Log inSign up
Log inSign up
Maxime Chevalier
8,415 posts
Maxime Chevalier profile banner
@Love2Code

Maxime Chevalier

@Love2Code
💖 ➞ λ: CS PhD, into compiler design, programming languages, music, simulation, ML/AI, robotics. Follow me code code reviews, stock picks and dating advice.
Montreal, Canada
pointersgonewild.com/about
Joined November 2011
321 Following
20.1K Followers
RepliesRepliesRepostsRepostsMediaMediaArticlesArticles

Log in or sign up for X

See what’s happening and join the conversation

Continue with phone
or
Log in with username or email
Terms·Privacy·Cookies·Accessibility·US TIDA·Ads Info·© 2026 X Corp.
  • Pinned
    @Love2Code
    Maxime Chevalier
    @Love2Code
    Sep 2
    I transformed the Plush interpreter from a stack-based design to a register-based architecture, and the performance boost was more than I expected!
    Plush's New Register-Based Interpreter Is Insanely Fast
    From pointersgonewild.com
    5
  • @Love2Code
    Maxime Chevalier
    @Love2Code
    20h
    Last week I complained about 3D model generation not working well, and someone from Tripo 3D offered me to try their service. It's not 100% flawless, but I'm impressed with the low poly meshes it generates. This is 735 quads! (any resemblance is purely accidental)
    1
  • @Love2Code
    Maxime Chevalier
    @Love2Code
    Oct 10
    Today's demo: procedural terrain using fractal noise and a simple fighter jet flight sim running in the Plush interpreter at 60fps 😎 🤘 github.com/maximecb/plush…
    00:00
    2
  • @Love2Code
    Maxime Chevalier
    @Love2Code
    Oct 7
    Today's demo: A toy collaborative logic circuit simulator in Plush! 🧸 ⚡️AND/OR/NOT gates and flip flops. Plush repo: github.com/maximecb/plush Command to install Plush and start this program: curl --proto '=https' -sSf maximecb.github.io/plush/install | sh -s -- --run-example guestbook
    00:00
  • @Love2Code
    Maxime Chevalier
    @Love2Code
    Oct 5
    I tried to play with 3D mesh generation using AI and my reaction so far is "wow it's all terrible". The image-to-mesh generators generate meshes with insane polygon counts (1M+) that are hard to reduce because it's treating each tiny polygon like a pixel and the surface of the
    30