---
title: "Discussion on \"Three rewrites of model selection\" | Hashnode"
description: "Discussion on \"Three rewrites of model selection\". There is a question at the heart of Neo's architecture that I have answered four different ways in seven weeks. The question is: given a piece of code that needs to call an LLM, how does that code kno"
url: "https://hashnode.com/posts/three-rewrites-of-model-selection/6a319a94756e23815388e7d8"
---

[Feed](https://hashnode.com/feed) /

Discussion

# [Three rewrites of model selection](https://nullstack.hashnode.dev/three-rewrites-of-model-selection?utm_source=hashnode&utm_medium=feed)

There is a question at the heart of Neo's architecture that I have answered four different ways in seven weeks. The question is: given a piece of code that needs to call an LLM, how does that code kno

Apr 21 · 12 min read · nullstack.hashnode.dev  
Read the article

-   [#architecture](https://hashnode.com/tag/architecture)
-   [#llm](https://hashnode.com/tag/llm)
-   [#refactoring](https://hashnode.com/tag/refactoring)
-   [#abstractions](https://hashnode.com/tag/abstractions)

## Written by

[Jason Price](https://hashnode.com/@nullstack)

@nullstack · Portland, US

  
[View profile](https://hashnode.com/@nullstack)

Trying something different here

Build from zero. No assumptions. Notes on engineering, AI, and turning ideas into systems.

## Responses

Markdown supported · ⌘↵ to post

No responses yet.