Grounds

AI research and software.

Grounds explores how neural networks learn, represent the world, and can be understood. We also build Workspace, a private AI application configured around a company's documents and workflows.

Explore Workspace Read the latest notes

Research

Research is at an early stage. Current explorations include training methods, interpretability, continual learning, and simulations.

Why it matters to us: Interpretability as a Guiding Principle, an essay from .

Grounds Workspace

A private AI workspace built around your company's documents and workflows. We configure your templates, reference material, and review steps. Your team can then draft from its best work in the company's own formatting, edit Word documents, and open cited sources beside a draft, all in one place.

See how Workspace works and how we set it up

A Workspace chat answer about a sample remodeling job that quotes the walkthrough notes and the company's tile and grout terms, with a link to each.
The cited terms document opened beside the answer, with the quoted sealing passage highlighted.
An answer in Workspace with its citation opened, and the cited document open at the highlighted passage. Sample company and data.

Latest

  1. Interpretability as a Guiding Principle

    Why understanding the internals of neural networks matters, why it is hard, and what Grounds is doing about it.

  2. Grounds goes live

    We are privately working with early customers and exploring core research directions.

All updates