We might all be AI engineers now

13 hours ago

The author enjoys coding, especially problem-solving and architecture decisions.
They now focus on building AI agents and tools, finding it more fun than traditional coding.
AI-generated code is often dismissed as slop, but when guided, it can outperform many developers.
The author is working on a complex project involving concurrent graph traversal, multi-layer hashing, AST parsing, and file system watchers, with AI handling much of the coding.
Debugging is enhanced by using multiple AI agents to quickly explore and solve problems.
The author no longer writes boilerplate code manually, saving significant time.
The real skill is knowing what to build and how it should work, not just writing code.
A strong foundation in software engineering is crucial to effectively use AI tools.
The barrier to entry for learning software engineering has never been lower, with abundant resources available.
Teams that understand and embrace AI are the ones the author wants to work with.

Hasty Briefsbeta