Pinned Tweet
Robotics stole hierarchical control from the brain.
But a significant gap still exists.
Read thread to find out 👉 xitter.cf/Vslowsprinter/status/2…
Robotics stole hierarchical control from the brain. But a significant gap still exists. Read on to see what’s missing 🧵 👇🏽
(Follow me (@Vslowsprinter) for regular AI x bio insights; only images are AI gen; thoughts my own)
This is age old principle that is going to propagate up the entire stack - when constraints start to matter - chips, models, routers, data, agents, tools, apps and everything in between get specialized and the system complexity increases. xitter.cf/Vslowsprinter/status/2…
In the '90s, general-purpose chips wiped out the specialists. The original Thinking Machines, Cray Computer, and a wave of others shut down, and their best ideas got absorbed into other chips.
Zain Asgar of @gimletlabs argues the conditions that made generalists win have changed. Dennard scaling (shrinking transistors made chips faster at the same power) ended around 2005, transistor costs stopped falling around 2012, and memory bandwidth never kept up with compute.
Meanwhile, token demand is projected to grow 24x by 2030 while available power grows about 2x. When power is the bottleneck, performance per watt wins, and that's where specialized chips shine.
Full piece from Zain on why "the time of specialization is upon us": a16z.news/p/why-hardware-spe…
8/9 Google’s recently demoed using reasoning to generate training experience (i.e. policy distillation).
The bigger gap, in my view, is making this loop continuous and online. Can a robot turn effortful deliberation and repeated execution into reliable, reusable skills?
Today’s hard problem should become tomorrow’s easy skill.
Follow me (@Vslowsprinter) for regular AI x bio insights)
🤖 Made with AI
Robotics stole hierarchical control from the brain. But a significant gap still exists. Read on to see what’s missing 🧵 👇🏽
(Follow me (@Vslowsprinter) for regular AI x bio insights; only images are AI gen; thoughts my own)
🤖 Made with AI
6/9 The cerebellum predicts the consequences of actions.
Comparing predictions with sensory feedback generates an error signal that is used to adjust motor system in real time. Fine control doesn’t require replanning the entire task.
5/9 But repeatable skills can’t be completely hardcoded.
Imagine picking up a cup. Its position, weight, orientation and distance may be different from the last time. Mastery means adapting a learned skill to the current situation.
4/9 'Policy distillation' amortizes the costs for frequent tasks in the brain.
When deliberation and refinement lead to mastery via repetition, the action policy becomes a repeatable skill requiring less conscious supervision. Practice makes execution cheaper.
3/9 Therefore, deliberation is effortful and computationally expensive.
It makes economic sense to recruit this analytical machinery only when a novel situation, complexity or high stakes demand it.
2/9 Planning and deliberation require recruiting diverse experts, simulating trajectories and evaluating counterfactuals.
Cortical, memory and basal ganglia brain regions interact throughout. Longer horizon simulations are even more expensive.
1/9 The brain evolved hierarchical planning and control. Why?
Because necessity is the mother of invention. And because the brain is energetically and compute constrained necessitating extreme efficiency.
Sleep turns today’s experience into tomorrow’s skill. Here’s how your brain decides what to replay, reinforce and remember.
To learn how, keep reading 👇
🤖 Made with AI
7/8 Synaptic changes in the brain happen within seconds, but require repeated reactivation and replay during rest and sleep to become long term learnings.
This is how today's experience becomes tomorrow's skills. All while you sleep!
For more such fascinating insights at the intersection of artificial and biological intelligence, follow me on X: xitter.cf/Vslowsprinter
