When did you start working on Mind?\u00a0<\/strong><\/p>\nSo, Mind is an interesting story and it actually comes back to R2.<\/p>\n
About two years ago, we were building a lot of confidence in R2. So if we\u2019re confident in R2 that means we think there\u2019s gonna be a lot of volume. If we think there\u2019s gonna be a lot of volume, it means we think we\u2019re going to need a lot of plants [factories].\u00a0<\/p>\n
And so effectively, I said, \u2018Boy, if we\u2019re gonna have to build four or five plants over the next decade, that means we\u2019re going to spend many, many billion dollars in capex. What are these plants going to look like? And we don\u2019t want to build a bunch of plants and then have them immediately be outdated.\u2019<\/p>\n
I kicked off an effort to study the future of manufacturing, and essentially came to the view that classic industrial robotics is going to continue to exist like what you see in our plant today, or in a Tesla plant or a Ford plant. But the idea of robotics with human-like skills is going to be really important.\u00a0<\/p>\n
We already had a strong thesis around AI in the physical world. I met with essentially every company that\u2019s working on some form of robotics that can do human-like skills.\u00a0<\/p>\n
I broadly characterize it as there\u2019s companies that fit into the category of existing industrial robotics companies. And I came to the view that they\u2019re going to continue to exist, but not be the types of companies that build robots that can do human-like tasks.\u00a0<\/p>\n
And then there\u2019s a new batch of companies that are all like startups, effectively. Most of which are focused on the home, surprisingly \u2014 which we could spend hours on. A very large number of these companies are focused on folding towels and doing dishes, a subset of those which are looking at industrial.\u00a0<\/p>\n
We just came to the view that the ingredients to be successful, none of them had all of those ingredients. And I felt I could build a better company.<\/p>\n
I\u2019m not going to build Rivian\u2019s future manufacturing dependency on companies that have never industrialized a product or understanding or have really spent time understanding industrial operations, or that don\u2019t have a data flywheel for training models, or that don\u2019t have supply chains because all those companies, we would need to go help them learn how to be a real company.\u00a0<\/p>\n
I made the decision to start a new company, and then had to figure out how the governing structure is going to be like.\u00a0<\/p>\n
And on the day I launched the company, I raised a seed round, and then we just did this Series A\u00a0 financing.<\/p>\n
When you originally founded it or registered, was it called something else?<\/strong><\/p>\nWhat was the project name? I\u2019ll tell you, nobody\u2019s asked me that. (pauses) So it was \u201cProject Synapse.\u201d\u00a0<\/p>\n
Why?<\/strong><\/p>\nA lot of reasons. This is a great question, no one has asked this. So the name of the company Mind, it sort of links to synapse and links to the brain.<\/p>\n
Like firing and misfiring.\u00a0<\/strong><\/p>\nHa, yeah exactly. And it was both because of the brain link. And then at the time, my kids were in a school called Synapse, and I was inspired by my kids and thinking about the business, so I linked them. I actually almost called the company Synapse, but it\u2019s just too funny of a word; I wanted something simpler, but yeah, it was almost synapse.<\/p>\n
It\u2019s funny, one of my board members just wrote me a note and still referred to it as Project Synapse.<\/p>\n
Who is on your board?<\/strong><\/p>\nMyself, Jiten [Behl] of Eclipse, Sameer from Accel, and Rivian has a board representative. It\u2019s a private company, so it\u2019s much easier than a public company.<\/p>\n
You talked about the different categories, is what you\u2019re planning on doing most closely related to what Boston Dynamics is working on, which is the Atlas humanoid, but pairing up with an AI Lab [Google DeepMind]?<\/strong><\/p>\nWe\u2019re building the models as well. We\u2019re building the models, the robotics, and the infrastructure to deploy at scale.<\/p>\n
Then on mechatronics and the robots itself, there has been a surprising emphasis on mimicking human biomechanics, or in some cases, even going further, making even more complex mechatronics.\u00a0<\/p>\n
I think what\u2019s missed in industrial [robotics] and this is one of the things we really see clearly, is the work happens with the hands. So, the hands are very, very important. Everything else, from a robotic system point of view, is to get the hands to the right place. And so the ability for the robots to do really complex motions, like, let\u2019s say, like a back flip that\u2019s actually just means the robot has a lot of unnecessary complexity in it for the vast majority of tasks.<\/p>\n
And I understand the purpose of showing that is to show the flexibility and capability (of the humanoid robots). But if you were to go into a Rivian facility, you will see very few people that have the type of flexibility that would enable them to do a back flip. And so it\u2019s just when you think about deploying at scale, you want to minimize the complexity, minimize the number of failure modes, reduce power consumption.\u00a0<\/p>\n
And so I do think many of the robotic systems, particularly humanoid systems, that are being thought about for manufacturing are way too complex relative to what they do. They\u2019re going to be human like, they\u2019ll have hands, there\u2019s a perception model at the top. There\u2019s going to be the ability to adjust and X, Y and Z; there\u2019s going to be locomotion to allow it to move. But I think, mimicking human biomechanics in a manufacturing environment misses some of the fundamental points of manufacturing, which is, it\u2019s all hand based.<\/p>\n
I\u2019m trying to visualize what this will eventually look like or what the different variants will be.<\/strong><\/p>\nSo, the hands are the most complicated part of robots, though.\u00a0\u00a0<\/strong><\/p>\n(Quickly jumping in) But it\u2019s also where you focus your energy, your dollars, the spending, everything should go into the hands.\u00a0<\/p>\n
And I think the other thing to notice, if you\u2019re building a business that is going to service a lot of different industrial sectors for manufacturing, there\u2019s not one set of hands that\u2019s going to be perfect.\u00a0<\/p>\n
If you think about the challenge of hands is torque amplification. If you\u2019re picking up a large four-inch diameter steel pipe, it\u2019s a different gripping solution than threading an M4 fastener into an aluminum casting. It\u2019s just very different skills and very different levels of dexterity.<\/p>\n
One of the other things that\u2019s happening is in robotics, which is very surprising, is in the biological world we so clearly recognize that evolution has allowed us to create, like very different optimal sets of biomechanics for different things. The best thing for swimming looks very different than the best thing for running and the best thing for climbing looks different than the best thing for lifting. And because of that, humans are not particularly good at swimming. Like, if you were to put me up against a dolphin, I would lose. If you put me up against a cheetah, I would lose.\u00a0<\/p>\n
We\u2019re more powerful, of course, in our brain. And so in manufacturing, to assume that the shape and form factor of a human is the optimal shape and form factor for all types of work, I think, just misses the fact that we didn\u2019t evolve in a plant. Now, plants have evolved around us, which is an important point. And there\u2019s a huge embedded brown field infrastructure that robots have to plug into. But the part that they have to plug into is they have to roughly fit in the space that humans allow for, and they have to have hands. I call that all out because we have a different perspective then I think a lot of the robots that often started from saying we\u2019re going to work in the home, which is a very different ODD [operational design domain].<\/p>\n
But like all these use cases that say we\u2019re gonna do everything a human does. it\u2019s a different ODD and therefore it\u2019s gonna lead to a different form factor than if you said, I\u2019m just gonna be focused on being a manufacturer.\u00a0<\/p>\n
And actually [in an industrial factory] you don\u2019t have to deal with stairs, you don\u2019t have carpet to tile transitions, you don\u2019t have to worry about tripping on a cat. You don\u2019t have to worry about stepping on a kid. So you just have much more friendly constraints for deployment at scale. And you can map the environment, and the environment doesn\u2019t change, right?\u00a0<\/p>\n
The one thing that would change the dynamic would be any interaction with humans [in a factory].<\/strong><\/p>\nWhich is an important point.\u00a0<\/p>\n
That\u2019s why the UI [user interface] is very important. We haven\u2019t shown what it\u2019s gonna look like, but I\u2019ll give you the brief [sic]: It needs to feel friendly, because it\u2019s gonna work a long time with humans. But it doesn\u2019t want to feel dopey and end up with a kick me sticker on its back, right?\u00a0<\/p>\n
A lot of robotics has also gone towards these very lean, muscular forms that look very Terminator like. I don\u2019t know why. I guess science fiction movies or something has caused the industrial design world to do that. There\u2019s a version of robots and they all sort of blend together. They\u2019re these very athletic sprinter body forms. I think there\u2019s a different form factor that can feel much more approachable, but not dopey.<\/p>\n
I\u2019m starting to see a thread with what you did with Also, and your thinking, in terms of modularity with Mind.\u00a0<\/strong><\/p>\nSome of that, yeah.\u00a0<\/p>\n
The other thing is Rivian. If you were to ask somebody in 2018 what a 1,000 horsepower truck looks like, and you squinted, you wouldn\u2019t think of an R1 [truck]. And so, I think the idea is you can have all this capability, but it doesn\u2019t have to look intimidating or scary. It can be friendly \u2014 it\u2019s a design belief and aesthetic that will carry into robotics as well.<\/p>\n<\/div>\n
\n
Source link <\/a><\/p>\n","protected":false},"excerpt":{"rendered":"If you haven\u2019t heard, Rivian f… <\/p>\n","protected":false},"author":1,"featured_media":27877,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":[],"categories":[249],"tags":[2091,2160,2829,952],"_links":{"self":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts\/27876"}],"collection":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=27876"}],"version-history":[{"count":0,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/posts\/27876\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=\/wp\/v2\/media\/27877"}],"wp:attachment":[{"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=27876"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=27876"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/microvibenews.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=27876"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}