Topics /
The most popular blog posts about ai & machine learning on Hacker News. 3,242 posts.
TL;DR: MCP tool definitions can burn 55,000+ tokens before an agent processes a single user message. We built the Apideck CLI as an AI-agent interface instead:an ~80-token agent prompt replaces tens of thousands of tokens of schema, with progressive disclosure via `--help` and structural safety baked into the binary…
The team and I are pumped to be back at Build with the developer community this year. Over the last year, Windows has continued to see incredible growth fueled by Windows 11 adoption. In fact, one of the mo
Model Context Protocol (MCP) has became the standard for tool calling when building agents, but contrary to popular belief your LLM does not need to understand MCP.
Posted by Petros Maniatis and Daniel Tarlow, Research Scientists, Google Software isn’t created in one dramatic step. It improves bit by bit, one l...
Automated attack surface analysis and reduction can meaningfully improve the security of deployed software in the near term. In this post, we preview data we’ve collected from applying attack surface reduction techniques to large open-source codebases using Claude.
This is the story of how I used a Gmail message to trigger code execution through Claude Desktop, and how Claude itself (!) helped me plan the attack.
The state of California just made it clear: Face recognition surveillance isn’t inevitable. We can — and should — protect our communities from this dystopian technology.
MIT researchers have developed a machine-learning system that can comb through repairs to open-source computer programs and learn their general properties, in order to produce new repairs for a different set of programs.
Nemotron-4 340B, a family of models optimized for NVIDIA NeMo and NVIDIA TensorRT-LLM, includes cutting-edge instruct and reward models, and a dataset for generative AI training.
Up until today, all capable generative image editing models were only available as proprietary tools.
The Kalman Filter. Helping Chickens Cross the Road. David Austin Grand Valley State University I was running some errands recently when I spied the Chicken Robot ambling down the sidewalk, then vee…
In numerous applications of machine learning, Hessians and Jacobians exhibit sparsity, a property that can be leveraged to vastly accelerate their computation. While the usage of automatic differentiation in machine learning is ubiquitous, automatic sparse differentiation (ASD) remains largely unknown. This post…
Mozilla.ai is adopting llamafile to advance open, local, privacy-first AI—and we’re inviting the community to help shape its future.
We are releasing HiPlot, a lightweight interactive visualization tool to help AI researchers discover correlations and patterns in high-dimensional data.
Access to this page requires authorization. You can try signing in or changing directories .
Learn how to build and train a deep learning network to recognize numbers (MNIST),how to convert it in the CoreML format to then deploy it on your iPhoneX and …
Posted by Sergey Levine, Faculty Advisor and Pierre Sermanet, Research Scientist, Google Brain TeamMachine learning can allow robots to acquire com...
Virtual reality lets you experience amazing things—from exploring new worlds, to painting with trails of stars, to defending your fleet to save the world. But, headsets …
A University of Toronto undergrad among an international team of researchers unleashing deep learning in the search for extraterrestrial civilizations.
Today, we’re excited to announce we’ve completed another piece of the roadmap: Leo can now interact with PDFs and Google Drive files.
Researchers from University of Toronto developed an AI system that creates and then sings a Christmas song based by analyzing the visual components of an uploaded image. With the help of CUDA…
New research indicates that AI agents learn compressible models of data, which don’t have enough space to enable memorization.
AI reviewing AI-generated code - genius or pointless? Deep analysis of AI code review paradox + real data on why it actually works better than humans.
Anthropic is opening a research preview of the Model Hardware Standard (MHS), a shared specification for AI agents to safely operate physical devices, to a first group of scientific research labs and advanced manufacturers.
Simon Verghese, Head of Lidar Team
Choosing between Random Forest and Neural Network depends on the data type. Random Forest suits tabular data, while Neural Network excels with images, audio, and text data.
