What I read.

Research institute investigating key trends and questions about the trajectory and governance of AI.

Independent model evaluations to understand the AI landscape and choose the best model for your use case.

Nonprofit dedicated to accelerating AGI development through human-calibrated benchmarks that measure the gap between human and AI capabilities.

Model Evaluation and Threat Research — measuring AI ability to complete long, complex tasks.

Independent benchmarks of LLMs for tasks that mimic real industry use cases in legal, finance, math, and more.

Comprehensive AI model rankings, timeline of AI and language models, and AI IQ testing research.

A small research group forecasting the future of AI.

Benchmark evaluating AI models on managing a simulated vending machine business over one year, testing long-term coherence and strategic negotiation.

A comprehensive visual guide covering everything you need to know about AI agents.

The AI software engineer — an autonomous agent that can handle full engineering tasks end-to-end.

Superintelligent AI agents for scientific discovery — specializing in literature search, synthesis, and chemistry experiment planning.

Startup focused on fully automating work through advanced AI agents.

A general-purpose AI agent capable of autonomous task completion across a wide range of domains.

Comprehensive guide to prompt engineering — understanding capabilities and limitations of large language models.

Anthropic's guide to prompt engineering for Claude, emphasizing best practices and why it's faster and more flexible than fine-tuning.

Discover, refine, build, and test prompts — with examples of effective prompting approaches from the Wharton Generative AI Labs.

Platform for democratizing AI knowledge and skills through expert-led learning, community collaboration, and up-to-date AI content.

Wharton professor Ethan Mollick examines the implications of the AI era for work and education.

Anthropic's agentic AI coding tool — the tool used to build this website and my other projects.

AI-powered UI generation — natural language to production React/Next.js components.

Google's guide to vibe coding — both exploratory and professional applications, including workflows and lifecycle stages.

Detailed, accessible 3.5-hour explanation of how large language models like ChatGPT are built, trained, and fine-tuned.