Newsletter
Innovator profile
SW

Simon Willison

Independent open-source developer and writer

LLM toolsPrompt injectionOpen sourceData journalism

Django and newspaper work

Simon Willison is a co-creator of the Django web framework[1]. Django's own documentation says that in fall 2003 the World Online developers, Adrian Holovaty and Simon Willison, moved from PHP to Python to build the websites of the Lawrence Journal-World, and began to extract a generic web development framework from that work[2]. Willison has written that he helped create what would later become Django at the Journal-World, and that he later spent two years on data journalism projects at the Guardian in London[3].

He has been blogging about web development and programming at simonwillison.net since 2002[1].

Lanyrd, Eventbrite and the JSK fellowship

In 2010 Willison co-founded Lanyrd, a Y Combinator funded company[1]. He joined Eventbrite when it acquired Lanyrd, and he was an engineering director at Eventbrite before he became an independent open-source developer[1].

In September 2019 he began a year-long fellowship with the John S. Knight Journalism Fellowships program at Stanford[3]. His stated aim for the fellowship was to grow an open-source ecosystem of tools that help data journalists collect, analyze and publish the data behind their stories[3]. He wrote that he wanted to make the kind of data reporting seen at well-funded newspapers more accessible to smaller publications without full-time software engineers[3].

Datasette and LLM

Willison is the creator of Datasette, an open-source tool for exploring and publishing data[1]. He has said Datasette was inspired by the challenges he saw at the Guardian[3]. The Datasette website describes it as a tool that takes data of any shape, helps people analyze and explore it, and publishes it as an interactive website and accompanying API[4]. The project is aimed at data journalists, museum curators, archivists, local governments, scientists and researchers[4]. It sits within a wider ecosystem of tools and plugins for working with structured data[4].

He also builds LLM, a command-line tool and Python library for working with large language models[5]. LLM can call models from OpenAI, Anthropic, Google and many other providers through their APIs, and can also run models installed on the user's own machine[5]. It stores prompts and responses in SQLite, generates and stores embeddings, extracts structured content from text and images, and can give models the ability to run tools[5].

Willison works full-time on open-source tools built around Datasette and SQLite[1]. He also does one day of work a week for Prime Radiant, Jesse Vincent's applied AI research lab[1]. He is a member of the board of directors of the Python Software Foundation[1].

Writing on prompt injection and AI agents

On 12 September 2022 Willison wrote about attacks, demonstrated by Riley Goodside, that use malicious inputs to make GPT-3 ignore its earlier instructions[6]. In that post he proposed the name prompt injection for this class of attack and compared it to SQL injection[6]. He also showed that the same technique could leak the original prompt[6].

In June 2025 he described what he called the lethal trifecta for AI agents: access to private data, exposure to untrusted content, and the ability to communicate externally[7]. He wrote that an agent that combines all three can be tricked by an attacker into sending private data to that attacker, because language models follow instructions found in the content they process[7]. He argued that users must avoid combining these tools themselves, writing that the LLM vendors are not going to save us[7].

His site also hosts a guide, Agentic Engineering Patterns, on getting good results from coding agents such as Claude Code and OpenAI Codex[8]. Its chapters cover topics including how coding agents work, using Git with coding agents, subagents, red/green test-driven development and agentic manual testing[8].

Known for

Sources

  1. Simon Willison: About me — checked 6 Nov 2026
  2. Django documentation: FAQ, General — checked 6 Nov 2026
  3. Simon Willison: My JSK Fellowship, September 2019 — checked 6 Nov 2026
  4. Datasette: An open source multi-tool for exploring and publishing data — checked 6 Nov 2026
  5. LLM documentation — checked 6 Nov 2026
  6. Simon Willison: Prompt injection attacks against GPT-3, September 2022 — checked 6 Nov 2026
  7. Simon Willison: The lethal trifecta for AI agents, June 2025 — checked 6 Nov 2026
  8. Simon Willison: Guides (Agentic Engineering Patterns) — checked 6 Nov 2026
Editorial profile compiled from public sources. Not an endorsement, and listing does not imply participation in this site. To correct or remove this profile, contact us.