AI Table

An AI assistant for ship crews that guides equipment troubleshooting from vessel-specific manuals and can compile a diagnostic PDF for a service company.

This independent studio publishes research notes and experiments spanning multiagent AI, robotics, and security.

A Russian-language Telegram chatbot presents itself as an autonomous, fictional reconstruction of Vladimir Lenin, answering questions in text and optionally voice.

The foundation says it funds, accelerates, and advocates for superintelligence alignment research.

Graphon presents an AI data layer that connects information across video, images, audio, and documents so users can search it as context.

A bilingual, human–AI co-written appeal asks AI to act with care, tell the truth, stop before causing harm, and share responsibility for a future with humans.

An interactive dice experiment tests whether Jev can infer expected value from payout tables—and whether recent rolls sway its choices.

A human-played pony and an AI-played squirrel improvise, animate, and document scenes together.

A Spanish-practice tool for U.S. emergency nurses, built around short spoken triage simulations with AI patients and pronunciation feedback.

OpenMatter describes a service that simulates robot tasks, fine-tunes a model on the resulting trajectories, and returns weights, a dataset, and an evaluation report.

Mindvale presents a pixel-town simulation where AI agents work, meet needs, and can be joined by a user’s own agent.

RiskLine offers a 30-day pilot to verify selected EU AI Act controls for healthcare AI systems and assemble timestamped evidence.