AI Explained is a London-based YouTube channel by a creator known as Philip that provides hype-free coverage of AI developments, capabilities, and safety topics for a general audience.
AI Explained is a London-based YouTube channel by a creator known as Philip that provides hype-free coverage of AI developments, capabilities, and safety topics for a general audience.
People
Updated 05/18/26Creator and host
Funding Details
Updated 05/18/26- Annual Budget
- -
- Current Runway
- -
- Funding Goal
- -
- Funding Raised to Date
- -
Org Details
Updated 05/18/26AI Explained is a YouTube channel based in London, England, created and hosted by a British creator who goes by Philip (publicly identified only as Philip L). The channel launched in early 2023 and quickly built an audience around its approach of applying journalistic rigor to AI coverage without hype or sensationalism. Philip's stated mission is to provide hype-free AI journalism at a moment when the field increasingly needs it, covering the arrival of increasingly capable AI systems with clarity and analytical depth. The channel covers a wide range of AI topics: major model releases from OpenAI, Anthropic, Google DeepMind, and Meta; LLM benchmark performance and its limitations; AI safety and alignment research; and the broader societal implications of transformative AI. Philip is known for producing thorough breakdowns of AI research papers and for investigative content on how frontier AI systems actually perform. In August 2024, Philip released SimpleBench, a multiple-choice text benchmark designed to expose limitations in LLM reasoning. SimpleBench contains over 200 questions focused on spatio-temporal reasoning, social intelligence, and linguistic adversarial robustness. On the benchmark, a human baseline of 83.7% significantly outperforms state-of-the-art models including o1-preview (41.7%). The benchmark is hosted at simple-bench.com and was covered widely in AI media. Philip also founded lmcouncil.ai, a multi-model workspace that allows users to compare simultaneous responses from multiple frontier AI systems (GPT, Claude, Gemini, Grok, and others). He runs an AI Insiders community of over 1,000 professionals across 30 industries, authors the Signal to Noise newsletter, hosts the AI Explained Official Podcast, and teaches a Coursera course on AI terminology. The channel had approximately 410,000 subscribers and 19.7 million total views as of early 2026, with 144 videos published. It is funded through YouTube ad revenue and Patreon memberships, with 1,251 paid Patreon members across two tiers ($9/month for AI Insiders content and $29/month for additional podcast and direct messaging access). No institutional grants from EA-aligned funders (SFF, LTFF, Open Philanthropy) were found.
Theory of Change
Updated 05/18/26AI Explained operates on the theory that public understanding of AI capabilities and risks is a crucial lever for better AI outcomes. By providing rigorous, hype-free coverage of frontier AI developments to a broad audience, the channel helps technical and non-technical observers accurately assess AI progress and the urgency of safety considerations. Philip also contributes directly to technical evaluation through SimpleBench, which aims to give the community better tools for measuring the true reasoning gap between AI systems and humans, supporting more calibrated expectations about AI capabilities.
Grants Received– no grants recorded
Updated 05/18/26Projects
Updated 05/18/26AI Explained Official Podcast is a show hosted by Philip, creator of the AI Explained YouTube channel, that offers hype-free coverage of frontier AI news and the arrival of smarter-than-human AI, drawing on his SimpleBench-driven evaluations of model reasoning.
LLM Council is a web platform that lets users access multiple leading AI language models—Claude, ChatGPT, Gemini and Grok—from a single interface and compare their responses side-by-side using a Council mode.
SimpleBench is a multiple-choice text benchmark introduced in August 2024 by Philip, creator of the AI Explained YouTube channel, in collaboration with Hemang, to evaluate large language models on basic common-sense reasoning tasks where humans still outperform them.
Discussion
No comments yet. Be the first to share your thoughts.