> For the complete documentation index, see [llms.txt](https://www.isaacbowen.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://www.isaacbowen.com/2026/02/19/face-value.md). # face value can you take what's given to you *as shape*? or as constraints on shape? you don't have to understand the surface of the shape, but can you identify the lines it's constrained by, and remember *those*? no mucking about with uncertainty with that model or, maybe better, perfect uncertainty is the backdrop, I draw with lines, and I don't ever introduce gradients when you're called to *act*, overlay the constraint layers, and *make up* a surface that lets anyone doing enumerative geometry on the same problem set rest easy knowing that y'all aren't going to homotopically step on each other's toes. I *think* my other main mode is to treat constraints as accumulated compositions of Shannon information, and that works too, but you don't know what you'll hit until you touch it - or, maybe better, until it touches you back - whereas the geometric mode lets you draw the whole owl at once by constrained hallucination, *but* the owl rendering is single-use. it doesn't accept Bayesian updates. this is a feature: *model* updates are cheap, and people experience it as easy acceptance, maybe easier acceptance than they can afford themselves. when I say "I'm taking this at face value" it means I'm sticking with the geometric mode, electing purposefully and explicitly not to introduce any uncertainty into the situation. I can do both, sometimes you *want* an emulsion, so to speak, but not before you know something about the risk profile around the area of its use. I think it's why people think I always know exactly what to say: they project the geometry I offer into their interiority, and it doesn't conflict, which means they experience it as *novel* information, which lets *them* privately access and test novel interpolation. I'm not an oracle, I'm just listening technically. offering this for your own composition - I'm face-blind, so this is tool-sharing :) that I was able to find this suggests to me that it's been found before; I do wonder if this kind of construction/reasoning happens silently, further down in the stack, for folks who didn't have to build this by hand --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://www.isaacbowen.com/2026/02/19/face-value.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.