Favorites

Collection
of links
from around
the web

Sometimes
my own
links

Receive new
favorites
via RSS

Collection of links from around the web.
Sometimes my own links.

Baba is Eval↩︎ 05 Jul 2025 (Lennart Finke, July 2025)

Brilliant! He had the idea to use Baba is You as an LLM benchmark similar to ARC-AGI. He reversed engineered the game code and made an MCP server to expose the game state. I'm gonna play with it!

AGI Is Not Multimodal↩︎ 08 Jun 2025 (Benjamin A. Spiegel, June 2025)

Great article. A more rational take on AGI and multimodality compared to the previous link.

The library came alive↩︎ 06 Jun 2025 (Togelius, June 2025)

Are LLMs going to be enough, or are embodiment and agency necessary? A poetic view.

Lily58 and Lotus58 keyboard layouts↩︎ 08 May 2025 (diegobit, May 2025)

I'm enjoying this couple of split keyboards. The link points to the repo with my custom layouts to be used on the Vial app and some explanation of my choices.

How I use LLMs↩︎ 30 Mar 2025 (Andrej Karpathy, Feb 2025)

If you are not conviced, or you don't understand LLMs, ChatGPT etc, please watch this video. Karpathy has an exceptional ability to explain complex topics in simple non trivial ways. He recently started a series targeted at a general audience that explains all the fundamental building blocks and provides the correct mental models to understand ChatGPT, how it works and how it's trained.

The Only Unbreakable Law↩︎ 01 Mar 2025 (Casey Muratori, Mar 2022)

A hierarchy that creates information silos between teams prevents them from sharing information effectively, forcing each to work around the lack of communication. This constrains the space of possible software designs to those that mirror the organizational structure. When the structure naturally reflects the problem domain, great. But when it doesn't, we're ruling out the best architecture from the start. This video illustrates that perfectly, explaining it through the lens of Conway's Law, with plenty of examples.

On the measure of Intelligence↩︎ 27 Feb 2025 (François Chollet, Nov 2019)

Fundamental paper for AI research. Chollet tries to find a definition of intelligence that is not naive, not something that shifts every time there is a new model that beats humans at some narrow task (like chess, at go, or that writes like humans). He aims for a definition that lets us discuss AGI in a practical way, not the holy grail that solves all the problems on earth, but a concrete goal we can actually work toward. Brutally summarized: For Chollet, intelligence is the ability to generalize quickly beyond what is known; in his words: "skill-acquisition efficiency".

Neural network training makes beautiful fractals↩︎ 26 Dec 2024 (Jascha Sohl-Dickstein, Feb 2024)

Really fascinating to see fractals here. What makes it so is that this is a real phenomenon, like visualizations that dive deeper and deeper into the world of small-scale physics, not just a mathematical object. It reminds me of the beautiful ending of Fez.

The Fibonacci Matrix↩︎ 30 Nov 2024 (Ian Henry, Jul 2023)

I love visualization that give a new intuition about something that is hard, solely with algebra, and I love clever algorithmic optimizations. This article has both. This article will also make you wonder why you never new about the clever ways to compute it and why you had to write the Fibonacci sequence in countless unoptimized ways in University.