Stance on GenAI
Our stance on the use of Generative Artifical Intelligence, like LLMs in our products, services, and our internal work. A clear statement of what we do, what we don't, and why.
Generative AI (aka. GenAI) comes in every flavor you could ask for, from chatbots to image, audio, video and code generation. It's an extremely useful and a very rapidly evolving technology, across the broad range of multimodal, it enables faster work while maintaining or improving quality.
GenAI expands what small teams like ours can do, and lowers the cost of expertise otherwise out of reach. We may make use of generative AI across our products, services, communications and strive to be transparent about it.
The individual incentive to use GenAI is often positive while the collective consequences can be a net-negative. We've thought carefully about this and believe that it deserves nuance.
What we believe The honest tension — useful individually, potentially harmful in aggregate Where the Math/Calculator analogy holds and where it breaks down The Jagged Frontier — why knowing the limits matters more than knowing the capabilities How we use it What we use it for and where What we don't use it for, and why Moderation in practice (make this concrete, not a principle) Our commitment We label AI-generated content The /resources/* guarantee The (i) button — how to check any page When we fall short (the small team caveat + contact) Commissioned work (clear handoff of responsibility) A note on why this document exists We're at the bleeding edge of this. That's exactly why we think the stance matters.
Last updated: