Blog Directory
Neural nets from first principles; the classic RNN and recipe posts.
Long, careful surveys: agents, RLHF, diffusion, hallucination.
The clearest visual explanations of LSTMs and representations.
The Illustrated Transformer, GPT-2 and BERT.
Ahead of AI: LLM research, explained with code.
ML systems in production, evaluation and agents.
Applied ML, recsys and LLM patterns from industry.
Evals, fine-tuning and debugging LLM products.
Daily hands-on notes on LLM tools and prompt injection.
Diffusion and generative modelling, in depth.
GPUs for deep learning, quantization, QLoRA.
Making deep learning go brrrr: performance from first principles.
Security of ML systems and honest takes on using LLMs.
Post-training, RLHF and open models.
GPT in 60 lines of NumPy.
CUDA matmul optimization, step by step.
The original paper, line by line, as runnable code.
Interpretability, alignment and model behaviour.
Model releases, reasoning and safety research.
Gemini, AlphaFold and research highlights.
Broad ML research across systems, NLP and vision.
Llama, JEPA world models and open research.
Practical guides on open models, training and inference.
Deep technical posts on training and inference.
Mechanistic interpretability research threads.
Berkeley AI Research: RL, robotics, vision, NLP.
Interactive, beautifully explained ML research (archive).