Skip to content
Sonic
AI
Sonic
AI
Home
Discover
Ask Sonic
Projects
Use with Claude or ChatGPT
Show me around
Request source or feature
Memory and Continual Learning: Engram's Dan Biderman and Jessy Lin, Sonic AI
Home
/
Training Data
/
Memory and Continual Learning: Engram's Dan Biderman and Jessy Lin
Training Data
Notify me
•
Jun 24, 2026
•
44:20
Interview
Memory and Continual Learning: Engram's Dan Biderman and Jessy Lin
From
Training Data
Sean
(Host, Training Data)
•
Shaun Maguire
(Sequoia)
•
Sonya Huang
(Host)
•
Jessy Lin
(Guest)
•
Dan Biderman
(Guest)
Get the full transcript next time Training Data releases an episode
Summary, key quotes, top claims, and the searchable transcript — emailed automatically. No card needed.
Sign up
Executive Summary
Engram is a research lab focused on continual learning, aiming to create AI models that constantly update their knowledge by integrating new information directly into their weights.
Their core technology uses adapter fine-tuning on a per-team or per-user basis, which can dramatically reduce inference costs (by up to 100x) compared to stuffing context into prompts (RAG).
Engram's approach requires white-box access to model weights, making it most suitable for open-source models or partnerships with closed-source providers.
The long-term vision is a future with millions of personalized models that act as an intelligent, associative 'brain state' or interface to a user's or company's entire data plane.
Continue your research
Keep pulling the thread on Dan Biderman and Jessy Lin.
Continual Learning vs. Context Stuffing
The Future of AI is Personalized
12
quotes
Transcript
Key Arguments
Analysis
Quotes & Entities
12
Related
Loading transcript...
Processed Jun 24, 2026
Daily intelligence brief →
yt-dlp + mlx-whisper + Gemini