The Future of Tesla AI and FSD: Arriving in 24 Months
Category Artificial Intelligence Saturday - September 2 2023, 11:28 UTC - 6 months ago Tesla's FSD is powered by OpenAI and uses distributed computing to deliver training with over 10,000 CPUs. This week, Tesla increased its available compute 7 times by enabling 10,000 Nvidia H100 chips, now with a total of over 100,000 A100 chips. Tesla plans to use the FSD for robotaxi level self-driving, factory, and real world settings with Teslabots in use worldwide by 2024.
Saturday - September 2 2023, 11:28 UTC - 6 months ago
Tesla's FSD is powered by OpenAI and uses distributed computing to deliver training with over 10,000 CPUs. This week, Tesla increased its available compute 7 times by enabling 10,000 Nvidia H100 chips, now with a total of over 100,000 A100 chips. Tesla plans to use the FSD for robotaxi level self-driving, factory, and real world settings with Teslabots in use worldwide by 2024.
Video only learning from end to end neural networks in Tesla FSD 12 is like the large language model Chatgpt. Adding more video data and compute will make Tesla FSD vastly better in months. Tesla FSD is also the basis for the real world AI for Teslabot. Tesla FSD, robotaxi level FSD and factory and real world usable Teslabots should all arrive within 24 months.
Tesla's journey to the forefront of Artificial Intelligence and Self-Driving Vehicles has been a long and arduous one. Although AI technology has advanced drastically in recent years, it wasn't until recently that Tesla fully jumped into the AI revolution. Since then, Tesla has rapidly improved upon its FSD technology to lead the industry in self-driving vehicle technology. Tesla FSD is powered by OpenAI, a deep-learning research lab, and uses distributed computing to deliver distributed training with over 10,000 CPUs.
Tesla achieved the world’s first fully autonomous driving AI system this past May, and since then has achieved remarkable milestones in AI and FSD technology. This week, the company declared that it would increase its available compute seven times, powering the FSD with 10 thousand Nvidia H100 chips and exceeding a total of 100 thousand Nvidia A100 chips of AI compute. This makes Tesla the first company to reach this staggering amount of AI compute power.
With the introduction of the world’s first driverless AI system, Tesla’s FSD technology is now being applied to more than just self-driving cars. The same technology which helps keep Tesla’s cars safely operate on the roads is now being applied to all aspects of their business. Tesla now uses AI to handle tasks like preventive maintenance, route planning, and process optimization in the factory. Teslabot also helps provide Tesla with data to further the development of its autonomous driving capabilities.
Tesla provides an example of how automation can assist with the training of AI models, often through the use of massive data centers that are required to house the massive amounts of data and compute required for such models. The company has built its own FSD-centric data centers which are compatible with all types of weather, designed to perform AI inference in sheet units that are resistant to extreme temperatures.
In the coming months, Tesla plans to roll out FSD to its robotaxi fleets, as well as in factory and real world settings. This will make Tesla the first company to achieve robot taxi level FSD. Tesla anticipates that these changes will drastically improve the safety and usefulness of self-driving vehicles. By 2024, Tesla plans to have Teslabots in use all around the world, with improved AI software and full self-driving capabilities.