davidgasquez.com
📚 Personal bits of knowledge
📝 Add Unix philosophy rules to AI models note
+11
Artificial Intelligence Models.md
+11
Artificial Intelligence Models.md
···
98
98
- The ability to ship to a preview environment.
99
99
- An instinct for what can be outsourced (to AI vs what needs human attention).
100
100
- An good (updated) sense of estimation.
101
+
- Some [[programming]] rules for code design from the Unix philosophy help here:
102
+
- Modularity: Write simple parts connected by clean interfaces.
103
+
- Clarity: Clarity is better than cleverness.
104
+
- Separation: Separate policy from mechanism. Separate interfaces from engines.
105
+
- Simplicity: Design for simplicity. Add complexity only where you must.
106
+
- Transparency: Design for visibility to make inspection and debugging easier.
107
+
- Fold knowledge into data so program logic can be stupid and robust.
108
+
- When a program has nothing surprising to say, it should say nothing.
109
+
- When you must fail, fail noisily and as soon as possible.
110
+
- Prototype before polishing. Get it working before you optimize it.
111
+
- Distrust all claims for "one true way".
101
112
- Build "partial autonomy" products where humans stay in the loop to verify output, rather than fully autonomous agents.
102
113
- [If you need to gather context, do that first in its own session](https://mariozechner.at/posts/2025-11-30-pi-coding-agent). Create an artifact that you can later use in a fresh session to give your agent all the context it needs without polluting its context window with tool outputs.
103
114
- In a world of abundant, near-unlimited coding capacity, human taste becomes most of the game.