Also hilarious if anyone thinks I've become a big fan of agent-led Rust out of any other consideration than purely scientific inquiry. Do you know how badly I would hate having to personally write Rust?? But when the premise changes, you need to update your priors.
John Hartquist retweeted
Reject non-understanding.
Slop grenades are not just code and PRs. We run a real risk that *reading* gets entirely discounted, because of the exhaustion of getting low-quality and unverified AI prose being thrown at you all the time.
There's a thread going viral about a 'perf improvement' attributed to a compiler change. The PR *description* itself has the AI telling this person it's *not* due to the compiler change, it's how the algorithms and data structures have changed. And the implication is that a similar win could have been attained otherwise.
I want AI in the service of understanding the universe and enhancing human cognition and creativity.
John Hartquist retweeted
Have you ever wanted to train LLMs in pure C without 245MB of PyTorch and 107MB of cPython? No? Well now you can! With llm.c:
github.com/karpathy/llm.c
To start, implements GPT-2 training on CPU/fp32 in only ~1,000 lines of clean code. It compiles and runs instantly, and exactly matches the PyTorch reference implementation.
I chose GPT-2 to start because it is the grand-daddy of LLMs, the first time the LLM stack was put together in a recognizably modern form, and with model weights available.
John Hartquist retweeted
Introducing AI Brainstore!
It’s a proof-of-concept of a brain for an AI agent.
Ask an agent a question, it checks its memories for an answer, otherwise it browses the web and learns the answer.
As it learns, it saves memories to its brain.
GitHub: github.com/mckaywrigley/ai-b…
John Hartquist retweeted
Nobody is better positioned to make a great living in the AI/ML space than software developers.
If you know how to build software, you have some of the most valuable skills necessary to make an impact.
John Hartquist retweeted
🔥 New (1h56m) video lecture: "Let's build GPT: from scratch, in code, spelled out."
youtube.com/watch?v=kCc8FmEb…
We build and train a Transformer following the "Attention Is All You Need" paper in the language modeling setting and end up with the core of nanoGPT.
John Hartquist retweeted
We’ve acquired the MuJoCo physics simulator (mujoco.org) and are making it free for all, to support research everywhere. MuJoCo is a fast, powerful, easy-to-use, and soon to be open-source simulation tool, designed for robotics research: dpmd.ai/mujoco-blog
John Hartquist retweeted
Announcing a new open-source JavaScript library for exploratory data visualization: Observable Plot! observablehq.com/@observable…
John Hartquist retweeted
Three.js Journey is live! 🎉
Learn how to create stunning 3D WebGL websites with the #threeJs library through more than 30 hours of video for $95 only (VAT incl.)
threejs-journey.xyz
#threejsJourney #webGL
John Hartquist retweeted
Our featured report today is by @johnhartquist, who shows us how to run a reproducible hyperparameter search.
He provides good intuition around which hyperparams are really key for getting your audio classification models to perform well.
Read it here: wandb.ai/jhartquist/fastaudi…
John Hartquist retweeted
The first featured report in the gallery today focuses on audio!
@johnhartquist wrote a beautiful report on finetuning ResNet-18 for audio classification on the ESC-50 dataset.
Report: wandb.ai/jhartquist/fastaudi…
#DeepLearning #dl #ArtificialIntelligence #100DaysOfMLCode


















