Currently working on Datasette and associated projects. Read my blog or follow @simon@simonwillison.net on Mastodon.
-
Datasette
- Half Moon Bay, CA
-
03:05
(UTC -07:00) - https://simonwillison.net/
- @simonw
Sponsors
Highlights
Block or Report
Block or report simonw
Report abuse
Contact GitHub support about this user’s behavior. Learn more about reporting abuse.
Report abusePinned
-
-
-
shot-scraper Public
A command-line utility for taking automated screenshots of websites
4,897 contributions in the last year
Less
More
Activity overview
Contribution activity
July 2023
Created 264 commits in 18 repositories
Created 9 repositories
- simonw/llm-plugins
- simonw/llm-gpt4all Python
- simonw/codespaces-devcontainer Dockerfile
- simonw/blogs.hn JavaScript
- simonw/llm-mpt30b Python
- simonw/llm-markov Python
- simonw/datasette-execute-selected Python
- simonw/llm-palm Python
- simonw/asgi-proxy-lib Python
Opened 2 pull requests in 2 repositories
simonw/llm
1
open
surprisetalk/blogs.hn
1
merged
Reviewed 1 pull request in 1 repository
simonw/datasette
1 pull request
Created an issue in simonw/llm that received 11 comments
Introduce conversations concept for continue mode
Split out from: #65 Needs documentation
1 task
Opened 50 other issues in 16 repositories
simonw/llm
11
open
13
closed
- Review plugin tutorial one last time
- Ensure things don't break if you load a conversation from an uninstalled model
- New logging schema for conversations and continue mode
- Implement chains and use them for OpenAI functions
-
Support multiple
stopoptions for OpenAI models - Cleanup before shipping 0.5
- Update plugins.md documentation page
- Support --system for every model using prompt concatenation
- Maybe models should have a default template
- Make user_dir() a llm.user_dir documented top-level function
- Command showing available options for installed models
- Switch plugin tutorial to use pyproject.toml instead of setup.py
- Plausible analytics for documentation
- Add ruff for linting
- Add mypy for linting
-
Combine
LLM_*_PATHvariables into one -
llm.get_model(alias)helper - Tests fail with pydantic 2
- Figure out truncation strategy for continue conversation mode
-
Add
rufffor linting -
--keyshould be used in place of the environment variable - CLI tool for showing the models available from OpenAI models.json
-
llm -sas a shortcut forllm --system - LLM should default to logging to SQLite
simonw/llm-markov
2
open
3
closed
simonw/sqlite-utils
3
open
1
closed
simonw/llm-gpt4all
2
open
1
closed
natbat/pillarpointstewards
2
open
simonw/llm-palm
1
closed
1
open
nomic-ai/gpt4all
1
open
pradyunsg/sphinx-inline-tabs
1
closed
dogsheep/pocket-to-sqlite
1
closed
simonw/datasette
1
open
simonw/llm-mpt30b
1
closed
pydantic/pydantic
1
closed
simonw/datasette-execute-selected
1
open
simonw/symbex
1
open
simonw/datasette.io
1
closed
simonw/datasette-render-markdown
1
closed
2
contributions
in private repositories
Jul 9







