simonw / simonw/llm

Add note to schemas tutorial about the date failure (and general unreliability of LLMs)

Open
#808 6 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

bug documentation schemas
Dominant language
Python
Stars
12.5k
Forks
998
Avg merge
3d 13h
Merged PRs (30d)
10

Description

In https://github.com/simonw/llm/blob/0.23/docs/schemas.md

https://bsky.app/profile/dracos.co.uk/post/3ljayeoea2224

The published date on that first AP article is 27th February, the last updated is 28th February, yet your output in that tutorial has 26th February. The date on the Katy Perry article is totally wrong.

On investigation, it turns out running AP content through strip-tags removes the dates entirely thanks to their web component:

Image

After strip-tags that ends up looking like this:

By 
JANIE HAR
Updated [hour]:[minute] [AMPM] [timezone], [monthFull] [day], [year]

This is a great opportunity to explain that you need to check this stuff thoroughly, but I don't know if I want to explain Web Components as part of the tutorial!

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start with docs/schemas.md and review the linked Bluesky report and the shown strip-tags output. Add a note explaining that the tutorial's extracted dates can be wrong or missing and that LLM output requires careful verification; done means the warning is clear without requiring a Web Components explanation.

Written by the indexing model from the issue text.

Assessment

Domain
documentation
Issue type
Documentation
Difficulty
2/5
Estimated time
1-3 hours
Activity status
Stale
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.