Hey Everyone,
This won’t be a very serious post, but I enjoyed the Optimus PR that recently dropped on X.
Tesla’s Dojo, Optimus and and AI efforts outside of Tesla (xAI) are worthy of our attention.
Tesla says Optimus can now sort objects autonomously
Its neural network is trained fully end-to-end: video in, controls out.
2 years ago, Tesla had never made a bipedal humanoid robot…
Now they are racing towards the top spot.
The motions are fluid, and the aesthetics is amazing as well.
Are the robots not coming to the AI party?
Adoption of Humanoid General Purpose Robots (HGPRs) might be a lot closer than AGI.
The Demo Video Optimus:
The recently Optimus demo reminds us that humanoid general purpose robots are likely closer than we think.
What if we are nowhere near AGI or functional self-driving cars and full autonomy but robots will become more useful? What sort of world would that be.
Tesla like Nvidia has seen its stock balloon over recent years based on future projections that may or may not be true or factual. Regardless, they can use these funds to reinvest in things like robotics and A.I. and for Elon Musk’s. X Corp.
After being humiliated by OpenAI in 2018, Elon Musk for his legacy needs to remain in the hunt for artificial intelligence supremacy. He will certainly try and he is the world’s richest person.
How Good will Optimus Be?
The following is from Tesla’s AI section of its website:
Tesla Bot
Create a general purpose, bi-pedal, autonomous humanoid robot capable of performing unsafe, repetitive or boring tasks. Achieving that end goal requires building the software stacks that enable balance, navigation, perception and interaction with the physical world. We’re hiring deep learning, computer vision, motion planning, controls, mechanical and general software engineers to solve some of our hardest engineering challenges.
I am optimistic about humanoid general purpose robots in general, but I think there are better pure-play robotics startups here, as compared to Tesla’s recent progress.
FSD Chip
Build AI inference chips to run our Full Self-Driving software, considering every small architectural and micro-architectural improvement while squeezing maximum silicon performance-per-watt. Perform floor-planning, timing and power analyses on the design. Write robust tests and scoreboards to verify functionality and performance. Implement drivers to program and communicate with the chip, focusing on performance optimization and redundancy. Finally, validate the silicon chip and bring it to mass production in our vehicles.
I don’t personally buy into Tesla’s FSD approach vs. the rest of the world here.
Dojo Chip
Build AI training chips to power our Dojo system. Implement bleeding-edge technology from the smallest training nodes to the multi-die training tiles. Design and architect for maximum performance, throughput and bandwidth at every granularity. Dictate physical methodology, floor-planning and other physical aspects of the chip. Develop pre-silicon verification and post-silicon validation methods to ensure functional correctness. Write compilers and drivers to optimize power and performance for our neural networks throughout the entire Dojo system. For more information about Dojo’s arithmetic formats and methods, download our latest whitepaper.
With Optimus, Tesla can get away with over-charging if it can be one of the first to market.
What Does Elon Musk Say?
Tesla with its Supercomputer Dojo, GPU-rich budget and ability to attract A.I. talent, X.AI and Tesla could be considered A.I. native companies.
Tesla Dojo is a supercomputer designed and built by Tesla for computer vision video processing and recognition. It will be used for training Tesla's machine learning models to improve its Full Self-Driving advanced driver-assistance system. According to Tesla, it went into production in July 2023.
There’s also a significant chance that Tesla won’t be a winner at all in robotics and A.I. and won’t manage to compete in those spaces. I don’t personally expect Tesla’s momentum to carry well further into the 2020s and its too reliant on its Shanghai mega-factory.
Elon Musk would certainly like you to think that Tesla is a major A.I. player though. But X.AI should take this mantle as it develops as a well-funded startup in its own right.
X.AI Corp., doing business as xAI, is an American startup company working in the area of artificial intelligence. It was founded on March 9, 2023 in Nevada by Elon Musk. According to the company’s website as of July 2023, "The goal of xAI is to understand the true nature of the universe."
Tesla has not succeeded thus far with FSD in producing cars that are fully autonomous breaking many business development promises of its CEO and public statements of recent years. In Supercomputers I have to imagine Dojo is near in performance with Microsoft or Meta.
I think the verdict is still out if Tesla is as A.I. native as it pretends to be. In July, 2022 Tesla AI and Autopilot leader Andrej Karpathy announced that he was no longer working for the electric vehicle maker. That was a major red flag. He got pulled back to OpenAI it seems. It’s a bit amusing since historically Elon Musk poached a famous Google Scientist for OpenAI.