Tesla AI Day 2022
Tesla2 HN recommenders · recommended in 2 months, 2022–2026
talk · advanced · 2022–2026 on HN
Tesla's full AI Day 2022 presentation, covering the Tesla Bot hardware and software, Full Self-Driving planning and perception, the occupancy network, data labeling, simulation and Dojo training infrastructure.
HN uses it as primary-source evidence on how Tesla's vision-based driving stack and training pipeline work.
Up next
5 mo on HNLeonard Susskind on The World As Hologram | TVO Docs
TVO Docs
7 HN recommender · 2013–2021
5 mo on HNRichard Feynman on Pseudoscience
Jacob Edward
5 HN recommender · 2016–2026
1 mo on HNMulti-Agent Hide and Seek
OpenAI
1 HN recommend · 2019–2024
8 mo on HNJim Keller: Moore’s Law is Not Dead
UC Berkeley EECS
7 HN recommender · 2019–2025
12 mo on HNHow Self-Driving Cars will Destroy Cities (and what to do about it)
Not Just Bikes
17 HN recommender · 2024–2026
7 mo on HNGoogle Self-Driving Car Project | SXSW Interactive 2016
SXSW
2 HN recommender · 2016–2018
6 mo on HNAndrej Karpathy - AI for Full-Self Driving at Tesla
Matroid
4 HN recommender · 2020–2022
5 mo on HNGeorge Hotz: Comma.ai, OpenPilot, and Autonomous Vehicles | Lex Fridman Podcast #31
Lex Fridman
7 HN recommender · 2019–2026
6 mo on HNTraffic Has a Perfect Solution. Humans Are the Problem.
CGP Grey
6 HN recommender · 2016–2026
4 mo on HNGeorge Hotz: Hacking the Simulation & Learning to Drive with Neural Nets | Lex Fridman Podcast #132
Lex Fridman
6 HN recommender · 2020–2026
3 mo on HNai.bythebay.io: George Hotz, Self-Driving Lessons from Comma AI
FunctionalTV
3 HN recommender · 2018–2021
3 mo on HNAndrej Karpathy: Tesla AI, Self-Driving, Optimus, Aliens, and AGI | Lex Fridman Podcast #333
Lex Fridman
3 HN recommender · 2022–2025
8 comments from Hacker News
> If the data from supervised driving can he used for unsupervised driving So you're unaware of the AI day presentations about their driving training pipeline. Learning about their engineering in that regard really changed my perception of the likelihood that they could solve the problem. Give it a watch: ▶ this video > failure to keep up with competitors like Waymo Looks like each of them lead in different ways to me. Waymo leads within certain specific city limits, Tesla leads on the open road. These seem to be direct consequences of their engineering decisions. Question is which will converge on working everywhere all the time.
It is pretty cool: ▶ this video They transitioned from 2D to 3D a couple years ago, major transition but it does seem like a critical step. We live in a 3d rather than 2d world.
>I've never really thought of Waymo as a robot in the same way as e.g. a Boston Dynamics humanoid, but of course it is a robot of sorts. So for the record, with this realization you're 3+ years behind Tesla. ▶ this video
Exactly. There's no better facsimilie for the sheer variety of public roads than, well, public roads. Tesla even has rigorous simulation in place for hostile situations that should never occur on public roads: A good general overview from two minute papers: link Tesla's AI Day 2022 video, timestamped around labeling and simulation ▶ this video
No, and they should be criticized for that. However, this is the transparency that makes up for it: ▶ this video
As stated by Tesla just a few days ago, the occupancy network is not currently shown in the customer-facing visualisation. (Some unrecognised objects are represented by generic placeholders. This is derived from, but not the output of, the occupancy network.) This video links to the relevant citation, and if you continue watching for around ten seconds you can see examples of the occupancy network visualised: ▶ this video
The Tesla AI day videos [1] go into some detail about this. They use multiple networks that are dedicated to specific tasks. [1] link (2021), ▶ this video (2022)
The replies to your comment don't seem to understand you at all. in the video link here ▶ this video he clearly states 16x dynamic range as a result of direct photon processing.