Intelligence meets information.
Language models, structured data, and the ways intelligent systems work with knowledge.
Models and data
Applied AI at the frontier.
Applied AI research lab focused on frontier intelligence.
Exploring what intelligence can do.
Explore research
What does a model actually know?
Model behavior & evaluation
How does intelligence meet information?
Language models & data
When can a model act usefully?
Agents & workflows
What changes outside the API?
Models & deployment
What makes capability usable?
Infrastructure
Where can intelligence make a difference?
AI in serious domains
We study models, data, and the systems that bring intelligence into the world.
Understanding models. Understanding their limits.
Model behavior, evaluation, and reliability. Questions about what a model can do, when it can be trusted, and what the evidence actually supports.
Explore model behaviorFrom the lab
Research notesOn governance at the substrate level
Governance should examine decisions made during model development as well as the conditions of deployment. A useful proposal states what an existing institution can already do and identifies the specific gap it would address.
Designing evaluation beyond public leaderboards
An evaluation program needs a way to change as models and tasks change. The useful design question is how to combine reusable comparisons, held-out work and targeted investigation within an explicit budget.
Compute, data and the shape of the next generation
Training and inference are separate, interacting uses of compute. Deciding where to spend a budget requires a workload, an evaluation target and an explicit account of data and serving costs.
A closer look at intelligence.
Explore the lab's interests, research questions, and approach to work that can be examined.
About Telasian