AutoGrad Changed Everything (Not Transformers) [Dr. Jeff Beck]

ML Street Talk · Beginner ·🧠 Large Language Models ·6mo ago

Key Takeaways

Dr. Jeff Beck discusses the future of AI, arguing that it should be modeled after the human brain, focusing on automatic differentiation and Bayesian inference, rather than relying on larger models like transformers and language models.

Original Description

Dr. Jeff Beck, mathematician turned computational neuroscientist, joins us for a fascinating deep dive into why the future of AI might look less like ChatGPT and more like your own brain. **SPONSOR MESSAGES START** — Prolific - Quality data. From real people. For faster breakthroughs. https://www.prolific.com/?utm_source=mlst — **END** *What if the key to building truly intelligent machines isn't bigger models, but smarter ones?* In this conversation, Jeff makes a compelling case that we've been building AI backwards. While the tech industry races to scale up transformers and language models, Jeff argues we're missing something fundamental: the brain doesn't work like a giant prediction engine. It works like a scientist, constantly testing hypotheses about a world made of *objects* that interact through *forces* — not pixels and tokens. *The Bayesian Brain* — Jeff explains how your brain is essentially running the scientific method on autopilot. When you combine what you see with what you hear, you're doing optimal Bayesian inference without even knowing it. This isn't just philosophy — it's backed by decades of behavioral experiments showing humans are surprisingly efficient at handling uncertainty. *AutoGrad Changed Everything* — Forget transformers for a moment. Jeff argues the real hero of the AI boom was automatic differentiation, which turned AI from a math problem into an engineering problem. But in the process, we lost sight of what actually makes intelligence work. *The Cat in the Warehouse Problem* — Here's where it gets practical. Imagine a warehouse robot that's never seen a cat. Current AI would either crash or make something up. Jeff's approach? Build models that *know what they don't know*, can phone a friend to download new object models on the fly, and keep learning continuously. It's like giving robots the ability to say "wait, what IS that?" instead of confidently being wrong. *Why Language is a Terrible Model for Thought* — In a provocativ
Watch on YouTube ↗ (saves to browser)
Sign in to unlock AI tutor explanation · ⚡30

Dr. Jeff Beck discusses the potential of modeling AI after the human brain, using automatic differentiation and Bayesian inference, to create more intelligent and efficient machines. He argues that current approaches, like transformers and language models, are limited and that a new approach is needed. This conversation provides a thought-provoking perspective on the future of AI and its potential applications.

Key Takeaways
  1. Understand the concept of automatic differentiation and its impact on AI
  2. Learn about Bayesian inference and its role in human cognition
  3. Design a model that incorporates Bayesian inference and automatic differentiation
  4. Develop a system that can handle uncertainty and learn continuously
  5. Apply this approach to real-world problems, like warehouse robots
💡 The brain works like a scientist, constantly testing hypotheses about the world, and AI models should be designed to mimic this process, rather than relying on giant prediction engines.

Related Reads

📰
Building Your First Model Context Protocol (MCP) Server with TypeScript and Zod
Learn to build a Model Context Protocol (MCP) server with TypeScript and Zod to integrate Large Language Models with external environments
Dev.to AI
📰
Beyond Code Generation: How OmniSVG Rethinks Vector Graphics with Vision-Language Models
Learn how OmniSVG uses vision-language models to rethink vector graphics beyond code generation
Dev.to · Shrijith Venkatramana
📰
LangGraph Checkpointing: Three Production Rewrites to Stop Losing State
Learn how to implement LangGraph checkpointing to prevent state loss in production environments
Dev.to · Elena Revicheva
📰
AI builder essentials: tokens, context windows and RAG 101
Learn the essentials of AI building with tokens, context windows, and RAG 101 to improve your LLM interactions
Dev.to · Tilde A. Thurium
Up next
5 Levels of AI Agents - From Simple LLM Calls to Multi-Agent Systems
Dave Ebbelaar (LLM Eng)
Watch →