Machina Labs

This Thursday, #MachineDesign is hosting a roundtable discussion on leveraging #robotics for gaining control & flexibility in meeting business goals. Featuring experts from Machina Labs, READY Robotics & Vention, this webinar will provide crucial #robot insights. https://bit.ly/4a6KHN9 Thanks to igus Inc. for sponsoring the event!

What if you could buy a car with your own custom designed grille?

When we first explored large-scale 3D printing at Relativity Space, I was influenced by Lockheed's efforts to build satellite tanks using Sciacky’s EBAM 3D printer. The transition from the traditional, multi-stage, and complex process of starting with Titanium forging to agile 3D printing was a game changer. However, when I dove into 3D printing tanks, I realized that while it was impactful, transforming wire or powder into thin-walled formed metal presented significant challenges. I am glad the folks at Relativity are pushing the boundaries there. Coming full circle, at Machina Labs, we have Lockheed as a partner and we are back at it. We switched to sheet metal, a considerably cheaper feedstock. This not only simplified the process but also enhanced the properties of the final product through a formative process. Last year, I discussed the advantages of Roboforming for tanks with SpaceNews. This year, we’re excited to collaborate closely with our partners to advance this technology, with promising developments on the way. https://lnkd.in/g3953Cp8

There are two types of successful startups in defense: Cowboys and Farmers. Cowboys, like Anduril, build new defense systems, disregarding legacy requirements to meet the needs of current conflicts. But our defense strategy still relies heavily on legacy systems. GAO found that only 1out of 49 aircrafts met mission readiness goals consistently over the past 11 years. No matter how many new weapons we build, without reliable legacy systems like the C-130 cargo aircraft, we can't support our new systems in theatre. Thats why we need Farmers: companies that ensure our established systems meet their goals. Machina Labs plays both roles: we help manufacture new defense systems faster while also repairing old aircraft like the C-130 to meet readiness goals.

repurposing and recycling is crucial in #space (e.g., to build #habitats and #infrastructure) ... Machina Labs' #Roboforming helps on that front from three perspective: 1) manufacture structures on earth that are designed with repurposing in mind (apart from usual use-cases, our toroidal tank pictured here can also be used for making the first of its kind, in-space waterslide 😄) 2) repurpose an "earthly" article/part in space (for example, add texture to a metal panel from a spent rocket to make a ramp to a habitat) 3) 1+2

For fellow engineers: This paper demonstrates how transfer learning on graph-based NNs lets us grasp the distinctiveness of each Roboforming machine—think differences in calibration, stiffness, etc.—or variations in sheet feedstock with just a quick model retraining with first few minutes of the process. Looking ahead, as we make larger model with transformer architectures, we won't need to retrain. Instead, we can use the initial layers of forming data as prompts, letting the model predict the rest of the run.

Joking aside we thought of high impact (something like bullets) for forming at some point but then decided to try other things first. Elon is so first principle, he tried bullets first 🤣

The following excerpt from an Ars Technica article on recent developments at #Tesla got me thinking about the need for "hardware updates" for consumer products (e.g., cars and cell phones) in an era of AI/software-defined products. "Customers not turned off by Musk's antics instead are losing interest with a product lineup of two EVs that are ancient in car years (the Models S and X) and two EVs that are merely old (the Models 3 and Y). The Models 3 and Y are also the only two vehicles that Tesla sells in volume. Any other automaker would have a second-generation Model 3 ready to go either this year or next, but at Tesla, the product pipeline is empty." Sticking with the car and #Tesla theme here (#Apple and others, you are not off the hook 😉). Car manufacturers produce new models for different reasons such as, + Market Expectations and Consumer Demand + Technological and Feature Updates + Regulatory Compliance + Economic Strategy + Leasing and Resale Value + Competitive Advantage Keeping your product's hardware changes to a minimum but improving its performance through rapid software upgrades is not only clever but also makes sense money-wi$e (manufacturing hardware is tough and expensive). BUT, we humans are flawed and like it or not, we do get bored too quickly. This is especially true if the looks of something stays the same. (Trust me, I see this everyday. I live in #LA.) So, as a business, is there a way to deal with this dilemma cost effectively and expeditiously? YES, by making your manufacturing and factories (the hard and expensive stuff to change) software-defined. At Machina Labs, we are working on software-defined factories of the future that allow hardware manufacturing to be simplified. The picture you see here is a #Roboformed hood for an electric pickup truck (width of the sheet here is approx. 60" / 1524 mm). No tooling is used for this fab, which makes going from design to part not cumbersome as in traditional routes. Now imagine the future, ... Disclaimer: I'm not advocating for consumerism (I still drive my faithful, 2013 car). Ars Technica article: https://lnkd.in/gfiwVM7Z

Our team will sweep into the customer’s facility with a blend of smiles and expertise, ready to execute the mission of getting the part airborne-approved.