Figure CEO Reveals Real World AI Plans

by

Figure CEO, Brett Adcock, revealed their plans for real-world humanoid robots. There will be more details in a video coming tomorrow.

Excited to share: Figure has entered into a commercial agreement with BMW 🤝

They hope to ship in 2024.

Figure’s humanoid robots will automate difficult, unsafe, and tedious tasks throughout the manufacturing process. This will enable companies to:

-increase productivity
-reduce costs
-create a safer work environment

BMW indicated a pilot operation for the next 12-24 months.

11 thoughts on “Figure CEO Reveals Real World AI Plans”

  1. But why a humanoid form factor? This feels too influenced by science fiction & pop culture, rather than rooted in solid practical engineering principles. Wheels are much more efficient to travel on than legs, in so many situations. Since when are wheels not general purpose? Even for handling stairs, legs should still be supplemented by wheels, since almost no place is entirely made of stairs.

  2. AI and immortality in 7 years. Supposedly.

    “In 1990, Ray Kurzweil predicted that by 2000, computers would outperform world chess champions and the age of portable computers and smartphones would commence. In 2010, he revisited his twenty-year-old predictions to see how many materialized. According to the data he provided in his article, out of 147 predictions, he believes 115 were wholly realized, 12 were partially fulfilled, and merely 3 were inaccurate (including the prediction anticipating the widespread usage of autonomous vehicles by 2009). Now, Kurzweil declares that we would gain immortality in the upcoming seven years.”

    See:

    https://www.msn.com/en-us/news/technology/seven-years-to-immortality-futurist-kurzweil-promises-a-brave-new-world-with-ai-and-nanobots/ar-BB1h04b1

  4. This will be motivation for the industrial sector to further embrace robotics, but we have a long way to go before robots can replace even 80% of that job base, much less 90 to 100%.

    While I’d love to see advancements to the point that we can all just sit back and relax (or be productive on my own projects, in my case lol), it seems like the semi-distant future– emphasis on distant. Am I wrong, though? Are we closer to that kind of world than it seems [to me]?

    • It is naive to think that after full automation people will sit and relax for a long time. In such a system, people without capital are ballast. The natural way to optimize is to throw the ballast in the trash. Whether it is a series of wars, outright genocide, birth control or expulsion from society is not important. Of course, you can say that your democracies will never do this. OK. In my country (Russia) or in China they will happily do this. And then China and Russia, deprived of social ballast, will be more effective and win in the future. Well, or some other players will win. And yes, that’s good.

  5. Then you simply don’t understand Factories in 2024. Tesla currently has 20,000 employees at the Austin GF and expects that to rise to 60,000 and that’s a brand new state of the art facility as automated as any in the world. .

    Here’s an illustration of a sort of job from Warehousing Distribution. It’s called Unitizing. Parts from suppliers mostly get prepped for presentation on assembly lines but a minority go into the supply of spare parts. Those have to be put together in kits and boxed or packaged. This is Unitizing. There are machines that set up boxes and close them so a human can drop in parts. Currently no automation can cope with the variety of different parts and kits and orientations – it has to be done by humans. The humans move about during the day. They work on different lines and change jobs. Optimus very likely could do this right now. There are tens of thousands of people doing similar things.

    Factories are currently full of jobs that require human perception and dexterity and flexibility. They cannot be handled by contemporary factory automation. Many could be done with Optimus style humanoid robots immediately as drop in replacements with some NN training mostly done by teleoperation by a human until the NN learns the job.

    The magic of humanoid robots is that there is nearly infinite potential demand so the focus can be on true mass production vs industrial machines which are by there nature pretty specialized at least in 2024.

    • Say I want to do unitizing.

      Why use five humanoids?

      Why not mount ten arms and twenty fixed cameras to move things in and out of boxes, then use a wheeled autonomous thing to move the objects from one line to the next. Use AI/ML to train their behavior.

      I don’t see the value of general AGI or legs for factories. Better ML should provide most of the value.

      • >I don’t see the value of general AGI or legs for factories.

        The humanoid robot is NOT just for factories.
        Factory usage is the FIRST use case because it is the most immediately profitable for the robot manufacturer.
        The humanoid robot is designed to replace HUMAN labor.
        These robots will use tools and equipment made for humans, in locations designed for humans, so they NEED to be human-like.

  6. Factory work for humanoid robots doesn’t make sense to me.

    Factory jobs are already highly automated.

    And how many remaining jobs aren’t automated because the job requires legs? How many aren’t automated because they require better controls/training mechanisms?

    If it is more of the latter, then humanoid form factors are mostly a distraction. Applying AI/ML to existing factory robot form factors would solve the right problem.

    • Good point.

      I can see Tesla putting bots in factories because they have the factories themselves and they represent a challenging environment to train bots for. More like using the factories to improve the bots than using the bots to improve the factories. I think companies like BMW want in on the publicity of having humanoids on the floor and giving Figure these “intern” positions is a way to do that.

Comments are closed.