Tesla chief Elon Musk recently announced his plans to host an ‘artificial intelligence day.’ It would happen in ‘about a month or so’, he tweeted.
There is still no clarity on the exact date as of now. Musk said the main agenda of the event is to showcase Tesla’s hardware and software capabilities, and also to recruit talent.
What to expect?
Earlier, Musk had said, “…it turns out that the only way to solve self-driving is to solve a big part of real-world AI.”
“We believe that an approach based on artificial intelligence for vision and planning, supported by efficient use of inference hardware, is the only way to achieve a general solution to full self-driving,” according to Tesla website.
In an earnings call in January 2021, Musk told investors that he was ‘highly confident the car will be able to drive itself with reliability in excess of humans this year,’ However, in May, Tesla engineers informed the California department of motor vehicles (DMV) that it might not achieve FSD by 2021 end. The DMV is currently investigating if the company violated regulations by falsely promoting its advanced driver-assistance systems as ‘fully self-driving’ or ‘FSD.’
On ‘Tesla AI Day’, Musk might throw light on latest advancements in the areas of FSD and Level 5 (L5) autonomy. ‘Will go over progress with Tesla AI software & hardware, both training & inference,’ said Musk in his Twitter post.
Also, there have been several controversies surrounding Tesla’s decision to remove radar from its cars. Musk has always been bullish on vision-based assistance systems. “LiDAR is a fool’s errand,” Musk said in April 2019, “Anyone relying on lidar is doomed.”
At ‘Tesla AI Day,’ the company might reveal innovations in the areas of computer vision, hardware, neural networks, code foundations, and evaluation infrastructure. At ‘Tesla Battery Day,’ held in September last year, Musk said a $25,000 car that drives itself would be possible in three years.
Last year, Tesla introduced ‘Full self-self driving (FSD) beta’ to vehicle owners as part of its early access programme. The latest update enabled drivers to partially access the automated driver-assist system on streets and local roads. Musk then said the company had handled the software update ‘very cautiously.’
The purpose of offering an early access programme was to iron out software bugs and use it as a testing platform. 824 vehicles were part of the pilot programme as of March 9, as per a DMV memo.
With these developments in the backdrop, Tesla is most likely to release FSD beta rollouts and FSD subscription plans for its customers at ‘Tesla AI Day.’
Tesla’s Gigafactory roadmap
Currently, Tesla operates out of four fully operational facilities — three in the US and one in China.
Last April, Tesla Gigafactory Texas received a second IDRA Giga Press machine. However, it is still unclear how many IDRA Giga Press machines Tesla would install in Texas. Two months ago, Musk said Tesla would need to hire more than 10K people for Giga Texas through 2022.
In Giga Berlin, Tesla is looking to install eight Giga Press machines for production. In March, Tesla got permission from Brandenburg’s State Office for the Environment to install the machines. In Asia, Tesla is reportedly eyeing countries like South Korea, Japan and India to set up Giga Factories.
Recently, Tesla incorporated a new Indian business unit and registered the company as Tesla India Motors and Energy Pvt Ltd with RoC Bengaluru. Last year, Union Minister Nitin Gadkari said Tesla is set to start its operations in the country in 2021 and would look at setting up a manufacturing unit.
Hopefully, at ‘Tesla AI Day,’ the company would reveal its plans about setting up Giga factories and research and development centres in India, as well as other locations.
Tesla is all-in on vision-only autonomous driving. The automobile giant has even started phasing out radar sensors in some of its electric vehicles. At CVPR 2021, Andrej Karparthy explained how it plans to do this by using an in-house supercomputer.
At ‘Tesla AI Day,’ Tesla is likely to highlight the features of its in-house supercomputer, alongside showcasing other hardware capabilities (FSD chip, GPUs, safety and security systems, etc.) in the self-driving paraphernalia.