# Superintelligence News

> Independent news and clear explainers on superintelligence: AI labs, research, government policy and the safety debate. Neutral news and plain-English explainers, edited by Umair (PhD in artificial intelligence, 10+ years in the field). The sources used are linked at the end of each article and AI assistance is disclosed. Every article shows its publication date and, if changed, an updated date.

## About
- [About us](https://superintelligence-news.com/about): who we are and what we cover
- [Editorial policy](https://superintelligence-news.com/editorial-policy): how we research, cite sources, correct errors and use AI tools
- [Corrections policy](https://superintelligence-news.com/corrections-policy): how errors are fixed
- [Authors](https://superintelligence-news.com/authors): the people and teams behind the articles
- [Contact](https://superintelligence-news.com/contact)

## Sections
- [Explainers](https://superintelligence-news.com/explainers): plain-English guides to the ideas behind the news

## Topics
- [Research & Models](https://superintelligence-news.com/categories/research): New AI models, research papers and benchmarks explained in plain English: what they can do, how they work and what they still cannot do.
- [Superintelligence](https://superintelligence-news.com/categories/superintelligence): What superintelligence is, how close today's AI is to it, who is trying to build it, and the debate over whether it can be made safe.

## Companies and people
- [Anthropic](https://superintelligence-news.com/companies/anthropic) (company): American artificial intelligence company, structured as a public benefit corporation and headquartered in San Francisco. Founded in 2021 by former OpenAI staff with a focus on AI safety, it develops the Claude family of large language models.
- [Google DeepMind](https://superintelligence-news.com/companies/google-deepmind) (company): Artificial intelligence research laboratory owned by Alphabet. Founded in the UK in 2010 and acquired by Google in 2014, it merged with Google Brain in 2023 to become Google DeepMind.
- [Nick Bostrom](https://superintelligence-news.com/people/nick-bostrom) (person): Philosopher known for his work on existential risk and the risks of superintelligence, and author of the 2014 book Superintelligence: Paths, Dangers, Strategies. He was the founding director of the Future of Humanity Institute at the University of Oxford.

## Articles
- [AI Agents vs Chatbots: What Is the Difference?](https://superintelligence-news.com/articles/ai-agents-vs-chatbots.md): Chatbots answer questions; AI agents pursue goals by planning, using tools and taking actions. Learn the differences, how agents work and where the risks are.
- [How Large Language Models Work: Tokens to Answers](https://superintelligence-news.com/articles/how-large-language-models-work.md): A clear walkthrough of how LLMs turn text into tokens, learn from data, and generate answers one token at a time, plus why they sometimes make things up.
- [What Is Superintelligence? A Plain-English Explainer](https://superintelligence-news.com/articles/what-is-superintelligence.md): What is superintelligence? Learn how ASI differs from AI and AGI, its potential benefits and risks, and why its arrival remains uncertain.

## Optional
- [Full text of all articles](https://superintelligence-news.com/llms-full.txt)
- [RSS feed](https://superintelligence-news.com/rss.xml)
- [Sitemap](https://superintelligence-news.com/sitemap.xml)