New public funding and company credits target biological threats. The next test is delivery, access controls and independent oversight.
Read the brief →
A virtual village where AI characters talk and socialize. An open-source starting point for experimenting with simulated worlds; supports several model providers, including OpenAI-compatible APIs.
An inspectable template collection for summarizing URLs, text, and PDFs. Adapt one input path first, then compare its summaries with the original documents.
A Next.js demo of voice agents that call tools and hand work to specialists. Compare the chat-supervisor and sequential-handoff patterns.
Making an experiment usable by someone else
Understanding realtime events and voice interactions
Discovering and trying community experiments
A practical approach to comparing AI use with delivery time, quality and business outcomes. Includes review and correction time, setup and support costs; its ROI example is illustrative, not a measured earnings result.
Details →A guide to setting a monthly project spending ceiling that pauses resources when reached. Useful for experiments that create cloud infrastructure; compare a real stop with a notification-only budget alert.
Details →Compare subscriptions, usage-based billing, tiers, and hybrid models. Useful for a software product whose model and support costs change with customer usage.
Details →Your first conversations with ChatGPT
Improving a prompt you actually use
A structured introduction to AI at work
Building a foundation through working examples
Keeping the evidence behind your work
Deciding what an always-on agent may do
Understanding AI beyond a single product
Reading a stack of sources
Moving from chat prompts to working agents
Allow Google Analytics to measure visits and link clicks? It’s optional. Privacy details
