Loading…
About
Nanofuze is a free place to learn AI: learning paths and three full courses on building agents, machine learning and deep learning, and robotics and physical AI. And because the field moves every week, one page shows everything that changed in AI, each item with its numbers and its source.
How you learn here
Four habits that research keeps finding work better than reading and hoping. The courses and paths on our learning site are built around them.
01
Short chapters, three to ten minutes each, written to be understood on the first read: the idea, a picture, an example and what to remember.
02
Every example marked as runnable was executed, and the real output is shown beneath it. The numbers in the text are the numbers you will get.
03
Quizzes inside chapters and questions at the end make you recall answers. Everything you answer joins a review queue that brings it back just before you would forget it.
04
Nano, the assistant, knows the chapter you are on. Ask it to explain simply, quiz you or set you an exercise, without leaving the page.
Everything personal (your progress, notes, daily goal and reviews) is stored only in your browser. The learning site (learn.nanofuze.com) asks you to sign in with Google; progress is never saved to your account. Each course says what was run and what was only checked, and where a fact comes from a public source and has not been verified against the original, it says so.
How the news works
01
Every hour we read public data: model price lists, the human-preference leaderboard, Hugging Face, the release feeds of the software that runs AI, and the United States Federal Register.
02
We compare each reading with the one before. A price that changed, a model that entered the top ten, a version that shipped, a rule that was published: each is a difference, so each is an item of news.
03
Importance, from 1 to 5, comes from written rules, not from a language model. A price cut of half or more is a 4. A model that takes first place on the leaderboard is a 5. The rules are public.
04
Each item shows its numbers, says how we know, and links to the data it came from. Nothing is copied or rewritten from anywhere else.
5 public data sources are read; 350 company profiles give the news its context.
Our approach
The chapters are written to explain, with examples that run. We tell you where a claim comes from, mark what we have not been able to check against a primary source, and say plainly when a result is a small toy example rather than a benchmark.
The news does not republish, summarise or rewrite articles from other websites. When a model appears on a price list, a leaderboard moves or a new version ships, that is a fact in public data, and we show the fact with its numbers. We say what the data shows and not why.
Read how every kind of news is made on the data sources page, the rules we hold ourselves to in our standards, and tell us about any mistake on the corrections page.
Meet Nano
Nano lives on the learning site.
Ask it to explain the chapter you are on, quiz you, or set you an exercise. It answers from the courses, the glossary and the live model data.
Say or type where you want to go, such as a learning path, the glossary or the AI news, and Nano takes you there.
Turn on "Hey Nano" to talk without touching anything. Voice is optional: everything works by typing too.
Tips from readers show us topics a course should cover and sources we should be watching.