Subvolts take New AI architectures learn continuously, mimicking human memory to avoid forgetting. Jun 3, 2026 - Developers
- Learn about novel architectures for building AI that can adapt and learn continuously, addressing catastrophic forgetting.
- Business
- Understand how advancements in continual learning could lead to more robust and adaptable AI applications for your business.
This interview explores a potentially transformative approach to AI learning that could significantly impact future model development.
Nested Learning: Ali Behrouz on the Quest for Continual Learning & Illusion of AI Architectures
Audio streams from Erik Torenberg, Nathan Labenz. Your place is remembered on this device.
From Subvolts
Summary and analysis by Subvolts editorial, AI-assisted.
This episode of The Cognitive Revolution features Ali Behrouz discussing his "Nested Learning" research for continual AI learning. Behrouz, a Cornell grad student and Google researcher, explains how his work allows AI models to learn new information without forgetting old knowledge, inspired by human memory. The conversation also touches on AI sleep, associative memory, and implications for AGI, privacy, and alignment.
Imagine AI that can learn new things without forgetting what it already knows, like how people learn over time. This research explores new ways to build AI that remembers better, which could lead to smarter and more helpful artificial intelligence.
This episode offers insight into a potential paradigm shift in AI model development, focusing on continual learning. It is for researchers and developers interested in building more adaptable and robust AI systems. Behrouz's work on "Nested Learning" directly addresses a core challenge in scaling AI, a problem all frontier labs like Google Gemini, Anthropic, and Meta AI are actively tackling to achieve more general intelligence.
- HostNathan Labenz
- GuestAli Behrouzgrad student, Cornell
- Guest: Ali Behrouz
- Episode Title: Nested Learning: Ali Behrouz on the Quest for Continual Learning & Illusion of AI Architectures
- Show: The Cognitive Revolution
- Length: 180 minutes
- Key Concept: Nested Learning enables models to adapt to new context while preserving core knowledge by updating different layers at different frequencies.
- Praise: Jeff Dean praised the work as a possible paradigm shift.
Ali Behrouz, grad student at Cornell and Google researcher, discusses his potentially transformative work on new architectures for continual learning in AI. His paper "Nested Learning," praised by Jeff Dean as a possible paradigm shift, enables models to adapt to new context while preserving core knowledge by updating different layers at different frequencies, inspired by human memory systems. The conversation also covers his latest work on AI "sleep" for memory consolidation, why he sees all deep learning as associative memory, and the profound implications of continual learning for privacy, alignment, and the path to AGI.
Mercury: The fintech trusted by ambitious companies and individuals to run their finances, with virtual cards, spending limits, merchant/category locks, and AI-friendly tools like API keys, MCP, and CLI. Check out Mercury at mercury.com
Sponsor:
Claude:
Claude by Anthropic is an AI collaborator that understands your workflow and helps you tackle research, writing, coding, and organization with deep context. Get started with Claude and explore Claude Pro at https://claude.ai/tcr
- continual learning
- Ali Behrouz
- AI learning new things over time without forgetting previous knowledge.
- AGI
- Ali Behrouz
- Artificial General Intelligence, AI with human-like cognitive abilities.
- associative memory
- Ali Behrouz
- A memory system where recalling one piece of information triggers related information.
1 h 27AI:AM: Where's the Steering Wheel? My Agent Set the Password, Hidden Valence, Proofs Find Linux Bugs
WithCameron Berg, Everett Hildenbrandt, Grigore Rosu, Phil Maher
2 h 192 h 021 h 28AI:AM: A Level We Shouldn't Pass? Notes from The Curve + Tokens vs. Salaries & Is SaaS Cooked?
WithThomas Sohmers, swyx, Evan Miyazono, Edward Hu
Same guests, other shows
1 h 55AI 2025 → 2026 Live Show | Part 1
WithZvi Moshowitz, Greg, Eugenia Kuyda, Ali Behrouz
Questions people ask
- What is Nested Learning?
- Nested Learning is a new AI architecture that enables models to adapt to new contexts while preserving core knowledge by updating different layers at different frequencies.
- What inspired this research?
- The architecture is inspired by human memory systems, aiming to allow AI to learn continuously without forgetting.
Page generated Oct 19, 2026. Summaries are Subvolts' own; the audio, artwork and notes belong to Erik Torenberg, Nathan Labenz.