# Here’s how I use LLMs to help me write code

<https://simonwillison.net/2025/Mar/11/using-llms-for-code/>

- Host: [simonwillison.net](https://links.thingelstad.xyz/hosts/simonwillison.net/)
- Saved: March 13, 2025 (2025-03-14T01:29:11Z)
- Record: https://links.thingelstad.xyz/links/10d08223e87d/
- JSON: https://links.thingelstad.xyz/links/10d08223e87d.json

## What Jamie wrote

Willison dove deep into LLM usage early on and has focused squarely on it for a couple of years now. This post is a great recap of what he has learned about using an LLM to help him write code. 

> Using LLMs to write code is **difficult** and **unintuitive**. It takes significant effort to figure out the sharp and soft edges of using them in this way, and there's precious little guidance to help people figure out how best to apply them.
> 
> If someone tells you that coding with LLMs is _easy_ they are (probably unintentionally) misleading you. They may well have stumbled on to patterns that work, but those patterns do not come naturally to everyone.

Willison’s comment near the end is great.

> I'm certain it would have taken me significantly longer without LLM assistance--to the point that I probably wouldn't have bothered to build it at all.
> 
> _This_ is why I care so much about the productivity boost I get from LLMs so much: it's not about getting work done faster, it's about being able to ship projects that I wouldn't have been able to justify spending time on at all.

We still have a lot of folks saying that computers are just going to code themselves with LLMs. Informed opinions are more aligned with Willison's. This is yet another huge boost in what developers can create. More layers of abstraction and power to create more things.

## Facts

- Kind: [Page](https://links.thingelstad.xyz/kinds/page/)

## Appears in

- [WT313: Weekly Thing 313 / Agentic, Shadows, WireMock](https://weekly.thingelstad.com/archive/313/), Notable

## Same host

- [Run LLMs on macOS using llm-mlx and Apple’s MLX framework](https://links.thingelstad.xyz/links/718b0d97cb99/) (simonwillison.net, February 18, 2025)
- [Hallucinations in code are the least dangerous form of LLM mistakes](https://links.thingelstad.xyz/links/1e3324e81167/) (simonwillison.net, March 3, 2025)
- [CaMeL offers a promising new direction for mitigating prompt injection attacks](https://links.thingelstad.xyz/links/f44b8768a2c3/) (simonwillison.net, April 17, 2025)
- [Backfill your blog](https://links.thingelstad.xyz/links/d0074f537816/) (simonwillison.net, May 1, 2025)
- [Expanding on what we missed with sycophancy](https://links.thingelstad.xyz/links/072f04967148/) (simonwillison.net, May 5, 2025)

## Saved the same week

- [The Fundamental Humanity of Reddit - The Atlantic](https://links.thingelstad.xyz/links/04b198406e5c/) (theatlantic.com, March 13, 2025)
- [Kill your Feeds - Stop letting algorithms dictate how you think - usher.dev](https://links.thingelstad.xyz/links/7f968f645384/) (usher.dev, March 13, 2025)
- [updown.io – Website monitoring, simple and inexpensive](https://links.thingelstad.xyz/links/406734e7dab5/) (updown.io, March 13, 2025)
- [Apple AI’s Platform Pivot Potential – Stratechery by Ben Thompson](https://links.thingelstad.xyz/links/178bbd8e263e/) (stratechery.com, March 13, 2025)
- [How Much Are LLMs Actually Boosting Real-World Programmer Productivity? — LessWrong](https://links.thingelstad.xyz/links/ca2bdddc0546/) (lesswrong.com, March 13, 2025)

## Elsewhere

- Wayback Machine: https://web.archive.org/web/*/https://simonwillison.net/2025/Mar/11/using-llms-for-code/