Marcelo de G. Malheiros
banner
mgmalheiros.bsky.social
Marcelo de G. Malheiros
@mgmalheiros.bsky.social
Patterns. Simulation. Graphics. Compilers. Creative Coding.

mgmalheiros.github.io

PhD in Computer Science. Academic researcher at UEM.

Also at https://mathstodon.xyz/@mgmalheiros
Reposted by Marcelo de G. Malheiros
Absolutely worth a read. Taking a shot at defining what makes a game AAA or Indie (without calling it 'indie') by using actual data.
The install footprint + length of credits seems to be as good a metric as any.
Love me some graphs
#gamedev
hushcrasher.substack.com/p/taxonomy-o...
By the way, what’s a AA?
Instead of fuzzy labels like 'indie' and 'AA', we propose a new data-driven system to classify video games based on their actual production scope. This new framework reveals the real economic and crea...
hushcrasher.substack.com
November 18, 2025 at 10:39 AM
Reposted by Marcelo de G. Malheiros
Astrophotographer snaps skydiver falling in front of the sun | Discussion
“The Fall Of Icarus”: You Have Never Seen An Astrophotography Picture Like This!
This is not photoshopped. That’s really a person falling in front of the Sun.
www.iflscience.com
November 17, 2025 at 10:40 PM
Reposted by Marcelo de G. Malheiros
Graphics Programming weekly - Issue 417 - November 15th, 2025 www.jendrikillner.com/post/graphic...
November 17, 2025 at 2:55 PM
Reposted by Marcelo de G. Malheiros
so you're saying it has out of the box support for git, vim and c++, and it's foss? ok fine I'll try zed
November 17, 2025 at 3:43 PM
Reposted by Marcelo de G. Malheiros
❤️ Love this ❤️

@mathforge.org has done what I’ve never been organised enough to do myself, and catalogued every #geometrypuzzle I’ve ever posted on social media. It looks amazing!
November 17, 2025 at 6:33 AM
Reposted by Marcelo de G. Malheiros
In Praise of Useless Robots | Discussion
In Praise Of Useless Robots
The most intriguing robots aren’t built to work, but to make us imagine other worlds.
thereader.mitpress.mit.edu
November 16, 2025 at 1:20 PM
Reposted by Marcelo de G. Malheiros
This year we have sold 31% of the advent calendars that we sold last year 😓

The art is so pretty and colorful 🥺
The facts are so good 🥺
Eels are so cool 🥺

Get an eel facts advent calendar at EelFacts.net
November 9, 2025 at 2:24 PM
Reposted by Marcelo de G. Malheiros
Running the "Reflections on Trusting Trust" Compiler | Discussion
research!rsc: Running the “Reflections on Trusting Trust” Compiler
research.swtch.com
November 16, 2025 at 2:40 PM
Reposted by Marcelo de G. Malheiros
Deezer ran a study with 9,000 listeners and found that only 3 percent could tell the difference between a fully AI-generated song and a human-made one.

I dig into the implications here:

medium.com/the-riff/dee...
Deezer Reports Only 3% of Users Can Accurately Identify AI-Generated Tracks
130 days' worth of AI “music” floods the platform every day
medium.com
November 14, 2025 at 7:26 PM
Reposted by Marcelo de G. Malheiros
Misinformation *is* central to democratic problems today, but the solution is *not* to correct such epistemic failures directly. Instead, we must address them at their roots, in social identity.

Or so I argue in a new post for @apaphilosophy.bsky.social.

blog.apaonline.org/2025/11/12/t...
The Problem is Epistemic. The Solution is Not. | Blog of the APA
Doubts about the wisdom of the masses are as old as philosophy itself. Yet interest in democracy’s “epistemic” merits has surged in the last decade—and it is no mystery why. Democracy is collapsing ar...
blog.apaonline.org
November 13, 2025 at 1:16 AM
Reposted by Marcelo de G. Malheiros
Go's Sweet 16 | Discussion
Go’s Sweet 16 - The Go Programming Language
Happy Birthday, Go!
go.dev
November 14, 2025 at 11:40 PM
Reposted by Marcelo de G. Malheiros
I think I understand how it can be that LLMs are both exceptionally good and quite terrible at programming. It's because there are two entirely different skillsets that we both call "good at programming." LLMs have only one of them.
blog.genesmindsmachines.com/p/llms-excel...
LLMs excel at programming—how can they be so bad at it?
My explanation for the mystery of why LLMs can be both exceptionally good and quite terrible at programming.
blog.genesmindsmachines.com
November 6, 2025 at 3:43 PM
Reposted by Marcelo de G. Malheiros
I found this knot pattern this week and generated a planar embedding by first using stereographic projection onto the disk, tracing, and tweaking—all in PowerPoint and Photoshop. I really enjoy this process as it allows me to experience the structure of the knot throughout the very slow process. 😊
November 14, 2025 at 2:25 AM
Reposted by Marcelo de G. Malheiros
Prefer Boring Technology | Discussion
3A29's webspace - Prefer boring technology.
in which it explains why it has a preference for boring technology.
itwont.work
November 14, 2025 at 2:40 AM
Reposted by Marcelo de G. Malheiros
Why Fei-Fei Li and Yann LeCun Are Both Betting on "World Models" | Discussion
Why Fei-Fei Li and Yann LeCun Are Both Betting on “World Models” — and How Their Bets Differ | entropytown
Gaussian splats, JEPA and Genie 3 — and why “world model” now means three different things at once.
entropytown.com
November 14, 2025 at 3:20 AM
Reposted by Marcelo de G. Malheiros
On my way home from #GodotFest. Thanks so much to everyone for this incredible event! Hope to see you all again soon :)

My presentation on foliage rendering in #Godot is already up on YouTube: www.youtube.com/watch?v=U2nB...

And I also shared my slides here: slides.com/hexaquo/plan...
Karl Bittner: Plants, Polygons and Pixels: Large-Scale Vegetation Rendering in Godot #GodotFest2025
YouTube video by GodotFest
www.youtube.com
November 13, 2025 at 1:42 PM
Reposted by Marcelo de G. Malheiros
The fact that OpenAI even tried to frame their defence in this absurd manner betrays how morally bankrupt they are. But perhaps also how desperate.
We didn't plagiarize, you made us plagiarize by asking questions to which we stole the answers.

"Because its output is generated by users of the chatbot via their prompts, OpenAI said, they were the ones who should be held legally liable for it – an argument rejected by the court."
ChatGPT violated copyright law by ‘learning’ from song lyrics, German court rules
OpenAI ordered to pay undisclosed damages for training its language models on artists’ work without permission
www.theguardian.com
November 14, 2025 at 12:12 PM
Reposted by Marcelo de G. Malheiros
I have updated my "Optimal Transport for Machine Learners" repository with a Pytorch illustration of Wasserstein gradient flows on pairwise interaction functionals (MMD distances) github.com/gpeyre/ot4ml
November 14, 2025 at 10:56 AM
Reposted by Marcelo de G. Malheiros
a warning from India for countries thinking of adopting a Digital ID model like Aadhar

rethinkaadhaar.in/campaign2025
Campaign 2025 — Rethink Aadhaar
rethinkaadhaar.in
November 14, 2025 at 1:01 PM
Reposted by Marcelo de G. Malheiros
Reposted by Marcelo de G. Malheiros
Visual Types
Interactive visualizations of TypeScript concepts. Behold pretty colors.
types.kitlangton.com
November 13, 2025 at 10:15 PM
Reposted by Marcelo de G. Malheiros
Proud to be part of this amazing tribute to HR GIGER by @spiridude.com and In official collaboration with the HR Giger Estate! 240 pages and over 100 artists!
The kickstarter campaign just started so you guys know what to do 🙂
www.kickstarter.com/projects/spi...
November 12, 2025 at 6:30 PM
Reposted by Marcelo de G. Malheiros
Someone at work just called the new Steam Machine a “GabeCube” and I can’t stop laughing.

steampowered.com/hardware
Steam Hardware
The Steam Hardware family officially expands in early 2026.
steampowered.com
November 12, 2025 at 7:27 PM