# TypeSafe says reliable decisions require code, not decision models

> Original title: so we heard you like decision models

- Source: TypeSafe AI
- Published: 2026-11-06T01:54:20.000Z
- Bad Influence: https://badinfluence.news/items/zyeri2rlae6wl8cszmpgy1x72
- Original: https://x.com/typesafeai/status/2118522339235389607

## Summary

TypeSafe AI argues that a model cannot make reliable decisions on messy data alone, so reliability requires code. It says reliable decisions come from composing reliable judgments in code, and that Jev makes those judgments. The post says calling Jev a decision model puts it at the wrong layer of abstraction and hurts reliability.

## Post

so we heard you like decision models

but how do you make reliable decisions against gnarly data with a model?

you don't: reliability requires code, and so reliable decisions are made by composing reliable judgments in code

Jev makes reliable judgments

## Quoted post @allietheicon

> Why don't we at TypeSafe call Jev a decision model?
> 
> If you think of Jev's questions as making *decisions*, I believe you'll use it at the wrong layer of abstraction and your reliability will suffer.
> 
> I wrote all about this on our site 🧵👇

https://x.com/allietheicon/status/2118508404192981040