An Economic Ode to Optimus

If you don’t have 40 minutes here it is in under 4.

Tesla’s vertical integration gives it a superb competitive advantage over all other humanoid robot makers, Not only does it have the technology to build and train the robots, it has the perfect sandboxes for the Optimus to play in Tesla’s own Mega, Giga, and Trillia Factories where it does not have to deal with pesky customers and nosy news media while reducing its own costs of manufacturing.

For the early investor in humanoid robots Tesla has the advantage of low risk with lots of other already profitable business lines supporting the stock’s price.

The $17 Trillion Opportunity (A Call Option on Infinity! Valuating Teslabot Pt.2: The Road to CHOAM)

The Captain

Are you ready to meet 2024?


Very interesting Denny and thanks for the link. Why do you think they are going after the lower paying jobs? It seems that it would be harder to make a robot that does manufacturing instead of making a robot that is a lawyer or an accountant. I have thought that making a robot paralegal or lawyer in a briefcase would be much simpler because you just have to have something that can follow and understand a rule base system and then spit out the correct response. You know like Deep Blue who won at chess. That would seem to be the low hanging fruit because you could charge more for the robot and then work your way down to the lower paying jobs.



Tesla/Musk reportedly put together a legal team to counter the Lies and Libel/Slander crowd.

With Musk’s penchant for AI assisted tech, why would he not include an AI legal assistant on his team?

ralph lph ph h


Of course I don’t the answer but for a lawyer or an accountant you don’t need the humanoid form. I would also guess that lower paying jobs also require simpler training and much lower risk of getting things wrong. With Americans suing at a drop of a hat that might be important. Finally, Tesla has lower paying jobs in house but not many spots for lawyers or accountants. The synergy just isn’t there.

The Captain


One of the most important events for investing in disruptive technologies is “Crossing the Chasm” because many if not most of them die in the Chasm. Investing pre-Chasm should be left to angel investors and venture capitalists, they know (or should know) the mortality rate. They have to get ten baggers to make up for the corpses.

Where on the Technology Adoption Lifecycle (TALC) are humanoid robots? The linked video should help.

Tesla Bot Expert Compares Top 5 Humanoid Robots

The Captain


Suppose it’s a business that isn’t a business?

The Metaverse.
Google Glasses.
Segue Scooters.
The Tesla Coil.
3D Movies.
Jet packs.

Remember how there were all going to change the world? Not everything that looks whiz-bang cool turns out to be whiz-bang cool.

1 Like

But most of those jobs have already been automated. Auto manufacturing plants are filled with robots already. That’s why so much of the energy in AI is their applications in industries and fields that haven’t previously been automated very much (accounting, programming, law, writing) - and not heavily automated fields like heavy manufacturing.

For Optimus to have a place in a Tesla factory - in any factory - it needs to be able to do a job that hasn’t already been replaced by a machine. Does Tesla have a lot of those jobs? What are the jobs that Optimus could do in a Tesla plant that aren’t today already being done by a robot?


Perhaps this job?

The robot does not have to replace every human in the factory, on the first day.
That can happen later, in stages, as the Teslabots gain functionality.

Remember Elons vision of the factory as “The Machine that Builds the Machines”?

I’ve seen a lotta “yeah but the “?” is too hard for a robot” type comments.
Tesla is re-engineering those.

“Yeah but the wiring harness”.
Tesla has a design that replaced the wiring harness, if I understand correctly, with a much more simple “buss” design.

I suspect the re-engineer specs require “for future robot assembly”.


Would you mind describing the job? I really would prefer not to have to watch a 10 minute video.

As for the above, no one’s saying the robot has to replace every human in the factory. But if this is actually a business sector that’s going to be important, it should be relatively simple to name a few factory jobs that aren’t already being done by robots that Optimus will be able to perform.

Sure :slightly_smiling_face:
That video shows a man assembling the front end components of a Tesla based on “modularity”.

There are 8 or 10 modules that are snapped into place. No nuts, screws, etc.
Everything snaps in from the top, ie no worker contortions required.

If you’ve not seen the video… My description does not really do the concept justice.
Open it, put it on 1.5x speed.
It will still be comprehensible.
Tesla is supposedly building this “simplicity” into every part of its cars.

I watch a fair number of Tesla videos.
Many of the authors say something like “and Tesla engineers are designing the process for eventual robot assembly”.

ralph is happy if someone can poke a hole in the investing thesis!


Okay. So what is the job that can’t be done by a “normal” robot, which has to be done by a human today, that Optimus will be useful to fill?

1 Like

Which jobs is not the heart of it.

The cost savings, the maths were interesting. Possibly $25 Billion in a decade. Maybe much more. That is just in the Tesla plants alone.

The math was deeper into the video.

BTW it is interesting how Musk goes out there and does it. GM and F CEOs are forty years behind in their approach. I think the CEO positions of GM and F should be volunteer jobs.


The topic seems to “shifting”?
I think “replacing a normal robot” is a different question than “replacing a human”.

If I understand “normal robot” tasks correctly, the normal, massive, stationary, floor mounted robot does one (1) task. Ie it picks up one (1) module (big, too heavy for a human*), puts it in place, the car moves to the next station where the next massive, stationary, floor mounted robot picks up its one (1) module, puts it in place, and the line moves forward one station… Repeat.

Is this the way Tesla robot assembly lines currently function?
I don’t see Teslabots replacing the massive, floor mounted robots.

Teslabots are destined to replace human workers. Not “normal robots”.

***(See comment below).

In this video, this assembly is (apparently) currently performed by a human, with the dexterity to snap 8 or 10 small, light weight modules into place. Ie, 8 or 10 “normal, floor mounted, stationary robots tasks”**.

The video suggests it’s a HUMAN job. Not a “normal massive, floor mounted robot job”.

I’ve not seen a video of a small, human size (not form factor, just human size), mobile robot doing anything other than moving tubs of “widgets” around an Amazon fulfillment center, or an automated grocery “fulfillment center”. These all performed one task - move merch from point A to B. In the grocery, other “automated” processes put items in the tubs.
(But see Digit, below).

Really. Watch the video. It’s worth the 10 minutes.


  • think something too big and unwieldy for a human to grab and “wave around, into place”.
    A module with two car seats, or a tire hub assembly, or a car “body” being dipped into a giant paint vat, etc.
    I’ve seen videos of robots doing these tasks.
    I’ve only seen videos. So… Is it believable?

** Musk/Tesla first attempt at an “all robot” factory failed and was quickly scrapped due to the inefficiencies of a bunch of “single task” robots and the extensive floor space required (around 2015-16?). Tesla went to “human assembly” of “human compatible” modules which reduced complexity and required floor space, and improved COGs (increased the profit margin).
Teslabots is, IMO, Musk/Tesla’s next attempt to roboticize these human-size compatible tasks.

This shift from single task robots and large floor area requirements to multi task human-size in a much smaller floor area and the efficiencies gained, is another “aspect” that has been repeated in multiple videos.
Ie, I’m parroting the videographers.

Take my descriptions/understandings with a grain of salt?
I’m not an engineer, and I’ve never been in an automated factory of any sort.

*** Digit, Atlas, Teslabot, the other two, etc in @captainccs linked video are also human worker replacements.

Boston Dynamics Atlas, and the “dog” bots are “Geo fence” constrained, and still largely controlled by a human operator. Reacts to QR codes in the environment.

AI Robotics’ "Figure" robot. Brad Adcock CEO. Might be partnered with Google for compute power.

Fourier. China based. GR-1 general-purpose humanoid robot.

Agility Robotics Digit.
Partnered with Amazon for compute power.
Digit is purported to be “testing” the “pick and place” human tasks in an Amazon fulfillment center. Digit is designed to replace a human worker.

Teslabot designed like the FSD in the car, to react on its own to environmental cues.

Do NOT watch the video Denny linked. And especially do NOT skip to 33:30 and watch the couple minutes.


Albaby if you watch the video’s that Denny and Ralph put up it will give you a better idea of what is going on. Why do you think that so many companies are trying to make a humanoid robot? Some people think it is for some kind of ego move that all of these companies are trying to satisfy. But in reality it is because with a humanoid form they can fit seamlessly into our world. When designing a job for a robot it can make that job unsafe for a human. Sometimes a human may have to do the job or interact on the job. So if they design the job to be safe for a human and make the robot fit that job, then any human can do it. So the jobs they want the robot to do is every job, any job you can think of. I suspect they will have normal robot’s designed for specific jobs but what they are trying to design, and if you watched Denny’s video’s, they think they are close to designing, is a robot that can interact on any job. What I found interesting is that they found out if they gave the robot a will to learn, that the robot didn’t need to be taught as much, it was able to learn the job faster.

All aboard!!! Next stop AI trained robots.


1 Like

That’s not what I was asking, though. I asked what the job was that has to be done by a human today (because the job can’t be done by a “normal” robot) that could be replaced by an Optimus?

There are small “normal” robots. A robot doesn’t have to be a massive device - many robots are compact. I went and watched the video about modular assembly, and there’s no reason that an “ordinary” robot couldn’t do the job of assembling those parts. It doesn’t have to be mobile, and probably doesn’t need to be. If you had a robot do that job, there’s no reason at all it would ever need to move beyond the small patch of the factory where that assembly was taking place. Why give it mobility? Why give it legs? Why not have it hardwired into a power source so it doesn’t need to have batteries and deal with charging, if it’s always going to do that job at that location 24/7? And you certainly don’t need an AGI for that type of basic assembly of rigid components together, either.

That’s exactly the sort of job you don’t need an Optimus for - a new job that could easily be done by an ordinary robot.

So, again, what’ are the jobs that exist in all these Tesla factories that can be performed by Optimus? It shouldn’t be too hard for someone to identify a handful of jobs that are candidates for replacement, if an Optimus with commercially likely technology (ie. no “Data” from TNG) were to be developed.

There’s lots of videos of people waxing enthusiastic about the possibilities of humanoid robots - especially from the companies trying to make them or the folks that are excited to invest in them. But that happens with a lot of technologies that don’t really end up having much commercial potential. I recall the same enthusiasm about Segway, Google glass (and other wearables), 3D television, the Metaverse, and a host of other devices/technologies (like Tesla’s own solar roof) that could have been enormous products, but the capabilities and limitations of the tech never quite lived up to the levels needed.

I watched the videos, and there’s no evidence whatsoever that they’re anywhere close to having a robot that can interact on any job. Munro is basing his opinion just on what he saw the robots do in the Tesla demonstrations - which honestly isn’t all that much of an improvement on what robotics companies have been doing for several years - but said he doesn’t have much insight on the “cognitive stuff.” But that’s always been the limiting factor. We’ve had the ability to build robots that can physically move a lot of stuff around like humans for years and years, but without the programming behind it to actually accomplish any useful tasks

The same reason why so many companies piled into “wearables,” why all the media companies tried to “pivot to video,” why companies tried to make 3D TV, why they all piled into "blockchain, why they all tried to make the “Metaverse” happen. The same reason why everyone and their mum wants to have an AI division right now. No one wants to be left behind if the hyped-up thing actually turns out to be the Next Big Thing, even though many (if not most) Next Big Things end up being relatively inconsequential.

Sure - if you had a humanoid robot that had a brain capable of doing the things that an unskilled or semi-skilled laborer can do, it would be very valuable. Such a brain does not exist, and is not likely to exist any time soon. Which makes the humanoid body purely an aesthetic feature, not a useful one.


What if you had a tent set up outside the factory floor like Musk did. The location could be changing.

True but I could say the same about other technologies that have done great, Cars, Computers, Planes, electric vehicles etc. I was just thinking that robots could be going around my block cutting grass and blowing out yards. But those jobs are low paying so how long would it be, before a robot was taking it over.

But Albaby it isn’t every company, I only see 5 right now. I also look at the money they are throwing at it. It isn’t a trivial amount and after listening to some of these video’s it looks like it is very close.

According to those video’s with experts in the field talking you are incorrect. They think this year we will see them being produced.


I thought the same thing at first.

I do not think robots will be at MCD any time soon. The workers are too low paid. Robots are expensive. Especially humanoid robots are expensive.

The costs are dropping. Pay in factories will be rising. Factory work is rougher than MCD in some conditions. More repetitive. The cost of robotic motions has dropped. Adding the full bones for movement and mobility is easier to do now. (In video game development the actor’s outer body parts are called bones. The movements rotate on them.)

The idea is a more general-purpose robot that learns from the AI network. Later on, as the economics shift the number of robots added to the workforce increases.

Musk is investing wisely for the future. He immediately gets savings.

White Castle–Flippy 2.

Well. That’s the question I answered in that post.

You originally asked for a job that a humanoid robot could do.
Could. I didn’t see "has to be done… I read it as “could be done by a robot” and more economically efficiently.

I didn’t see it asking for a job for which a robot was specifically designed.
That was an example of a human job that I think a humanoid robot COULD do.

NASA is working with an Austin robotics company to build Valkyrie, humanoid robots that can clean solar panels or inspect (damaged, malfunctioning) equipment in space, so that astronauts can focus on other, more intelligent jobs.

Your question, here, seems more “why not just build the perfect robot for THIS specific job?” This robot would NOT be flexible enough for truly “humanoid tasks”.

Did you answer your own question:

POV: a single perfect-for-THIS-ONE-job “dumb” robot vs a multifunctional robot that easily and intelligently “fits” into a world/workplace designed for the human form factor.

A multifunctional robot form factor simplifies the R&D and maintenance, and allows that form factor step into many different “jobs”.
Musk/Tesla is famous for “simplify” and “the best part is no part”, etc.

Why give it “legs”? So that it can fit into a world designed FOR HUMANS? Stairs are an example.
(Yes there are Segway tech chairs, carts, etc that do stairs. But Segways have already been lampooned as failed, here.)

But, really, if the goal is to create a robot that does only that one station, I suspect a “small normal” connected directly to a power source robot could be designed that would fill THAT one station better than a multifunctional robot. No legs. Put it on a track, or wheels, or suspend from an overhead frame.

Others have already mentioned the ability to quickly exchange a non functional (damaged, broken, low battery, etc) robot for a functional one, and keep the process moving.
With a unique robot for each task or “station” it might be more time consuming, and less doable?

Is there an assumption that the job that requires only one human assembling the modules would require only 1 (dumb) robot to accomplish all the tasks? Is this a valid assumption? Or would that station require several different single purpose robots or robotic arms?

I think what I see is a difference in POV.
Musk is a visionary, purportedly thinking 8 3D moves into the future. His goal is to build a robot that is multifunctional, that can perform tasks that are not imagined, today.

I see people complain that Musk makes grandiose claims and does not deliver.

We, as children, are advised to shoot for the stars and if you just hit the moon, well that’s a great achievement, too.
Are these critics really gonna tell “us” to shoot for our shoe laces… Cause our dreams are not achievable?
They are complaining that “Musk dreams too big”!

I also see a man who is in the process of delivering what he’s promised, AND adding new promises.

I see critics demanding the PERFECT “tool, thing, whatever they think Musk promised”.
It MUST BE PERFECT when "released… Or it’s not “what was promised”.

But that suggests a lack of understanding of Musk’s “First Principles” mantra.
First Principles, IF I UNDERSTAND CORRECTLY, says to deliver functional STEPS as each becomes available, so the “customer” can get some use out of parts of their product before the full product is ready (perfect) for delivery.
Musk/Tesla is doing that!

I find it exciting to see each “step” that Tesla reveals.

So, again, this discussion looks like a “POV” difference to me.



But you would not mind rainphakir spending 10 minutes on you? My lawyer used to charge $400 an hour.

The Captain