Hubs

Knowledge cutoff is the date beyond which an AI model has no training data, and therefore no built-in awareness of events, releases, or facts that came after it. This hub tracks how leading models’ cutoff dates compare, why the gap matters for accuracy, and how techniques like web search and retrieval help close it.

LG AI Research is the artificial intelligence research arm of the LG Group, developing the EXAONE family of large language models and pursuing AI for corporate and industrial applications. As part of one of South Korea’s largest technology conglomerates, it aims to compete with leading Chinese and US labs by releasing open-weight models. This hub tracks LG AI Research’s model releases, research milestones, and role in the global AI landscape.

LimX Dynamics is a Shenzhen-based robotics company developing humanoid robots for industrial and commercial use, part of China’s fast-growing wave of embodied-AI startups. This hub tracks its funding rounds, product releases, and progress toward a public listing.

Liquid AI is a startup building compact, efficient models designed to run AI agents on phones and other devices without a cloud connection. This hub follows Liquid AI’s on-device models and releases.

Llama is Meta’s family of open-weight AI models, released with downloadable parameters so developers can run and fine-tune them on their own infrastructure. It is one of the most widely used open model families in the industry. This hub follows Llama releases, capabilities, and how it compares to other open and closed models.

A large language model (LLM) is an AI system trained on vast amounts of text to understand and generate human language. LLMs power ChatGPT, Claude, Gemini, and most modern AI assistants. This hub collects news and explainers on how LLMs work, how they are trained, and how the leading models compare.

As governments move to regulate artificial intelligence, AI companies, industry groups, and advocacy organizations are actively lobbying to shape the resulting laws and policies. This tag tracks who is spending on influence, what positions they take, and how that lobbying affects AI regulation worldwide.

Machine learning is the branch of AI where software learns patterns from data instead of following hand-written rules — the foundation beneath most modern AI systems. This hub covers the core ideas, methods, and news, from training and models to how machine learning shows up in real products. Expect plain-language explainers.

Mark Zuckerberg is the co-founder and CEO of Meta, which he has steered into a heavy bet on artificial intelligence alongside its social apps. This hub follows Zuckerberg’s decisions on Meta’s AI models, infrastructure spending, and strategy. Expect news and context on what his moves mean for the industry.

The Model Context Protocol (MCP) is an open standard that lets AI models connect to outside tools, files, and apps through a common interface instead of custom code for each one. This hub covers MCP’s adoption, tooling, and what it enables for AI agents.

Meituan is a large Chinese technology company, best known for on-demand services, that has become an active builder of open-source AI models. This hub tracks Meituan’s AI research, model releases, and its place in China’s AI landscape.

Meta is the US technology company behind Facebook, Instagram, and WhatsApp, and one of the largest developers of artificial intelligence. It builds the open-weight Llama models and the proprietary Muse family, runs the Meta AI assistant across its apps, and invests heavily in AI data centers and research. This hub gathers news and explainers about Meta’s models, products, and strategy.

Microsoft is one of the world’s largest technology companies and a central player in AI, through its Copilot assistants, the Azure cloud, and a deep partnership with OpenAI. This hub gathers news and explainers on Microsoft’s AI products, investments, and strategy.

Midjourney is an AI image-generation service that turns text prompts into images. This hub follows Midjourney’s product updates, pricing, creative uses, and the copyright disputes surrounding AI-generated art.

MiniMax is a Chinese AI startup building large language and multimodal models and consumer AI products. This hub tracks MiniMax’s models, funding, and its place among the fast-growing group of Chinese AI companies. Expect news and plain-language explainers.

Mira Murati is a technology executive and former chief technology officer of OpenAI, where she helped lead the development of ChatGPT and other products. She now runs Thinking Machines Lab, an AI startup staffed largely by former OpenAI researchers. This hub collects news and explainers about Murati and her work.

Mistral AI is a French artificial-intelligence company known for its open-weight language models and a strong presence in the European AI landscape. It builds general-purpose and specialized models, including tools for coding and mathematics. This hub collects news and explainers about Mistral’s models, releases, and its place among the leading AI labs.

Mixture of Experts (MoE) is an AI model design that holds hundreds of billions of parameters but activates only a small fraction of them for each query. This lets very large models run quickly and cheaply, and it underpins many of today’s frontier systems. This hub explains how MoE works and why it matters.

Model distillation is a technique in which a smaller, faster AI model is trained to imitate a larger, more capable one. This hub explains how distillation works, why it matters for cost and efficiency, and the disputes it raises when done without permission.

Moonshot AI is a Chinese AI startup best known for its Kimi chatbot and large language models, which emphasize very long context windows. This hub tracks the company’s model releases, funding, and product updates.

Multimodal AI describes models that process more than one kind of input — text, images, audio, and video together — rather than words alone. This lets a single system describe a photo, answer questions about a chart, or generate video from a prompt. This hub explains how multimodal models work and follows the systems that offer it.

NACE is the European Union’s standard classification system for economic activities, used to categorize businesses for statistics, business registration, and regulatory compliance. This hub covers what NACE codes mean and how AI tools are making it easier for businesses to find and assign the right one.

Naver is the South Korean internet company behind the country’s dominant search engine and the HyperCLOVA X family of AI models, as well as products like the LINE messaging app. This hub tracks Naver’s AI models, products, and business moves as it competes with US and Chinese AI labs.

Neuracle Technology is a Shanghai-based company developing implantable brain-computer interfaces. Its best-known product is the NEO system, the first implantable BCI device to win full commercial market approval in China. This hub tracks Neuracle’s device rollout, clinical trial results, and regulatory milestones.

Neural networks are the layered, brain-inspired computing structures that underlie most modern AI systems, from language models to image recognition. This hub covers how neural networks work, the architectures and training methods behind them, and news on the research pushing them forward. Expect plain-language explainers.

Our daily news briefs: short, source-linked updates on the most important AI developments, written in Georgian and English.

Natural language processing (NLP) is the branch of AI that lets computers understand, generate, and work with human language — the foundation behind chatbots, translation, and large language models. This hub covers the methods, research, and products built on NLP.

A non-human identity (NHI) is a digital identity issued to a machine rather than a person — a service account, API key, bot, or increasingly an AI agent acting on a company’s behalf. As organizations deploy more autonomous agents, these identities now outnumber human ones, creating a fast-growing attack surface. This hub covers the rise of non-human identities and the tools and practices used to secure them.

NotebookLM was Google’s AI-powered research and note-taking assistant, letting users ask questions and get summaries based only on the documents they uploaded. Google rebranded the tool as Gemini Notebook in 2026, adding features such as code execution. This hub gathers coverage published under the original NotebookLM name.

Nscale is a European AI cloud company that builds GPU-based data center infrastructure for large-scale AI training and inference. Positioned as a sovereign alternative to US hyperscalers, it has signed major compute deals with Microsoft, NVIDIA, and OpenAI to expand capacity across the UK and Europe.