Topics /
The most popular blog posts about ai & machine learning on Hacker News. 6,334 posts.
Sparse Autoencoders (SAEs) have recently become popular for interpretability of machine learning models (although sparse dictionary learning has been around since 1997). Machine learning models and LLMs are becoming more powerful and useful, but they are still black boxes, and we don’t understand how they do the…
AI is now able to decompile games and software in a matter of days. That's scary... but kind of exciting.
Executives love AI while ICs stay skeptical. The divide comes down to non-determinism: who is evaluated on judgment versus deterministic execution.
Anisha Sharma Marketing Specialist, RISC-V International Anisha is part of the RISC-V International marketing team, responsible for managing social media and tracking the latest updates from our members. She brings...
Thriving in a world of agents
The comfort we get when offloading our cognitive load to LLMs is bad for us. Cognitive load should exist, and if we reduce it too much – if we stop thinking – we can actually unlearn how to think. Kids who always choose the easy route and copy their homework from other students eventually find […]
In the past I used AI code editors for all of my programming, but I stopped using it and recommend others to consider this as well
If you asked someone in 2018 what a "small model" was, they'd probably say something with a few million parameters that ran on a Raspberry Pi or your phone. Fast-forward to today, and we're calling 30B parameter models "small"—because they only need ...
comma.ai gets punished for being open, so why would anyone even bother?
AutoGPT decided to drop the support for vector databases: Pinecone, Milvus, Redis, Weaviate - all gone. What has happened to the original idea of efficiently storing agent's memories?
We run a multi-tenant Rails application with sensitive data and layered authorization. In this post, I walk through how I added the first AI agent tool using RubyLLM, Pundit policies, and our existing Algolia search, without introducing a parallel system or loosening constraints.
DeepSeek R1, available through the model catalog on Microsoft Azure AI Foundry and GitHub, enables businesses to seamlessly integrate advanced AI.
We used Multislice Training to run the world’s largest LLM distributed training job on a compute cluster of 50,944 Cloud TPU v5e chips.
Our latest ANN release supports scales of 100+ billion vectors in a single search index, with 200ms p99 query latency at 1k QPS and 92% recall.
AI agents are here—from ChatGPT Operator to coding tools like Devin and Lovable, there’s a rapid growth of products that can interact with websites on our behalf. It’s time to start building for a new persona: the autonomous agent. Instead of just designing experiences for human users, we need to think carefully about…
Research has begun to reveal many algorithms can be expressed as matrix multiplication, suggesting an unrealized connection between linear algebra and computer science. I speculate graphs are the missing piece of the puzzle. Graphs are not only useful as cognitive aides, but are suitable data structures for a wide…
We trained a model to predict speech from ultrasound video of the tongue, no speaking required. It reads open text, generalizes across people, and reaches a 15.6% word error rate on open-vocabulary speech.
Google has chosen to send a response to my article to Barry Schwartz of Search Engine Roundtable. Here is my response. (1) On the March 2019 core update claim in the piece: This is baseless speculation. The March 2019 core update was designed to improve the quality of our search
Posted by Philip Sun, Software Engineer, Google Research Suppose one wants to search through a large dataset of literary works using queries that r...
This isn’t a joke, and it’s becoming a big problem
Why neurons have thousands of synapses, a theory of sequence memory in neocortex Hawkins & Ahmad, Front. Neural Circuits 2016 It all began with a fascinating lunchtime conversation with Martin …
This post will give some example comparisons running Llama 2 uncensored model versus its censored model.
For two weeks last July, I cocooned myself in a hotel in Portland, OR, living and breathing probabilistic programming as a “student” in the probabilistic programming summer school run by DARPA. The…
This is a high level post about algorithms (especially mathematical, scientific, and data analysis algorithms) which I hope can help people who are not researchers or numerical software developers better understand how to choose and evaluate algorithms. When scientists and programmers think about efficiency of…
What in the world is Explorative Modeling?
The frequentist paradigm enjoys the most widespread acceptance forstatistical analysis. Frequentist concepts such as confidence intervalsand p-values dominat...
Applying photo geolocation estimation to the game of GeoGuessr.
