Every blog post that took off on Hacker News, ranked by the votes of its best submission. 3,092 posts.
The CEOs of OpenAI and Anthropic have both claimed that human-level AI is just around the corner — and at times, that it's already here. These claims have generated enormous public attention. There has been some technical scrutiny of these claims, but critiques rarely reach the public discourse. This piece is a sketch…
Recently, I have been testing how well the newest generations of large language models (such as GPT-5 or Claude 4.5) handle natural language, specifically counting characters, manipulating characters in a sentences, or solving encoding and ciphers. Surprisingly, the newest models were able to solve these kinds of…
A year ago, Mihail Eric wrote a blog post detailing his experiences working on AI inside Amazon: How Alexa Dropped the Ball on Being the Top Conversational System on the Planet. It’s a great …
In the previous article in this series we distinguished between two kinds of unsupervised learning (cluster analysis and dimensionality reduction) and discussed the former in some detail. In this installment we turn our attention to the later. In dimensionality reduction we seek a function $f : \mathbb{R}^n \mapsto…
It's been 7 months since my last commentary on the field, and as it became regular appearance in this blog (and in fact many people apparently enjoy this form and keep asking for it),
Marko Elez , a 25-year-old employee at Elon Musk’s Department of Government Efficiency (DOGE), has been granted access to sensitive databases at the U.S. Social Security Administration, the Treasury and Justice departments, and the Department of Homeland Security. So it should fill all Americans with a deep sense of…
The past week was a momentous occasion for protein structure prediction, structural biology at large, and in due time, may prove to be so for the whole of life sciences. CASP14, the conference for …
One of my favorite ¹ concepts for multi-dimensional data visualization is the Chernoff Face . The idea here is that for a dataset with many...
Protein “language” is a lot like human language. Given the similarities, researchers have been building and training language models on protein sequence data, replicating the success seen in other domains, with profound implications. In this post, I will explore how transformer models have been applied to protein data…
Practical guide to improving LLM finetuning with better instruction datasets, covering data curation, prompt-output pairs, synthetic data, and experiment ideas.
Matthew Honnibal is a computational linguist from Sydney based in Berlin, Germany. He's the author of the spaCy Natural Language Processing library and the co-founder of Explosion.
The lead in the last month or so was an indicator of having good chances, but there was a huge shuffling of ranks in the last week and some last minute casualties.
In high-dimensional vector spaces, numerical optimization is omnipotent.
For some inane reason, Github classifies me as a "major open source maintainer", which means I get a free copilot subscription.1 I've been using it for a...
We have been learning quickly from how people are using Sora and taking feedback from users, rightsholders, and other interested groups. We of course spent a lot of time discussing this before...
From magical to frustrating in months. Why AI coding agents feel like dial-up internet and what ultra-fast inference could unlock for developer productivity.
Serious learning requires friction, frustration... and other humans
A little AI tells detector helps us promote good writing.
Back in 2009 I posted a simple Mandelbrot fractal viewer on the web: a single HTML file with inline Javascript. Just 329 lines of code, each pixel a tiny table cell. Click to zoom. Watch it iterate. That was about it!
I’ve been coding entirely with Claude Code for the past two months. At first it was exhilarating. I was speeding through tasks. I was committing like mad.
The evolution of the AI landscape, and YC's Copilot era.
I wrote earlier about a neural net-powered dungeon crawling text adventure game called GPT-2-Adventure in which gameplay is incoherent and dreamlike, as you encounter slippery sign text, circular passages, and unexpected lozenge rooms. A PhD student named Nathan trained the neural net on classic dungeon crawling…
Building a simple neural net in Java.
Be careful what you wish for...
Shape From Silhouette Reconstruction from a short cell phone video, showing reconstructed masks.
Deep neural networks have been shown to be vulnerable to the adversarial example phenomenon: all models tested so far can have their classifications dramatically altered by small image perturbations . The following predictions were for instance made by a state-of-the-art network trained to recognize celebrities :
