So Elmo did a conference where he made bold promises for the future and people are reassured? Like, obviously, this is a guy that keeps his word and has a record of correct predictions....

@AIBear also, totally unrelated to the crashing stock

How far into 2026 do you plan on resolving this if we hear no news?

@SimonWestlake I'll try to resolve as soon as possible, hopefully it'll be a clear yes or no at the end of the year but I could maybe wait until till mid to late January if we were waiting on more news.

https://arstechnica.com/ai/2024/10/reports-teslas-prototype-optimus-robots-were-controlled-by-humans/

@JaimeSantaCruz Wild that people keep betting yes on this

@SimonWestlake what would thousands of humanoid robots do on a factory 🤷🏽‍♂️

For anyone who wants to go a bit bigger and further out with their bet:

https://manifold.markets/DavidFWatson/will-there-be-over-10000-optimus-ro

my priors on this are no, I bought no originally at 15, i remain of that opinion, feels lower priority for tesla too than self driving or language models for xai

I don't think they'll solve the logistics issues (deploy, storage, charging, control, etc) of having that many robots working by then, let alone a manufacturing and assembly up for Optimus. Even if the entire Optimus assembly line is staffed by Optimus robots they'll spend a year doing it.

Why is this not lower? It’s clearly grifter marketing, those robots don’t do anything useful right now

Does 500 showroom display/demonstration models (at 500 showrooms) plus 500 in factory roles count to resolve yes?

If they have models giving a tour of the display like a human worker would or if the tasks they're performing are actually accomplishing something for the factory then sure. If it's just a demonstration of a possible use case I wouldn't really count that as performing productive labor in a factory.

Isn't this incredibly likely? There are 140,000 salaried workers at Tesla as of 2023, only a few thousand of these are engineers, most are in assembly roles within factories.
If you take the simplest 1% of tasks (not jobs, just tasks) then you can introduce robotic labor for ~170-200 workers in each of their 6 factories and hit this target.
If these bots simply move boxes between shelves and perform other minimally complex actions they are likely to hit this number.
There's also a strong incentive to do so as a way of gathering training data for the Optimus platform.

Another data point for context: Amazon already has 750,000+ robots working in their fulfillment centers.

1,000 doesn't seem hard to hit at all; I think this is just a market on whether Optimus will be minimally usable by 2026.

Yeah that's exactly my thinking, the Agility Robotics models are doing minimally useful work today and don't appear dramatically ahead of recent Optimus demos.

The main problem is that basically wherever you can use humanoid robots, you'll improve efficiency and reduce cost by using a different shape. Also the safety aspects of humanoid robots are terrible. Do you want a potentially unstable several hundred kg hard thing moving freely around humans? No? Then you lose the last advantage of human shape. So to me the most likely scenario for a YES resolution is Tesla doing that as a marketing gimmick and/or to use the robots nobody wants to buy.