Every blog post that took off on Hacker News, ranked by the votes of its best submission. 3,241 posts.
Stability AI emphasizes the importance of open models in a letter coinciding with the United States Senate hearing on AI oversight. Openness and transparency will improve safety, foster competition, and ensure strategic leadership in critical AI capabilities. Discover how open models can democratize
This guide provides an overview of practical Object Detection applications, its main challenges as a Machine Learning problem and how Deep Learning has changed the way to tackle it.
Keras was initially released a year ago, late March 2015. It has made tremendous progress since, both on the development front, and as a community.
Copilot may be a glimpse into the future of AI, but it comes with a lot of legal and security baggage.
YC is at our best when backing founders with an ethos of building for others, including strong open source developers like the founders of Continue [https://continue.dev] and many others. Earlier this week, a startup in the current batch, PearAI, was criticized online for an inappropriate approach to open source…
NVIDIA today launched the NVIDIA Vera CPU, the world’s first processor purpose-built for the age of agentic AI and reinforcement learning — delivering results with twice the efficiency and 50% faster than traditional rack-scale CPUs.
We use ChatGPT to allow a user to intuitively control multiple robots such as drones, robot arms, home assistant robots using only language.
We introduce AMD's AI Tensor Engine for ROCm (AITER), our centralized high performance AI operators repository, designed to significantly accelerate AI workloads on AMD GPUs
How can we build machines with human-level intelligence? There’s a limit to how far the field of AI can go with supervised learning alone. Here's why self-supervised learning is one of the most promising ways to make significant progress in AI.
A new post-training training quantization paradigm for diffusion models, which quantize both the weights and activations of FLUX.1 to 4 bits, achieving 3.5× memory and 8.7× latency reduction on a 16GB laptop 4090 GPU. Code: https://www.github.com/mit-han-lab/nunchaku
Posted by Adam Gaier, Student Researcher and David Ha, Staff Research Scientist, Google Research, Tokyo When training a neural network to accompl...
Sarvam open-sources 30B and 105B reasoning models trained in India: MoE LLMs with leading Indian-language benchmarks; weights on AI Kosh and Hugging Face.
As I was getting ready to write this post, I thought back to my childhood (largely spent watching TV) and some of the popular computer and robot voices of the 1960’s and 1970’s. In just a few minutes, pleasant memories of HAL-9000, B9 (Lost in Space), the original Star Trek Computer, and Rosie (from The […]
Understanding the safety impact of Automated Driving Systems (ADS) is crucial for their widespread adoption, yet robust real-world evaluation remains a critical area of development. By leveraging insurance industry third-party auto liability claims, we analyze the Waymo ADS performance across 25.3 million fully…
Political bias in AI measures where every major AI model stands on charged political and ethical questions: run many times, no web search, plotted with error...
We used agents to port all of Git to library-first, memory-safe Rust and it passes the entire C Git test suite.
Applications are now open for OpenAI Grove Cohort 2, a 5-week founder program designed for individuals at any stage, from pre-idea to product. Participants receive $50K in API credits, early access to AI tools, and hands-on mentorship from the OpenAI team.
Machine Learning has made huge strides in the last few decades, but it remains largely inaccessible to the average developer. This post examines some of the reasons why.
A safety-filtered pretrain can duck a charged word without refusing. It puts a fraction of the probability an open-data pretrain puts there. We call that...
Here's a common question we get from prospective Cline users: "How does Cline handle large codebases? Do you use RAG to index everything?" It's a reasonable question. Retrieval Augmented Generation (RAG) has become the go-to solution for giving AI systems access to large knowledge bases. But for Cline, we've taken a…
Doing small network scientific machine learning in Julia 5x faster than PyTorch ...
Astra is the first OpenAI model to meet the Critical cybersecurity capability threshold under the Preparedness Framework, with stronger safeguards for release.
Update (October 2024): In response to numerous requests from external developers, we have open-sourced a version of our Enzyme to React Testing Library (RTL) conversion tool. You can now find it on npm, along with detailed instructions on how to integrate and use it in your projects. In the world of frontend…
Finding that 70% of attention heads and 20% of feed-forward networks can be excised with minimal effect on in-context learning suggests that large language models are undertrained.
Trillium, Google’s sixth-generation Tensor Processing Unit (TPU) is now GA, delivering enhanced performance and cost-effectiveness for AI workloads.
Repository drawgent by yanndegat.tngl.sh on Tangled.
We’ve seen incredible adoption of SAM 3 over the last few months, and during that time, we’ve been working behind the scenes on updates to improve video processing efficiency. Today, we’re pleased to introduce SAM 3.1.
