Tesla AI Day 2022 was a follow-up to Elon Musk’s promises last year to produce a robot named Optimus that could handle dangerous and repetitive tasks. There’s a lot of synergy between the self-driving technology Tesla’s electric vehicle arm has been working on and the challenges this Optimus robot faces. Let’s take a closer look at Tesla’s new bot and the progress of their AI work.
Before you dive into the replay, you should keep in mind that this presentation is primarily geared towards recruiting talent for Tesla. If you’re looking for a series of flashy product announcements, you’re going to be disappointed. That said, if you want to peek behind the curtain at Tesla and get a glimpse of what might one day become available, be sure to check it out.
Optimus Prototype Revealed
The show kicked off with Tesla’s robot Optimus going out for a walk and waving to the audience. The team told the crowd that this was the first time she had been shown untethered.
The object recognition engine it uses to interact with the world is the same one used in Tesla vehicles. A video clip showed a working prototype picking up a box in Tesla’s offices, as well as navigating a busy office full of people and furniture.
A second prototype made exclusively with parts produced by Tesla was also released, although it still had a few weeks to go before it could walk on its own. Elon Musk said Optimus is designed to be a high-volume robot, costing less than $20,000 and available within three to five years. They will be tested for usefulness at Tesla factories before moving to prime time, according to Musk.
Musk was keen to emphasize the need for safety when developing this robot, but was also very optimistic about the economic benefit robots could bring to society.
Much of the time spent on Optimus involved in-depth discussion of the mechanics of individual aspects of the Bot Tesla, with particular attention paid to individual fingers and knees. “Biologically-inspired design” was a phrase that emerged as a way to describe how Tesla engineers put the parts together.
Tesla’s fully autonomous driving progress
Tesla engineers detailed much of the cloud and on-vehicle processing that occurs for their self-driving processes. Much of these neural networks are designed to be able to make the safest decisions possible under tight deadlines and complex landscapes. Tesla pumps 30 petabytes of footage through three supercomputers to build his model training.
Tesla also said the fully autonomous driving (FSD) beta has grown from 2,000 customers in 2021 to 160,000 customers in 2022. Musk again confirmed a global beta for fully autonomous driving by the end of this year, with great improvements along the way. next month, pending testing in various weather conditions.
Tesla also spent a lot of time describing its custom Dojo platform. There is a plot more details than I can go into here, but obviously Musk sees the possibility of Dojo being used for cloud computing for training neural networks in the future.
The end of the presentation was a Q&A session with the audience. Here, Musk reiterated the need to create a useful and scalable robot as soon as possible. As for its evolution, he openly considered giving Optimus a compelling personality or having a complementary ecosystem where other manufacturers could get involved.
Musk also touched on politics, backing a government agency overseeing AI developments and universal basic income in a world where robots were doing much of our manual labor.