care about the code (opens in a new tab)). Hallucination clusters where real work happens: next to a well-known concept, asking for specific details such as an internal library or a just-released API version (pitfalls of agentic development podcast (opens in a new tab)).
Start here
- The junior learning path: How AI changes entry-level work, hiring signals, and the deliberate practice needed to build future senior engineers.
- Verification literacy: Why probabilistic code generation makes review, testing, and awareness of model failure modes foundational skills.
Related
- Getting Started and Learning: where foundations fit into the first weeks of practice.
- Software Engineering: the values and principles that agents make matter more.
- Architecture work in the agentic era: where the engineer’s job moves when agents write the code.
- Quality and Verification: why checking, not generating, sets the pace.
- Senior Engineering Skills: the skills around the code, such as writing the design docs and conventions agents now build from.
Resources
Articles
- CTOs Reveal How AI Changed Software Developer Hiring in 2025 (opens in a new tab) - Final Round AI interviews with twelve CTOs, who prize critical thinking, systems thinking and business-context translation over prompt engineering.
- I still care about the code (opens in a new tab) - Essay arguing that LLMs are inferrers, not compilers, so their output stays a standing risk that every use must weigh.
Posts
- The fundamentals were always important, and AI made them urgent (opens in a new tab) - LinkedIn post framing the skills gap of the agentic era as one of engineering fundamentals rather than of AI skills.
Talks and videos
- A Developer’s Guide to Not Getting Replaced by Agents (opens in a new tab) - AI Native Dev podcast episode on agentic pitfalls, locating hallucination next to well-known concepts where specific details are asked for.
- The AI-Native Software Engineer (opens in a new tab) - JavaScript Conferences by GitNation talk naming the seventy-percent problem: agents get most of the way and juniors miss the rest.
In this section
- The junior learning path
AI absorbs the small, low-risk tasks juniors once learned on, so teams have to create learning work on purpose. Hiring shifts with it: interviewers now test whether a candidate questions and audits flawed AI-generated code.
- Verification literacy
What an engineer needs to review generated code well, and why that code needs a different kind of review, for anyone who checks agent output.