What is Advanced Machine Intelligence or AMI Labs?

2026-03-11 · Source: AI Supremacy · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Robotics & Autonomous Systems, Emerging Technologies & Innovation · Depth: Advanced, quick

Summary

The article discusses a shift in AI research away from the current LLM-centric path and AGI marketing towards a "Physical AI" era, projected to begin in 2027. This new direction emphasizes Superhuman Adaptable Intelligence (SAI), which leverages self-supervised learning for generic knowledge acquisition from unlabeled data and world models for planning and zero-shot transfer. Prominent AI critic Yann LeCun, a co-founder of AMI Labs, advocates for this alternative, highlighting the need for AI agents to predict action consequences. This second wave of AI startups, including Prometheus Project, Core Automation, and World Labs, focuses on practical applications in robotics, real automation, and physical sciences, moving beyond predictive and generative systems to meet the demands of a future "Machine Intelligence Economy."

Key takeaway

For research scientists evaluating future AI development, consider the emerging "Physical AI" era, which prioritizes Superhuman Adaptable Intelligence over AGI. Your research should explore world models, self-supervised learning, and practical applications in robotics and automation to align with this evolving landscape and address the demands of the Machine Intelligence Economy.

Key insights

A new wave of AI research is shifting from AGI to Superhuman Adaptable Intelligence and Physical AI.

Principles

• AI agents need world models for planning.
• Self-supervised learning acquires generic knowledge.
• Specialized AI startups drive practical innovation.

Method

Superhuman Adaptable Intelligence (SAI) uses self-supervised learning for generic knowledge from unlabeled data and world models for planning and zero-shot transfer.

In practice

• Explore Physical AI and robotics applications.
• Investigate world model-centric AI startups.
• Focus on real automation solutions.

Topics

• Superhuman Adaptable Intelligence
• Physical AI
• World Models
• Robotics
• Self-supervised Learning

Best for: Research Scientist, AI Researcher, AI Scientist, Entrepreneur

Related on AIssential

• A peek inside Physical Intelligence, the startup building Silicon Valley’s buzziest robot brains — Physical Intelligence aims to build general-purpose robotic intelligence through extensive data collection and cross-embodiment learning.
• In Japan, the robot isn’t coming for your job; it’s filling the one nobody wants — Japan's physical AI push is a necessity-driven response to severe labor shortages, aiming for global market leadership.
• What is Physical AI? How Robots Learn & Adapt in Real Life — Physical AI integrates perception, reasoning, and action in the physical world, driven by advanced models and efficient compute.
• After ChatGPT, Liam Fedus Is Building This — Connecting AI to the physical world via robotics and experimentation is essential for scientific and technological advancement.
• Physical AI Matters More Than Humanoid Robots — Physical AI integrates perception, reasoning, and action to enable AI systems to interact effectively with the real world.
• The Spectacular Failure of AI in the Physical World — Current AI models fail in dynamic real-world physical applications due to their inability to adapt to rapidly changing conditions.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AI Supremacy.