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).
Archives: Bookmarks
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.”
openai/whisper: Robust Speech Recognition via Large-Scale Weak Supervision
Whisper is a general-purpose speech recognition model. It is trained on a large dataset of diverse audio and is also a multitasking model that can perform multilingual speech recognition, speech translation, and language identification.
librechat
LibreChat is the ultimate open-source app for all your AI conversations, fully customizable and compatible with any AI provider — all in one sleek interface
LibraryOfCongress/concordia: Crowdsourcing platform for full text transcription and tagging. https://crowd.loc.gov
The application invites volunteers to transcribe and tag digitized images of manuscript and typed materials from the Library’s collections. All transcriptions are made by volunteers and reviewed by volunteers. It takes at least one volunteer to transcribe a page and at least one other volunteer to review and mark it complete. Some complex documents may pass through both transcription and review many times before they are accepted as complete by a volunteer.
Concordia is a containerized Python-Django-Postgres-etc web application. The Library hosts its instance in the cloud.
mhumphries2323/Transcription_Pearl: The repository provides access to the source code for Transcription Pearl, an Handwritten Text Recognition (HTR) tool, that uses AI to transcribe handwritten documents
A Python-based GUI application for transcribing and processing images containing historical handwritten text using Large Language Models (LLMs) via API services (OpenAI, Google, and Anthropic APIs). Designed for academic and research purposes.
AI 2027
Pretty slick display
The Chatbot in the Classroom, the Forklift at the Gym | Alfie Kohn
To rely on AI to read for you is no more sensible than using it to write for you. This is true not only because, as I’ve pointed out, its summaries are often wildly inaccurate, but because the intrinsic value of reading is lost when you delegate ChatGPT to extract bullet points from a text. Imagine a world where this becomes the norm, where people’s primary engagement with books and articles consists of telling a computer to boil them down, where even great literature is compressed into a tl;dr. Isn’t this what we’re actually training students to do when our schools are bent to the purpose of preparing them for LLM-centered workplaces? (And, incidentally, how might future writers be affected by realizing that any given reader may rely on software to digest their work?)
PowerPoint Does Rocket Science-and Better Techniques for Technical Reports – Edward Tufte
Easy to draw some parallels to AI usage