- ergs0204/LatexInNotebooklm: A tempermonkey for rendering latex in Notebooklm
A Tampermonkey userscript that adds LaTeX rendering capability to Google’s Notebooklm using KaTeX. This allows you to write mathematical formulas with standard LaTeX syntax using $formula$ delimiters. - Jacob Andreas @ MIT
I’m interested in language as a communicative and computational tool. People learn to understand and generate novel utterances from remarkably little data. Having learned language, we use it acquire new concepts and to structure our reasoning. Current machine learning techniques fall short of human abilities in both their capacity to learn language and learn from language about the rest of the world. My research aims to understand the computational foundations of language learning, and to build general-purpose intelligent systems that can communicate effectively with humans and learn from human guidance. - [2401.03910] A Philosophical Introduction to Language Models — Part I: Continuity With Classic Debates
Large language models like GPT-4 have achieved remarkable proficiency in a broad spectrum of language-based tasks, some of which are traditionally associated with hallmarks of human intelligence. This has prompted ongoing disagreements about the extent to which we can meaningfully ascribe any kind of linguistic or cognitive competence to language models. Such questions have deep philosophical roots, echoing longstanding debates about the status of artificial neural networks as cognitive models. This article — the first part of two companion papers — serves both as a primer on language models for philosophers, and as an opinionated survey of their significance in relation to classic debates in the philosophy cognitive science, artificial intelligence, and linguistics. We cover topics such as compositionality, language acquisition, semantic competence, grounding, world models, and the transmission of cultural knowledge. We argue that the success of language models challenges several long-held assumptions about artificial neural networks. However, we also highlight the need for further empirical investigation to better understand their internal mechanisms. This sets the stage for the companion paper (Part II), which turns to novel empirical methods for probing the inner workings of language models, and new philosophical questions prompted by their latest developments.
- Zoom, video conferencing, energy, and emissions – David Mytton
This is the kind of complexity I think is required to think through the ecological impact of technology in context . . . no one will be happy with this.—
A lifecycle analysis from the Danish government shows that compared to a standard plastic bag, an unbleached paper bag is better when considering impacts on climate change, but it would need to be reused 43 times to have a positive impact on all environmental factors. This is because “environment” means many things: pollution, ozone depletion, causes of human cancer, acidification of freshwater, resource depletion, and others. The worst thing you can do is replace a plastic bag with organic cotton which requires 20,000 reuses before it is better for the environment.
AND
If this happened in the US between two people in New York, the emissions factor is similar to the UK at 0.28839 kgCO2 per kWh but if it was between two people in Chicago, that would be 0.56191 kgCO2 per kWh. Location matters.
- Belief, Uncertainty, and Truth in Language Models – YouTube
- Deloitte was caught using AI in $290,000 report to help the Australian government crack down on welfare after a researcher flagged hallucinations | Fortune
- Grade Inflation: What Goes Up Must Come Down | Opinion | The Harvard Crimson
“Grades A and B are sometimes given too readily — Grade A for work of no very high merit, and Grade B for work not far above mediocrity … One of the chief obstacles to raising the standards of the degree is the readiness with which insincere students gain passable grades by sham work.”This is from the “Committee on Raising the Standard” in 1894.
- n8n-io/n8n: Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.
- Defeating Nondeterminism in LLM Inference – Thinking Machines Lab
What might be more surprising is that even when we adjust the temperature down to 0This means that the LLM always chooses the highest probability token, which is called greedy sampling. (thus making the sampling theoretically deterministic), LLM APIs are still not deterministic in practice (see past discussions here, here, or here). Even when running inference on your own hardware with an OSS inference library like vLLM or SGLang, sampling still isn’t deterministic (see here or here). - Socrates, Memory & The Internet | Issue 122 | Philosophy Now
In his book In Search of Memory (2006), Kandel writes that for a memory to persist, “the incoming information must be thoroughly and deeply processed. This is accomplished by attending to the information and associating it meaningfully and systematically with knowledge already well established in memory.” Without paying attention to a thought or experience, the neurons involved lose their excited state of electrical activity within a few seconds, and the memory is gone, leaving maybe only a small trace in the mind. - Socrates on the Invention of Writing and the Relationship of Writing to Memory : History of Information
“For this invention will produce forgetfulness in the minds of those who learn to use it, because they will not practice their memory. Their trust in writing, produced by external characters which are no part of themselves, will discourage the use of their own memory within them. You have invented an elixir not of memory, but of reminding; and you offer your pupils the appearance of wisdom, not true wisdom, for they will read many things without instruction and will therefore seem [275b] to know many things, when they are for the most part ignorant and hard to get along with, since they are not wise, but only appear wise.”