Tesla AI Day 2 — Minority Report

Sam Lin
4 min readOct 3, 2022

While Tesla AI Day 2022 is to recruit AI & robotics engineers, there are some disappointments because Tesla bot walks barely. It does not even run nor dance as bots from Boston Dynamics, etc. But, why may Tesla make a difference? How can Tesla reinvent Humanoid Robots better?

Moffett Field Bay Trail, Mountain View, California

A genesis, maybe

The president of the Institute of the Future, Roy Amara simplified his insights as Amara’s Law: “We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run”. Because we think in a linear term, but adopt new tech in an S-curve way.

Similarly, it’s too early to estimate where Tesla AI Day 2 will lead the industry & market because of many unknown-unknown. Elon Musk models Tesla as a series of startups from EV, Self-Driving to Humanoid Robots. Even EV is more mature around the tipping point, there is no easy way to rationalize Self-Driving & Humanoid Robots as a regular business.

Teslas’ “startups” on Amara’s Law & S Adoption Curve

So, why does Tesla has a higher chance for these successful ventures? They play 2 cards better at least.

  1. A “wartime CEO”, who has a good track record to innovate & ship highly useful & loved products to many people. Good companies may hit 1 or 2. Whereas, great companies do them all.
  2. A visionary founder, who can tell inspirational visions to attract top & driven talents to build them. This year, they even get many engineering leads to share their stories & progress in detail. A good way to demonstrate “the strong Autopilot team continues the momentum” after the previous lead, Andrej Karpathy left. Sure, they far become a wow factor yet. In reality, only the speed & persistent progress count in the longer run.
Tesla Bot Brain & Battery Pack

To evolve or to revolutionize

In a ‘Star Trek’ Utopia, people are free to pursue their passions instead of working for payday. When the dream comes true, the Humanoid Robot is definitely a part of the equation. Robotic technologies have significant advances in the past 2 decades indeed. For example, Boston Dynamics bots can dance much better than I do. However, when may they be really useful to most consumers?

The same process of collect data & training for Self-Driving & Humanoid Robot.

This Sep., Tesla AI Day 2 shows a few fundamental changes. They may start a new level toward the end game.

  1. Humanlike eyes & brain: Tesla’s Self-Driving bets on cameras as digital eyes & neural networks as silicon brains may be retooled for Humanoid Robots. It’s easy to see the logical sense. The question is how to make that reality. And, how much them can be leveraged across Self-Driving & Humanoid Robot domains.
  2. A general-purpose robot with 200+ degrees of freedom. New capabilities can be added Over-The-Air powered by “Software 2.0” & Dojo super training farms. In a simple term, new features & code can be data driven too, instead of algorithm driven by programmers.
  3. Democratizing Humanoid Robots by a more affordable price: under $20K. Which can make more applications feasible. Tesla is leveraging off-shelf components to find “the fastest path to a useful robot, that can be made at volume” as Elon Musk answered.
Biologically Inspired Design

Full Disclosure

The opinions stated here are my own, not those of my company. They are mostly extrapolations from public information. I don’t have insider knowledge of those companies, nor a whatever expert. Up to you to take it with a grain of salt or two 😉.



Sam Lin

A Taiwanese lives in Silicon Valley since 2014 with my own random opinions to share. And, they are my own, not those of companies I work for.