Next-token prediction in JavaScript
What inspired this project today was watching this amazing video by 3Blue1Brown called "But what is a GPT?" on Youtube (https://www.youtube.com/watch?v=wjZofJX0v4M - I highly recommend watching it). I added it to the repo for reference. When it clicked in my head that "knowing a fact" is nearly synonymous with predicting a word (or series of words), I wanted to put it to the test, because it seemed so simple. I chose JavaScript because I can exploit the way it structures objects to aid in the modeling of language. For example: "I want to be at the beach", "I will do it later",…
In plain words
Next-token prediction in JavaScript is a language model implementation that predicts the next word in a sequence based on training data. Built in JavaScript, it leverages the language's object structure to efficiently store and retrieve word patterns for prediction tasks. The project demonstrates how language understanding can be modeled as a word prediction problem, using JavaScript's fast object lookup to find known sentence patterns rather than recursive text searching.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
What inspired this project today was watching this amazing video by 3Blue1Brown called "But what is a GPT?" on Youtube (https://www.youtube.com/watch?v=wjZofJX0v4M - I highly recommend watching it). I added it to the repo for reference. When it clicked in my head that "knowing a fact" is nearly synonymous with predicting a word (or series of words), I wanted to put it to the test, because it seemed so simple. I chose JavaScript because I can exploit the way it structures objects to aid in the modeling of language. For example: "I want to be at the beach", "I will do it later", "I want to know the answer", ... becomes: { I: { want: { to: { be: { ... }, know: { ... } } }, will: { ... } }, ... } in JavaScript. You can exploit the language's fast object lookup speed to find known sentences this way, rather than recursively searching text - which is the convention and would take forever or not work at all considering there are several full books loaded in by default (and it could support many more). Accompanying research yielded learnings about what "tokens" and "embeddings" are, what is meant by "training", and most of the rest - though I'm still learning jargon. I wrote a script to iterate over every single word of every single book to rank how likely it is that word will appear next, if given a cursor, and extended that to rank entire phrases. The base decoder started out what I'll call "token-agnostic" - didn't care if you were looking for the next letter... word... pixel... it's the same logic. But actually it's not, and it soon evolved into a text (language) model. But I have plans to get into image generation next (next-pixel prediction), using this. Overall the concepts are similar, but there are differences primarily around extraction and formatting. Goals of the project: - Demystify LLMs for people, show that it's just regular code that does normal stuff - Actually make a pretty good LLM in JavaScript, with a version at least capable of running in a browser tab
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 16d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 26d ago · cactuscompute.com
Launched alongside, April 2024
the whole month →
Supabase▲2,328The Postgres developer platform is now generally available
Dev tools · 2024 · supabase.com
Cal.com Platform▲1,602Build your pixel-perfect booking experience with Atoms
Dev tools · 2024 · cal.com
deco.cx 2.0▲1,080Build web apps 10x faster with Deno, JSX, TS & Tailwind
Dev tools · 2024 · decocms.com
