121
OpenAI o1 system card
This report outlines the safety work carried out prior to releasing OpenAI o1 and o1-mini, including external red teaming and frontier risk evaluations according to our Preparedness Framework.
Top posts
Every blog post that took off on Hacker News, ranked by the votes of its best submission. 673 posts.
122
Microsoft's Recall AI feature is now indefinitely delayed
Update Oct. 31, 2024: We are committed to delivering a secure and trusted experience with Recall. We recently shared updates to the security and privacy architecture for Recall in a
123
Extracting concepts from GPT-4
Using new techniques for scaling sparse autoencoders, we automatically identified 16 million patterns in GPT-4's computations.
124
My finetuned models beat OpenAI's GPT-4
Finetunes of Mistral, Llama3 and Solar LLMs are more accurate for my test data than OpenAI’s models.
125
What I learned from looking at 900 most popular open source AI tools
[Hacker News discussion, LinkedIn discussion, Twitter thread]
126
Planner programming blows my mind
Picat is a research language intended to combine logic programming, imperative programming, and constraint solving. I originally learned it to help with vacation scheduling but soon discovered its planner module, which is one of the most fascinating programming models I’ve ever seen. First, a brief explanation of…
127
LLMs use a surprisingly simple mechanism to retrieve some stored knowledge
Researchers find large language models use a simple mechanism to retrieve stored knowledge when they respond to a user prompt. These mechanisms can be leveraged to see what the model knows about different subjects and possibly to correct false information it has stored.
128
Introducing Copilot+ PCs
An on-demand recording of our May 20 event is available. Today, at a special event on our new Microsoft campus, we introduced the world to a new category of Windows PCs designed for AI, Copilot+ PCs. Copilot+ PCs are the fastest, most intelligent Windows PCs ever built. With powerful new silicon capable of an…
129
How Meta trains large language models at scale
As we continue to focus our AI research and development on solving increasingly complex problems, one of the most significant and challenging shifts we’ve experienced is the sheer scale of co…
130
Do AI companies work?
The market needs to be irrational for you to stay solvent.
131
Zed AI
From the Zed Blog: Powerful AI-assisted coding powered by Anthropic's Claude, now available.
132
133
How I run LLMs locally
I document how I run Large Language Models locally.
134
Eagle 7B: Soaring past Transformers
A brand new era for the RWKV-v5 architecture and linear transformer's has arrived - with the strongest multi-lingual model in open source today
135
Hello OLMo: A truly open LLM
Introducing OLMo 7B, a truly open, state-of-the-art large language model.
136
AI's $600B Question
The AI bubble is reaching a tipping point. As we continue to follow the GPUs, navigating what comes next will be essential.
137
Getting 50% (SoTA) on Arc-AGI with GPT-4o
You can just draw more samples
138
Can AI do maths yet? Thoughts from a mathematician
So the big news this week is that o3, OpenAI’s new language model, got 25% on FrontierMath. Let’s start by explaining what this means.
139
Opus 1.5 released: Opus gets a machine learning upgrade
This 1.5 release is unlike any of the previous ones. It brings many new features that can improve quality and the general audio experience. That is achieved through machine learning. Although Opus has included machine learning — and even deep learning — before (e.g. for speech/music detection ), this is the first time…
140
Many AI safety orgs have tried to criminalize currently-existing open-source AI
<p>I've seen a few conversations where someone says something like this:</p> <blockquote> <p>I've been using an open-source LLM lately -- I'm a huge fan of not depending on OpenAI, Anthropic, or Google. But I'm really sad that the AI safety groups are trying to ban the kind of open-source LLM that I'm using.</p>…
141
Use Prolog to improve LLM's reasoning
On one side, LLMs show unseen capabilities in reasoning, but on the other - reasoning in LLMs is not ideal.
142
Diffusion models from scratch, from a new theoretical perspective
This tutorial aims to give a gentle introduction to diffusion models, with a running example to illustrate how to build, train and sample from a simple diffusion model from scratch.
143
Byte Latent Transformer: Patches Scale Better Than Tokens
We introduce the Byte Latent Transformer (BLT), a new byte-level LLM architecture that, for the first time, matches tokenization-based LLM performance at...
144
Francois Chollet is leaving Google
Today, we're announcing that Francois Chollet, the creator of Keras and a leading figure in the AI world, is embarking on a new chapter in his career outside of Google.
145
Web scraping with GPT-4o: powerful but expensive
tl;dr; show me the demo and source code!
146
Ollama is now available on Windows in preview
Ollama is now available on Windows in preview, making it possible to pull, run and create large language models in a new native Windows experience. Ollama on Windows includes built-in GPU acceleration, access to the full model library, and serves the Ollama API including OpenAI compatibility.
147
AI generated security reports about curl
I have held back on writing anything about AI or how we (not) use AI for development in the curl factory. Now I can't hold back anymore. Let me show you the most significant effect of AI on curl as of today - with examples. Bug Bounty Having a bug bounty means that we offer … Continue reading The I in LLM stands for…
148
TinyML: Ultra-low power machine learning
Explanation of TinyML, what it is used for and the main techniques it uses
149
150
Google’s AI thinks I left a Gatorade bottle on the moon
Google’s NotebookLLM is the first AI podcast I’d actually listen to. Give it a web page or document, and it will generate you a podcast about it.