← Originals
Simon Willison

Simon Willison

6 interviews

What they argue

Simon Willison is a thinker who focuses on the structural, economic, and technical evolution of the AI industry. He explores the shift from raw model novelty to system utility, the dynamics between closed and open-source development, and the practicalities of building on modern AI infrastructure.

  1. The Slow Takeoff Philosophy

    AGI should be deployed incrementally to manage safety risks and ensure controlled integration into the global economy. This explains why certain features, such as full autonomous loops, may currently be missing from AI systems.

    From AGI is Being Achieved Incrementally (OpenAI DevDay w/ Simon Willison, Alex Volkov, Jim Fan, Raza Habib, Shreya Rajpal, Rahul Ligma, et al) (episode)

    • Says much the same Greg Brockman — Shift to Safety and AlignmentThe transition into the AGI era marks a shift from a 'compute-first' race to a 'safety-and-alignment-first' era. The primary challenge is ensuring that increasingly…Greg Brockman on What it Means to Cross Into The Age of AGI

  2. The Three Pillars of Customization

    To build successful AI applications, developers must leverage three distinct layers: the System API for stateful interactions, Dialogue, and other layers of customization. This framework helps in building on the new Assistant and GPT infrastructure.

    From AGI is Being Achieved Incrementally (OpenAI DevDay w/ Simon Willison, Alex Volkov, Jim Fan, Raza Habib, Shreya Rajpal, Rahul Ligma, et al) (episode)

    • A related idea Aman Sanger — The Abstraction Stack ModelTo manage complex systems with AI, developers should use a mental model of navigating different levels of code. This includes viewing the system as a set of logical…#447 – Cursor Team: Future of Programming with AI
    • A related idea Greg Eisenberg — The Shift to Personal Agent PlatformsThe AI landscape is shifting from a 'chat' interface to a 'personal agent' platform, exemplified by OpenAI's Dots. This creates a massive distribution channel where…OpenAI DevDay: Dots, Agents & $100B Opportunities

  3. The Open Source Structural Advantage

    Open source is a structural advantage for innovation because the community can iterate on smaller, stackable models at a velocity that exceeds the bureaucratic constraints of giants like Google. The 'moat' of being a large, closed-source company is evaporating because of this speed.

    From No Moat: Closed AI gets its Open Source wakeup call — ft. Simon Willison (episode)

  4. The Commodity Analysis Framework

    To find a competitive advantage in the AI era, one must evaluate their position across three tiers: the Model Layer, the Data Layer, and the Application Layer. The Model Layer is increasingly viewed as a commodity.

    From No Moat: Closed AI gets its Open Source wakeup call — ft. Simon Willison (episode)

    • A related idea Dylan Patel — The Infrastructure MoatThe primary competitive advantage in the current AI era is the ability to operate and own highly efficient, custom-built infrastructure that competitors cannot match.…The State of Silicon and the GPU Poors - with Dylan Patel of SemiAnalysis
    • A related idea Alistair Pullen — Sovereign AI Competitive FrameworkTo compete with billion-dollar labs, entities should follow three pillars: infrastructure leverage, architectural differentiation, and specialized RL. This includes…Watching America Run Away With AI - Alistair Pullen (Cosine AI)

  5. The Collapse of the GPT-4 Barrier

    The belief that only massive, proprietary infrastructure could produce GPT-4 level intelligence was proven false in 2024. Numerous organizations have now released models that match or beat GPT-4's performance at a lower cost.

    From [Ride Home] Simon Willison: Things we learned about LLMs in 2024 (episode)

  6. Horizontal Expansion of Capabilities

    While the public expected a 'GPT-5' level step-change in 2024, actual progress was a horizontal expansion where models became cheaper, faster, and more multimodal. This represents a shift from raw model intelligence to a massive explosion in cost efficiency and inference-time compute.

    From [Ride Home] Simon Willison: Things we learned about LLMs in 2024 (episode)

  7. The Slop Filter

    This framework evaluates whether AI-generated output is useful or just noise by identifying if the content is unrequested and checking if a human has reviewed it for accuracy. It serves as a tool to filter out low-quality AI generation.

    From [Ride Home] Simon Willison: Things we learned about LLMs in 2024 (episode)

    • Says much the same Startup Ideas Podcast Host — The 'No AI Slop' SkillThis is a tool designed to act as an editor that strips out 'AI-sounding' patterns like overly polished or formulaic writing. It aims to preserve the user's unique…These 5 Github Repos are a goldmine

Every idea names the summary it came from. Written by a local model from this mind’s own work, never from memory.

Interviews 3

Open Simon Willison in Originals