I am toying with the idea of using an LLM to write bits of code - not a whole program - then gluing the bits together into a larger whole using various technologies at my fingertips.
The first small experiment was to take a practical problem I am trying to solve, then write the code for part of the solution in English sentences.
I’ve devised a divide-and-conquer solution to my practical issue. I selected ⅒ of the whole project and generated code snippets in a Claude 4.0 ‘project’.
...
The general skeleton for a reactor is, roughly:
defreact save image copy script ❲
⟪
| *:
⎨
| "":
‛query in English’
‛query in English’
send ("", ...) # optional output of a mevent to a named port (“” in this case)
⎬
⟫
❳1
... pdf below ...
See Also
Email: ptcomputingsimplicity@gmail.com
References: https://guitarvydas.github.io/2024/01/06/References.html
Blog: guitarvydas.github.io
Videos: https://www.youtube.com/@programmingsimplicity2980
Discord: https://discord.gg/65YZUh6Jpq
Leanpub: [WIP] https://leanpub.com/u/paul-tarvydas
Twitter: @paul_tarvydas
Bluesky: @paultarvydas.bsky.social
Mastodon: @paultarvydas
Substack: paultarvydas.substack.com
The substack editor mangles code indentation, but the PDF version correctly retains indentation.