TLDR
Grant Sanderson (3Blue1Brown) demonstrates his custom Python animation library Manim by walking Ben Sparks through a live coding session. They build a 'Hello World' example, then animate the Lorenz Attractor — a chaotic 3D system — showing Grant's interactive workflow: writing code in Sublime Text, testing instantly via an IPython terminal with a checkpoint system, and rendering to MP4. Grant also explains the split between his personal Manim version and the community-maintained fork.
For you
PRO members can sign in and request more information — or their own version of this digest, tuned to what they care about.
Sign in to unlockSurprising Insights
Grant uses ChatGPT to generate boilerplate for numerical solvers (e.g., SciPy's solve_ivp wrapper), saving hours of API friction.
The interactive IPython embedding has a scope bug: functions defined inside don't see outer variables unless you run `globals().update(locals())` — a 'cursed' hack Grant admits is terrible practice but necessary in this dev mode.
Grant avoids GitHub Copilot for Manim because he thinks in code, not English; he prefers dumb autocomplete over AI suggestions that guess wrong.
Manim's 'anything can transform into anything' philosophy means a letter 'H' can smoothly morph into a circle by matching point sets — a core design principle.
The 'glow dot' and 'tracing tail' effects are reusable, polished primitives Grant built years ago and still uses; they handle fade, stroke, and lifetime automatically.
Key takeaways
Manim is a programmatic, Python-based animation engine built by Grant for 3Blue1Brown; everything is code, not keyframes.
There are two versions: Grant's personal fork (cutting-edge, less documented) and Manim Community (better tested, documented, recommended for newcomers).
Grant's workflow centers on an interactive loop: write code in a text file, highlight lines, hit a shortcut to execute them in a live IPython session attached to the scene — like a hybrid of script and notebook.
A custom 'checkpoint paste' system caches scene state at labeled points, letting him re-run mid-scene sections without replaying from the start.
The Lorenz Attractor demo shows how Grant uses SciPy's ODE solver for the math, then maps numerical solutions into Manim's coordinate system for animation.
Chaos is visualized by evolving 10 nearly identical initial conditions; they track together briefly then diverge wildly, illustrating sensitive dependence.
Camera moves are animated programmatically (e.g., slow 3D orbit), and LaTeX equations can be color-coded by variable and pinned to the camera frame.
Final rendering is a command-line call (with pre-run error check), outputting 4K MP4s dropped into Final Cut for editing.
Notable quotes
“The most common question I get about 3Blue1Brown is, what do I use to animate the videos? And the short answer is that I wrote a custom Python library, its name is Manim, so it's all programmatic and it's also very bespoke.”
▸ Notable quotes · 0:00“I've always posted all of the code that I make for videos. It's openly visible on GitHub. And I also made the tool itself, the underlying tool, Manim, open source. But I don't, as a personality type, really have the constitution to manage an open source project.”
▸ Notable quotes · 1:46“Some viewers might be wondering why I don't just have it all in a Jupyter notebook. But I really like having the scene in a single text file and being able to interact with it through scripts and things like that.”
▸ Notable quotes · 6:33“A different philosophy also with Manim was like anything can transform into anything.”
▸ Notable quotes · 7:52“The way I started here, I went to ChatGPT... write me a Python function using some numerical ODE solver to find a solution of the Lorenz equations.”
▸ Notable quotes · 12:35“One of the reasons I wanted to make any kind of behind the scenes... the way I used to use Manim... is you're always running it from the command line... So it was later in the game... to then also have kind of the interactive shell version, such that the process of creating just is like highlight the code, see what the code does.”
▸ Notable quotes · 43:02“Thank you for showing me I really enjoyed seeing the back end like it's a very human experience coding and what and the messiness is part of it isn't it”
▸ Notable quotes · 52:34Chapters
0:00 — Introduction: What is Manim and the two versions: Grant explains Manim is his custom Python animation library, open-sourced but personally maintained; a community fork exists with better docs and testing.
2:38 — Hello World: Basic scene structure and interactive workflow: Grant shows a minimal scene: a Python class with a construct method, adding shapes, and using a Sublime Text shortcut to send code to a live IPython terminal for instant visual feedback.
6:30 — Checkpoint paste: Notebook-like iteration in a text file: Grant explains his 'checkpoint paste' system that caches scene state at labeled comments, letting him re-run mid-scene sections without full replay — a hybrid of script and notebook.
11:00 — Project: Animating the Lorenz Attractor (chaos theory): They build a 3D Lorenz Attractor animation: define ODEs, use SciPy to solve, map points to Manim coords, draw curves with ShowCreation, add tracking dots with updaters, and visualize divergence of nearby initial conditions.
33:30 — Camera control and 3D rendering: Grant shows how to animate the camera frame (slow 3D orbit) using a clipboard-captured orientation and the `.animate` syntax for smooth transitions.
35:40 — Polishing: Colors, tails, and LaTeX equations: They refine colors (color gradients matching initial conditions), add fading tracing tails via a reusable 'tracing_tail' primitive, and overlay LaTeX equations with per-variable coloring pinned to the camera.
42:00 — Rendering to MP4 and final workflow: Grant runs the CLI render command (with pre-run error check), outputs 4K MP4, and mentions dropping into Final Cut. He also shows example scenes repo for learning Manim patterns.
52:00 — Conclusion and resources: Grant wraps up, links GitHub repo with all video code, promises workflow docs in README, and hints at future Manim live streams on Patreon.
Speakers
Grant Sanderson — Creator of 3Blue1Brown and Manim
Ben Sparks — Mathematician and Numberphile contributor