About Envision

Making scientific papers click through interactive simulations.

Why I Built This

I kept hitting the same wall with research papers. I'd read the papers, follow the math, nod along—and still not really get it. The understanding was fragile. Ask me to explain it a week later and I'd struggle.

Then I'd find a good visualization or interactive demo, and suddenly everything clicked. Seeing attention weights shift as I changed the input. Watching noise gradually destroy and then recover an image. That's when concepts moved from "I've read about this" to "I understand this."

Envision is my attempt to create more of those moments—for myself first, and hopefully for others who learn the same way.

A hat tip to Distill.pub—the gold standard for interactive scientific communication. Their work showed what's possible when you treat explanation as a craft. Envision aspires to that same spirit, focused specifically on making individual papers accessible.

What Envision Is

Each page on Envision takes a single influential paper and turns its core ideas into something you can play with. Not a summary. Not a tutorial. An interactive explanation that lets you build intuition by experimenting.

Simulations, not animations

You control the parameters. Change inputs and see what happens. The goal is exploration, not passive watching.

Papers, not topics

Each page focuses on a specific paper—its key insight, its mechanism, its impact. Always linked back to the original source.

Intuition before math

Build understanding through interaction first. Formulas come after you already have a mental model of what's happening.

Sources cited

Empirical claims link to papers. Equations reference their source. You can always verify and go deeper.

How It's Made

Every simulation on Envision is bespoke. There's no template that spits these out. Each one is crafted to match the specific paper's concepts—what visualization will make this idea click?

Yes, I use AI tools to help with code and writing—shout out to my intern Claude. But there's always a human in the loop—reading the paper, deciding what to emphasize, testing whether the simulation actually teaches, iterating until it feels right. The goal isn't volume; it's quality.

I'd rather publish a handful of visualizations that genuinely help people understand than infinite AI-generated slop.

The Structure

Each paper follows a similar arc, designed to build understanding progressively:

  1. 1

    The Problem

    What challenge was the paper trying to solve? Why did existing approaches fall short?

  2. 2

    The Key Insight

    The core idea—usually one sentence that makes everything else follow.

  3. 3

    The Mechanism

    How it actually works. This is where the interactive simulation lives.

  4. 4

    The Math

    Formulas and equations—but only after you have intuition for what they mean.

  5. 5

    The Impact

    What did this paper enable? Why does it matter?

Get In Touch

This is very much a work in progress. I'd love to hear from you:

  • • Spotted an error? Please tell me. Accuracy matters.
  • • Want a specific paper? I'm always looking for suggestions.
  • • Have feedback on a simulation? What helped? What confused you?
  • • Just want to say hi? That's great too.

Use the feedback button in the bottom-right corner of any page. It's the fastest way to reach me, and I read everything.

No login required. Just type and send.