S04E02 - Programming with AI - with Simon Willison
Mar 31, 2025
auto_awesome
In this conversation, Simon Willison, creator of Datasette and co-creator of Django, discusses his passion for building open-source tools for data journalism. He shares insights on blending AI with coding to empower journalists and the benefits of local large language models. Simon also dives into the transformation AI brings to programming and education, stressing the importance of foundational knowledge. Plus, he shares playful interactions with AI, revealing its creative potential and limitations. Expect intriguing thoughts on technology and journalism!
The podcast emphasizes the importance of developing tools that empower data journalists, particularly in supporting impactful storytelling amidst declining funding in traditional news.
Simon Willison discusses the transformative influence of AI models like GPT-3 on data journalism, enabling non-programmers to engage dynamically with technology and data analysis.
The speaker highlights the need for a playful approach to exploring AI capabilities, encouraging experimentation to foster creativity and confidence in using these tools effectively.
Deep dives
The Mission to Support Data Journalism
The goal of building software specifically for data journalists is emphasized. The speaker aims to create tools that assist journalists in telling impactful stories through data, particularly in the context of investigative reporting. This aim reflects a strong desire to democratize access to powerful data analysis tools, especially for small or emerging news organizations. By developing resources that can help journalists without programming skills, the speaker hopes to enhance the quality and accessibility of data journalism in a landscape where traditional news funding is diminishing.
Navigating the AI Landscape
The introduction of AI models, particularly GPT-3, significantly influenced the speaker's perspective on coding and data journalism. This model's functionality, which allows for sentence completion and code generation, changed the way journalists could engage with technology. The speaker discusses the challenges of lacking programming knowledge within journalism, suggesting that while teaching coding is one path, leveraging AI tools to assist those without programming skills may be a more effective approach. This shift in perspective highlights the potential for AI to transform how journalists interact with data and technology.
Challenges in AI Utilization
The speaker acknowledges the criticisms and limitations of AI models, particularly around their reliability and accuracy. Many users have had negative experiences with early AI models, leading to skepticism about their utility. Understanding the strengths and weaknesses of these tools is essential for harnessing their full potential in practical applications. The ability of AI models to generate useful information while also exhibiting flaws underscores the importance of user experience and familiarity with the technology.
Potential Applications in Local Newsrooms
The discussion highlights structured data extraction as a major application area for AI in journalism. By utilizing AI to analyze large volumes of unstructured data—such as police reports—smaller news organizations can access insights that would otherwise be prohibitively complex to extract. This capability not only supports investigative work but also enhances the efficiency with which small teams can operate. The speaker envisions a future where local newsrooms equipped with AI can produce high-quality reporting that rivals traditional media outlets.
Embracing Playfulness with AI Tools
The speaker advocates for a playful approach to exploring AI capabilities, as engaging with these tools in a lighthearted manner can lead to deeper understanding and creativity. By experimenting with various prompts and tasks, users can glean valuable insights about the limitations and strengths of AI, making the learning process enjoyable. This playful exploration fosters confidence in how to effectively utilize AI in both writing and coding contexts. The encouragement to 'try it out' reflects a broader strategy for integrating AI into professional practices without being overwhelmed by hype.
Simon is the creator of Datasette, an open source tool for exploring and publishing data. He currently works full-time building open source tools for data journalism, built around Datasette and SQLite.
Prior to becoming an independent open source developer, Simon was an engineering director at Eventbrite. Simon joined Eventbrite through their acquisition of Lanyrd, a Y Combinator funded company he co-founded in 2010.
He is a co-creator of the Django Web Framework, and has been blogging about web development and programming since 2002 at simonwillison.net
We talked to Simon about his goal of building tools for data journalists, what he's learned about tinkering with, and writing about, AI models for years, his excitement about their code-generating capabilities, how to get the most out of all of these tools, and what generative AI tools have to do with pelicans.
