Breaking News
John Bolton Pleads Guilty To Retaining National Security Information
Can Two Hours Of School Really Be Enough? | MacKenzie Price #491 | The Way I Heard It
Tucker Carlson: "This World Is Not Run By Humans!" Trump Has Supernatural Powers
How They're Building an Off-Grid Community in the Private Realm (Trust + PMA + Ministry)
Top Tech News
'Groundbreaking' Potential Lupus Cure Sends Patients into Remission, Allowing Dreams...
Speculations on What Could Show Physics Beyond the Standard Model
SpaceX Orbital Travel and Orbital Hotels Need Starfall – Getting Back Safe and Cheap is Exciting
Lizard-inspired wiggly wheels let Mars rover swim through sand
Fact Sheet: President Donald J. Trump Ushers in the Next Frontier of Quantum Innovation
Researchers at Johns Hopkins University just let an AI-guided robot remove a dead pig's gallblad
World's first consumer wing-in-ground effect aircraft takes flight
America's Military Readiness Depends On Deployable Nuclear Power
License Plate Cameras Are About To Start Tracking A Lot More Than Just Your Car
Heads up: Apparently the government is hiding cameras inside fake utility boxes
Tesla Dojo, Supercomputers and AI
There does not seem to be a limit for neural nets to utilize more resources to get better and faster results.
Tesla is motivated to develop bigger, faster computers that are precisely suited to their needs.
The Google TPU architecture has not evolved as much over the last 5 years. The Google TPU chip is designed for the problems that Google runs. They are not optimized for training AI.
Tesla has rethought the problem of AI training and designed the Dojo AI supercomputer to optimally solve their problems.
If Tesla commercializes the AI supercomputer that will help to get to lower costs and greater power with more economies of scale.
One of the reasons that TSMC overtook Intel was that TSMC was making most of the ARM chips for cellphones. TSMC having more volume let them learn faster and drive down costs and accelerate technology.
99% of what neural network nodes do are 8 by 8 matrix multiply and 1% that is more like a general computer. Tesla created a superscalar GPU to optimize for this compute load.