From Neural Networks to Digital Brains: The Next Leap in AI
summary
inait builds biologically accurate digital brains — full neuron and synapse simulations validated against wet‑lab results — and claims 300 papers validating the simulation. these digital brains use a local, biology‑inspired learning rule instead of backprop, enabling 30–50× faster learning on interactive tasks (cart‑pole and Pong learn in minutes and games instead of hours). forecasting is already productized with a Microsoft partnership and an Excel upload demo, small networks of 500–1,000,000 neurons run on a laptop for real‑time robotics experiments.
AI-written summary · about
chapters (14)
- 0:00Intro
- 0:22How do we bridge the gap in robotics to a natural physical movement?
- 2:29Evolution
- 4:41Digital brians
- 5:56What does a neuron look like?
- 6:37How does a neuron compute?
- 9:34What does it look like in a network?
- 9:59AI with digital brains
- 14:29Biological learning
- 15:49Gen I: Forecasting
- 17:42Gen II: Acting (demo)
- 20:48The future
- 21:40Takeaways
- 23:46Outro
related talks
32mHow Neuroscience Makes AI Intelligent
50mAI deciphers the brain code: from real synapse to a neuron's digital twin
29mDeep Learning, Neuroscience and the future of AI
21mOrganoid Computing – Can Nature Save AI?
36mBrain's Neural Network as a service
41mArtificial Intelligence: Let's Think About Learning
41mBrain-inspired computational primitives for reliable neuromorphic intelligence
01h 16mRethinking Computing with Neuro-Inspired Learning: Algorithms to Hardware