Follow us today...
A couple of weeks ago Tesla held its long-awaited event called "AI Day" (artificial intelligence day). In addition to revealing some details about the long-awaited Tesla Bot - about which we reported some time ago - Elon Musk's company demonstrated the capabilities of its supercomputer, which is internally called Dojo. This demonstration was carried out by presenting a series of digital models of what some electric models based on the Tesla Cybertruck and the Tesla Semi would look like, conquering planet Mars.
Days before the event, Elon Musk himself warned that this would not be an event suitable for all audiences, mainly because it would talk about very technical issues and focused almost exclusively on experts in software, robotics and artificial intelligence. And so it was: the presentation lasted about three hours in which much of the optimal content for "almost all" audiences was shown in the first hour, while the remaining two hours went directly to developers and experts in the hight-tech field.
During the event, Tesla made a clear emphasis on its projection as a technology company by demonstrating the capabilities of its artificial intelligence software systems, as well as its technical resources with respect to – very - advanced software engineering. One of these sections was focused on advances on the Dojo supercomputer, which has hardware entirely designed in-house by Tesla, with absolutely no third-party components. This is something new, since up until now this enormously capable computer used, among other elements, graphics processors manufactured by Nvidia.
With all this artificial intelligence software and hardware equipment the company hopes that one day the supercomputer itself will be able to design its electric cars on its own, without having to resort to the – up until now - successful human team of "Tesla Design Studio"; although for this "it seems that it still has long way to go before matching the team", according to Rajiv Kurian, principal engineer at Tesla, at the public presentation of "AI Day".
Kurian is currently focused on the co-design and integration of hardware and software for micro-architecture. During the presentation, the person in charge confirmed the important steps that this Dojo supercomputer is currently taking, mainly in the artificial intelligence field. One of its main applications will be the so-called Stable Diffusion, which is a text-to-image transfer model that can generate previously detailed digital representations from writing.
The images that illustrate this article about some hypothetical electric vehicles on the surface of the planet Mars have come out of this very process. Some of them are completely impossible to come true, although others could very well inspire the Tesla design team to eventually present a totally new and functional vehicle.
Rajiv Kurian, principal engineer at Tesla who focuses on hardware-software codesign for micro-architecture, explained in a presentation how Tesla is making the best of its Dojo supercomputer AI accelerator through software. https://t.co/s1LJRMPA95 via ...— Ujigami - Zero Defect Manufacturing (@ZeroDefectMfg) October 7, 2022
However, the system still has a long way to go in order to reach the goal proposed by Elon Musk and the Tesla team; although it clearly demonstrates the potential that Dojo already has as of today. In the coming years, with more advances in this field, the company will show new and even more interesting capabilities, as per the announcements made during AI day.
Source: electrek
All images courtesy of Tesla Inc.
Nico Caballero is the VP of Finance of Cogency Power, specializing in solar energy. He also holds a Diploma in Electric Cars from Delft University of Technology in the Netherlands, and enjoys doing research about Tesla and EV batteries. He can be reached at @NicoTorqueNews on Twitter. Nico covers Tesla and electric vehicle latest happenings at Torque News.
Follow us today...
