Tesla gave an replace on its humanoid robotic program, which is named Tesla Bot or Optimus. New photos of prototypes have been fairly spectacular – making the challenge look much less like a sideshow and more and more like a doubtlessly actual product.

When Elon Musk first introduced the Tesla Bot, many laughed it off as a sideshow or distraction to Tesla’s extra essential mission to speed up the appearance of sustainable power.

The CEO hyped it up by describing how a lot worth it could create by fixing the labor disaster, however like Tesla’s autonomous driving effort, everybody can see the worth of humanoid robots – the issue is individuals have points seeing Tesla making it a actuality.

It didn’t assist that the most recent demo at Tesla AI Day final 12 months was lower than spectacular.

On the time, Tesla had a really early prototype that didn’t appear to be a lot. It was barely capable of stroll round and wave on the crowd. That was about it.

The corporate additionally had one other extra refined-looking prototype, but it surely wasn’t even capable of stroll in time for the presentation.

Right here’s what it appeared like:

Tesla claimed to have an important alternative to develop this humanoid robotic as a result of it may leverage loads of current {hardware} developed for its electrical automobiles and software program from its self-driving know-how.

However it wasn’t clear how a lot effort was put into the challenge although Musk claimed it grew to become a high precedence at Tesla early final 12 months.

Now at Tesla’s 2023 shareholders assembly in the present day, Musk gave an replace on Tesla Bot that included loads of new footage of a number of prototypes:

The footage included 5 Tesla Optimus prototypes, and so they have been seen performing easy duties, strolling across the workplace in addition to in different Tesla services the place Cybertrucks have been round.

The prototypes have been waking slowly, however they seemed to be secure.

Whereas the duties that they have been performing have been not likely spectacular, Tesla appeared to have made loads of progress in creating the fingers:

Tesla additionally gave a glimpse on the robots detecting and memorizing their setting.

Musk once more claimed that the “Optimus stuff is extraordinarily underrated”. The CEO mentioned that the demand could possibly be as excessive as 10 to twenty billion items.

He went so far as “confidently predicting” that Optimus will account for “a majority of Tesla’s long-term worth”.

Electrke’s Take

I’m nonetheless skeptical about this challenge, however I’ve to present credit score the place credit score is due. This seems to be like a major enchancment over the past demo, which was nearly 8 months in the past.

The fingers, that are arguably essentially the most troublesome a part of a humanoid robotic, are actually spectacular right here.

Now I believe we’re nonetheless at the least about 3 years from a helpful product, however that in itself could be superb.

Thoughts you, that timeline can also be once I assume that Tesla may have a helpful self-driving car, which is sensible since Elon says that Tesla is leveraging its AI improvement for self-driving for Optimus.

Whereas we will argue on the timeline, I wouldn’t wager towards Tesla on this. On the {hardware} aspect of issues, they do have an important benefit in leveraging their present EV {hardware}.

When you concentrate on it, there’s no main engineering downside that must be solved with a purpose to create a humanoid robotic. It simply must be effectively packaged in addition to designed and manufactured in a manner that the robotic doesn’t price greater than $100,000.

Folks didn’t assume that was attainable with electrical automobiles and Tesla did it. I believe they will do it for humanoid robots.

The AI aspect is the tougher problem. That’s why I’m speaking about ~3 years for a helpful product. That’s accounting for the truth that Tesla has already made loads of errors in creating AI for self-driving. These are errors that it received’t make for the robotic and at one level, it’s going to run out of errors to make.

Additionally, a helpful product doesn’t imply a robotic that may change a major proportion of human laborers. It signifies that on the price, it can change some staff. It should probably be a decade earlier than the capability can lengthen to a lot of duties and the price of acquisition and operation make it a priceless choice for deployment at scale.

Much like what Tesla did with electrical automobiles timeline-wise.