LMM Labs is exploring what comes next in AI—from more capable agents to better-connected, dependable systems. Building on our working applications, we use human creativity and technical judgment to guide research into ways to accelerate learning, development, and practical problem-solving.
Collapse graphic ▴ · or press Escape
Research directions
- Agentic workflows: Plan, act, evaluate, and refine—with human oversight at meaningful decision points.
- Connected knowledge: Explore retrieval-augmented generation (RAG), data integration, and tool connections through protocols such as MCP.
- Collaborating agents: Investigate how specialized agents can divide tasks, share findings, and coordinate results.
- Reliable results: Develop repeatable evaluations, source checks, privacy protections, and safeguards appropriate to each application.
- Efficient AI: Compare models across providers, alongside routing and caching, to balance quality, cost, and response time. OpenAI is our primary focus, with other models also in use and under evaluation.
- Observable systems: Use traces, logs, and metrics to understand behavior, diagnose failures, and improve performance.
These themes span active experiments and future exploration. We will share what we learn as the work develops.
From exploration to working applications
Our work has progressed from conversations and prompt experiments to agents that help develop, test, and deploy working applications. These prototypes let us put ideas into practice, learn from the results, and identify the next questions worth investigating.
LMM Labs Prototype Showcase →
Explore the working applications emerging from this effort.
Articles and explorations
Our earlier explorations provide context for this evolving work, from neural networks and their mathematics to increasingly autonomous organizations.
- ChatGPT – A Closer Look
- Introduction to the Singularity
- Artificial Neural Networks (ANNs): What Are They, Why Are They Important?
- AI and the Human Brain
- The Iterative Organization
- The Math That Makes Neural Nets Possible
- Introduction to the Autonomous Company (AC)
- The Fourth Industrial Revolution (4IR)
- Introduction to Quantum Computing
- Building a Consulting Practice: A Quick-Start Guide
Developed with the assistance of generative AI. We share experiments and ideas as work in progress; testing, source verification, and technical judgment remain part of the process.