Breaking News
The Road to War in Ukraine - The History of NATO and US Military Exercises With Ukraine - Part 1&2
Farming = Rebellion Against the Machine | Joel Salatin
DOGE Is Now in Charge of U.S. National Parks
@Benz_Pilled ANIMATION: The Reuters Kerfuffle
Top Tech News
Cramming More Components Onto Integrated Circuits
'Cyborg 1.0': World's First Robocop Debuts With Facial Recognition And 360° Camera Visio
The Immense Complexity of a Brain is Mapped in 3D for the First Time:
SpaceX, Palantir and Anduril Partnership Competing for the US Golden Dome Missile Defense Contracts
US government announces it has achieved ability to 'manipulate space and time' with new tech
Scientists reach pivotal breakthrough in quest for limitless energy:
Kawasaki CORLEO Walks Like a Robot, Rides Like a Bike!
World's Smallest Pacemaker is Made for Newborns, Activated by Light, and Requires No Surgery
Barrel-rotor flying car prototype begins flight testing
Coin-sized nuclear 3V battery with 50-year lifespan enters mass production
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.