Chris Kiehl: Programming Can Still Be Fun, Even If AI Could One-Shot Your Game
Chris Kiehl argues programming can still be joyful outside paid work, though reactions split on whether AI is shifting satisfaction toward product building or away from the coding people loved.
Chris Kiehl argues that enjoying programming remains legitimate even as AI changes the economics of paid, hand-written code. Work is not life, he writes, and the social pressure to treat coding as pointless fades when you disconnect from doomscrolling.
He still programs on weekends for fun, explores topics without needing a practical payoff, and says the reward comes from solving problems and understanding hard ideas. He also enjoys vibe-coding, treating it as a different kind of fun.
Key takeaways
- He is learning game programming by hand, currently a buggy Breakout clone with terrible collision detection; Claude could one-shot it, but that is not the point.
- One weekend he explored how far type safety in Java could be pushed using ideas from dependently typed languages like Agda; another went into recursion schemes, which was frustrating and left him unsure of the ideas' value.
- He argues that many beautiful programming ideas only become clear after years of baking in the mind, so exploration can be rewarding without an immediate application.
- He describes a social pressure to stop enjoying programming, which he attributes to consuming a doom stream; the feeling disappears when he disconnects.
- He also has a suite of vibe-coded apps for which he has never looked at a line of code, and suspects doing so would horrify him.
What people think
Analysis details
19 comments analyzed. Analyzed .
The sample selects active discussion branches and includes available parent comments. It may omit parts of the full thread. The analysis included 19 of 75 usable stored comments because of the input limit. Selected themes do not measure community opinion or how common a view is.
AI reshapes job satisfaction and career identity
AI coding tools are changing what developers do, with some reporting a shift from writing code to managing agents and verifying output. The change affects both paid work and hobby projects, as vibecoded tools can quickly replicate community contributions. The debate includes whether programming as a hobby can replace lost job satisfaction.
Divergent assessments of AI code quality
Claims that AI writes better code than humans are disputed. Some report Claude Sonnet produces code matching or exceeding their own, while others call the claim a lie. A proposed explanation is that LLMs appear better to less skilled programmers, so evaluations depend on the evaluator's baseline.
Using AI selectively for disliked tasks
One suggestion is to write code yourself and delegate disliked tasks to AI. Another describes switching to manual mode for interesting algorithms while using AI for product design and avoiding fragmented documentation.
Cognitive load and burnout from managing agents
Juggling multiple agents, verifying their work, and remediating problems spreads attention thin and reduces downtime, raising burnout concerns. The role shifts toward engineering or product management.
Business context determines whether AI code is better
Early-stage businesses may value speed and time-to-market over tight, performant code, making LLMs preferable. For well-factored code, LLMs may not match humans, though that may matter less than assumed.