AI-powered
podcast player
Listen to all your favourite podcasts with AI-powered features
Comma AI's mission is to solve self-driving cars while delivering shippable intermediaries. They aim to create fully autonomous vehicles that can drive with the same level of safety, speed, quality, and comfort as human drivers. The Comma 2, their flagship product, is a piece of hardware equipped with cameras and sensors. It functions as a person replacement and offers features like adaptive cruise control and lane-keeping. Comma AI emphasizes the importance of revenue and capitalism, striving to provide value and consent-based transactions. They are bullish on the long-term prospects of cryptocurrencies, recognizing the potential of blockchain and smart contracts in redefining systems and providing certainty. George Hotz, the founder of Comma AI, has an aversion to deliberate intentional approaches to life and finds it difficult to distill his own learning process into actionable advice.
The Comma 2 is a hardware device that functions as a person replacement in vehicles. Equipped with cameras and sensors, it offers features like adaptive cruise control and lane-keeping, aiming to make driving a more relaxed experience. It utilizes machine learning models and neural networks to interpret input data, execute commands, and manage the vehicle's performance. Comma AI focuses on delivering shippable intermediaries to keep the company accountable and ensure real progress in solving self-driving cars. The goal is to make driving a seamless and safe experience while generating revenue along the way.
OpenPilot is a software component developed by Comma AI that connects various sensors and manages the execution of commands in the Comma 2 hardware. It encompasses the neural network models, data logging, data fusion, and the implementation of driving policies. OpenPilot bridges the gap between the neural network's predictions and the actual execution of driving functions, including adjustments for banked roads and the management of device resources. Its focus is to continuously improve the driving capabilities and ensure a smooth and reliable autonomous driving experience.
Comma AI leverages large amounts of high-quality user data to train their machine learning models. They adopt an end-to-end learning approach, training on actual driving behavior rather than relying on explicit lane detection or object detection tasks. While there is an abundance of data, edge cases pose challenges in selecting the most relevant training examples. Finding the cleverest way to select data is an ongoing question in the field, with the recognition that driving is full of unique situations and scenarios. Comma AI actively works on improving data selection to enhance the performance of their autonomous driving system.
The speaker emphasizes the importance of selecting data well when creating intelligent systems. In reinforcement learning, the data depends on the weights, which is different from supervised learning where the weights depend on the data. The goal is to eventually have the system learn and pick its own data. This approach is seen as a form of reinforcement learning framework, where the system constantly trains and improves based on user disengagements and mistakes.
The speaker discusses the possibility of exponential improvement in self-driving technology. While exponential improvement does not mean radical changes in hardware or processing power overnight, it does suggest that progress in self-driving cars will continue to advance gradually. The gap between current technology and human driving abilities is still significant, but with continued development, several 10x improvements are expected before achieving fully autonomous level five capabilities.
The podcast explores the different approaches taken by comma AI and Tesla in developing self-driving cars. Comma AI focuses on machine learning and a more holistic end-to-end approach, while Tesla adopts a multitask learning approach, breaking down driving tasks into smaller, more specialized tasks. Both approaches have their advantages, with comma AI prioritizing simplicity and the use of human driving behavior data, while Tesla focuses on mastering specific tasks and using common expertise in driving.
Learning programming languages like Assembly, C, Python, Haskell, and Verilog can provide a comprehensive understanding of different programming paradigms and their applications.
Love is seen as multifaceted, with different forms such as opiates or oxytocin, and a broader abstract love for humanity. Hotz prioritizes ethical, productive love and emphasizes that love is about self-discovery.
Hotz believes that all drugs have lessons to offer, and highlights the value of exploration and experiences. Hotz also asserts that insights gained from psychedelics are personal and self-reflective rather than providing universal truths.
Hotz's advice is to focus on real-world problem-solving rather than seeking advice. He emphasizes the importance of experience and encourages individuals to go out and program, tackle projects, and learn from their own efforts.
George Hotz (geohot) is a programmer, hacker, and the founder of Comma.ai. Please support this podcast by checking out our sponsors:
– Four Sigmatic: https://foursigmatic.com/lex and use code LexPod to get up to 40% & free shipping
– Decoding Digital: https://appdirect.com/decoding-digital
– ExpressVPN: https://expressvpn.com/lexpod and use code LexPod to get 3 months free
EPISODE LINKS:
Comma.ai’s Twitter: https://twitter.com/comma_ai
Comma.ai’s Website: https://comma.ai/
George’s Instagram: https://www.instagram.com/georgehotz
George’s Twitch: https://www.twitch.tv/georgehotz
George’s Twitter: https://twitter.com/realgeorgehotz
Comma.ai YouTube (unofficial): https://www.youtube.com/channel/UCwgKmJM4ZJQRJ-U5NjvR2dg
PODCAST INFO:
Podcast website: https://lexfridman.com/podcast
Apple Podcasts: https://apple.co/2lwqZIr
Spotify: https://spoti.fi/2nEwCF8
RSS: https://lexfridman.com/feed/podcast/
YouTube Full Episodes: https://youtube.com/lexfridman
YouTube Clips: https://youtube.com/lexclips
SUPPORT & CONNECT:
– Check out the sponsors above, it’s the best way to support this podcast
– Support on Patreon: https://www.patreon.com/lexfridman
– Twitter: https://twitter.com/lexfridman
– Instagram: https://www.instagram.com/lexfridman
– LinkedIn: https://www.linkedin.com/in/lexfridman
– Facebook: https://www.facebook.com/LexFridmanPage
– Medium: https://medium.com/@lexfridman
OUTLINE:
Here’s the timestamps for the episode. On some podcast players you should be able to click the timestamp to jump to that time.
00:00 – Introduction
07:02 – Will human civilization destroy itself?
09:49 – Where are the aliens?
14:36 – Tic Tac UFO and Bob Lazar
17:04 – Conspiracy theories
19:07 – The programming language of life
23:28 – The games that humans play
31:58 – Memory leaks in the simulation
34:29 – Theories of everything
36:14 – Ethereum startup story
44:02 – Cryptocurrency
53:28 – Self-help advice
57:08 – Comma.ai
59:02 – Comma two
1:07:50 – Tesla vs Comma.ai
1:16:53 – Driver monitoring
1:30:34 – Communicating uncertainty
1:32:22 – Tesla Dojo
1:38:50 – Tesla Autopilot big rewrite
1:45:09 – How to install the Comma Two
1:49:44 – Openpilot is Android & Autopilot is iOS
1:58:59 – Waymo
2:10:12 – Autonomous driving and society
2:12:24 – Moving
2:15:29 – Advice to Startups
2:28:32 – Programming setup
2:31:32 – Ideas that changed my life
2:39:37 – GPT-3
2:42:57 – AGI
2:47:00 – Programming languages that everyone should learn
2:53:33 – How to learn anything
2:56:05 – Book recommendations
3:04:28 – Love
3:06:17 – Psychedelics
3:08:38 – Crazy
Listen to all your favourite podcasts with AI-powered features
Listen to the best highlights from the podcasts you love and dive into the full episode
Hear something you like? Tap your headphones to save it with AI-generated key takeaways
Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more
Listen to all your favourite podcasts with AI-powered features
Listen to the best highlights from the podcasts you love and dive into the full episode