Directory
Published directory entries matching your search.
SwarmVault is a local-first tool that turns documents, code, transcripts, notes and URLs into a markdown wiki and a local knowledge graph for AI agents. It follows a three-layer architecture based on a pattern described by Andrej Karpathy. A first run requires no API keys, using a built-in heuristic provider that runs locally and offline.
A GitHub Blog post by Nicole Choi, based on discussions with Alireza Goudarzi and Albert Ziegler, about the emerging architecture of LLM applications. It covers five steps to building an LLM app, including focusing on a single problem and choosing the right LLM, and notes that LLM outputs are probabilistic and not predictable.
The copyright conundrum of AI art is an article from The Verge examining copyright issues related to generative AI. It covers image generation tools including Midjourney, Stable Diffusion, and DALL-E, discusses questions of AI art ownership, and notes that copyright concerns impact generative AI development around training data and output control.
The Radical AI Podcast is a show distributed via iTunes, Spotify, and iHeartRadio. It publishes episodes, show notes, and transcripts, alongside community collaborations, panel content, and series focused on technology, power, and democracy-related AI topics.
A visualisation of major large-language models ranked by performance using the MMLU benchmark, which consists of 16,000 multiple-choice questions across 57 academic subjects. It also presents selected lawsuits filed against AI companies and notes that MMLU has some critiques.
A page from AI Digest presenting a timeline of AI forecasts, describing what to expect in AI capabilities, potential harms, and society's response, according to the median prediction of forecasters on Metaculus and Manifold. Forecasters predicted when statements about the future would come true or the probability they would occur by a given date. The page notes the forecasts have limited precision and were last updated June 26, 2026.
A blog post presenting findings and tips on fine-tuning Stable Diffusion with the Dreambooth technique using Diffusers. It describes experiments with hyperparameters and settings, recommending a low learning rate and progressively more training steps, and notes that faces need more steps and prior preservation.
A guide describing the AdaMPI model, which converts single, in-the-wild 2D images into 3D photos. It covers inputs and outputs and a step-by-step usage walkthrough, and notes limitations such as complex scenes with intricate 3D geometry that may not convert perfectly.
Research report examining seven years of truck-involved fatalities, 2018 through 2024, across 19 major US cities, based on FARS data. It compares each city's 2018-2019 baseline average with its 2023-2024 average and notes that of 17 Vision Zero adopters, 10 improved and 7 did not.
Voice Dream Reader is a text-to-speech app that reads PDFs, web pages, and books aloud. It supports importing PDFs, EPUBs, Word documents, PowerPoint files, and web pages, with highlighting, notes, offline listening, and syncing across iPhone, iPad, and Mac. Accessibility features include VoiceOver support, Dynamic Type, Voice Control, and braille display support.
A Gary Marcus article on his Marcus on AI Substack asking when the GenAI bubble might burst and how that could happen within 12 months. It cites survey findings on generative AI and security, notes $50B in and $3B out, and argues that hallucinations remain and that the industry is built on hype.
A Hackaday article by Matthew Carlson explaining why LLaMa matters. It describes LLaMa as a transformer language model from Facebook/Meta research spanning 7 billion to 65 billion parameters, and notes that llama.cpp runs inference with 4-bit quantization.
An essay by Arjun Ramani and Zhengdong Wang arguing that transformative AI is hard to achieve, assembling technical, social and economic arguments. It notes that trend growth of GDP per capita in frontier economies has never exceeded three percent per year, and that major technical hurdles remain despite rapid progress in some subfields.
Zoo is a web guide or update reference covering changes, screenshots, notices, or community notes.
Zoo Code is a GitHub repository for a tool that provides a team of AI coding agents integrated directly into a code editor. The repository includes information about migration from Roo Code to Zoo Code, new features added since that migration, and version notes for v3.84.0.
