Posted on Monday 1 Feb 2021. 2,011 words, 13 links. By Matt Webb.
An untitled post from Mar 2011
Posted on Thursday 17 Mar 2011. 1,130 words, 10 links. By Matt Webb.
What Children Can Do and AI Can't
Psychologist Alison Gopnik tells three stories about intelligence — a golem, a pot of stone soup, and a digital child — to explain what AI still can't do.
Functions as Objects - Part 1 | Daryl Ducharme | Offprint
This was originally posted on Jan. 15, 2007 and is being posted here for historical reasons
Why mental metaphors do not help us understand chatbot mistakes
The function of chatbots like OpenAI’s ChatGPT is based on detecting probabilistic patterns in the training data. This makes them vulnerable to generating factual mistakes in their outputs. Recently, it has become commonplace in philosophical, scientific, and popular discourses to capture such mistakes by metaphors that draw on discourses about the human mind. The two most popular metaphors at present are hallucinating and bullshitting. In this paper, we review, discuss, and criticise these mental metaphors. By applying conceptual metaphor theory, we provide numerous reasons why they do not succeed in providing us with a better understanding of factual chatbot mistakes. We conclude by calling for justifications of the epistemic feasibility and fruitfulness of the metaphors at issue. Furthermore, we raise the question what would be lost if we stopped trying to capture factual chatbot mistakes by mental metaphors.
The AI Aesthetic
Writing about the big beautiful mess that is making things for the world wide web.
Seven Sketches in Compositionality: An Invitation to Applied Category Theory
This book is an invitation to discover advanced topics in category theory through concrete, real-world examples. It aims to give a tour: a gentle, quick introduction to guide later exploration. The tour takes place over seven sketches, each pairing an evocative application, such as databases, electric circuits, or dynamical systems, with the exploration of a categorical structure, such as adjoint functors, enriched categories, or toposes. No prior knowledge of category theory is assumed. A feedback form for typos, comments, questions, and suggestions is available here: https://docs.google.com/document/d/160G9OFcP5DWT8Stn7TxdVx83DJnnf7d5GML0_FOD5Wg/edit
Magic Paper
Working notes by Michael Nielsen, November 2017. Followup to (but doesn't require) my notes on Chalktalk.
An appreciation for (technical) architecture
Posted on Saturday 28 Mar 2026. 995 words, 14 links. By Matt Webb.
Aaron Steven White
computational semanticist. into modular synths and rum. https://aaronstevenwhite.io
Mindless Machines, Mindless Myths | Los Angeles Review of Books
Erik J. Larson thinks about “Mindless: The Human Condition in the Age of Artificial Intelligence,” which traces Robert Skidelsky’s philosophical reckoning with AI, automation, and the illusion of progress.
:quality(75)/https%3A%2F%2Fassets.lareviewofbooks.org%2Fuploads%2FMindless.jpg)
Mindless Machines, Mindless Myths | Los Angeles Review of Books
Erik J. Larson thinks about “Mindless: The Human Condition in the Age of Artificial Intelligence,” which traces Robert Skidelsky’s philosophical reckoning with AI, automation, and the illusion of progress.
:quality(75)/https%3A%2F%2Fassets.lareviewofbooks.org%2Fuploads%2FMindless.jpg)
Five Manifestos for the Beautiful World: The Alchemy Lecture 2023
The Alchemy Lecture 2023
On thinking machines
While Chiron Codex is about the application of LLMs and AI-augmented tools, we also need to understand their meaning to us, each other, and society. I have three topics: intelligence,...
Position: Stop Anthropomorphizing Intermediate Tokens as Reasoning/Thinking Traces!
Intermediate token generation (ITG), where a model produces output before the solution, has become a standard method to improve the performance of language models on reasoning tasks. These intermediate tokens have been called \say{reasoning traces} or even \say{thinking traces} -- implicitly anthropomorphizing the traces, and implying that these traces resemble steps a human might take when solving a challenging problem, and as such can provide an interpretable window into the operation of the model's thinking process to the end user. In this position paper, we present evidence that this anthropomorphization isn't a harmless metaphor, and instead is quite dangerous -- it confuses the nature of these models and how to use them effectively, and leads to questionable research. We call on the community to avoid such anthropomorphization of intermediate tokens.
