Ethan Rublee

🤖 Transform Your Robot with Advanced Object Detection! Are you diving into the world of robotics and grappling with object detection challenges? We've got just the breakthrough you need! Discover how to elevate your robot's perception capabilities using custom ROS messages. Our latest blog post is a treasure trove of insights and step-by-step guidance on publishing your object detection results over ROS, making your robot smarter and more responsive. Why is this a game-changer? 🚀 >> Seamless Integration: Learn the nuts and bolts of using custom ROS messages tailored to your project's needs. >> Enhanced Performance: Elevate your robot's object detection capabilities to new heights with practical, easy-to-implement strategies. >> Community and Support: Join a thriving community of innovators, and leverage the collective wisdom of experts in robotics. >> Dive into the blog post here: Custom ROS Messages for Object Detection Let's push the boundaries of what our robots can do. Share your thoughts and experiences with object detection and ROS in the comments below! 👇 #ROS #Robotics #ObjectDetection #MathWorks #Innovation #TechCommunity #matlab

4

🤖 Can people love #robots if they (appear) to love us? 🤖 I recently toured Miso Robotics #Flippy pop-up store in Pasadena and I can tell you, I love the robots, all of the robots (irrespective of their "feelings" for me)! From back-of-house fry stations to servers, burger makers, robot servers, #autonomousdelivery bots, #drones, #bringthemon! But I'm way bullish, so they don't need to win me over. If they create efficiencies and cost savings for my #restaurant friends, I'm all in. 🤖 This The New York Times article by Meghan McCarron highlights robotics in the #restaurantindustry. "The crux is, if you feel the robot loves you, you allow the robot to cook with love.” Michael Giebelhausen, a professor at the Wilbur O. and Ann Powers College of Business at Clemson University Clemson University 🤖 Mindy Shoss, an organizational psychology professor at the University of Central Florida worked on a study that looked at how automation may affect workers, making them anxious about losing their jobs. The key to doing this is the way the robotics are introduced and how the employees are trained. 🤖 Benson Tsai of Stellar Pizza says part of the magic was hiding the robots from customers. “I’m a firm believer that automation should be felt and not seen. I want them to feel it in their wallet and taste the higher quality of ingredients.” Tsai's automated pizza truck employed one human. “Humans still do one thing very well, and that’s smiling.” 🤖 Bear Robotics Juan Higueros “In our own restaurant, we had four different languages being spoken. People are crowded trying to get everything done, rushing to get orders out and food on time to customers — back-of-the-house is a battleground." Denny's server robots are alive and well. 🤖 Darian Ahler, a food tech consultant, believes that automation “needs to act like a toaster” (as opposed to trying to do stuff humans do well). He helped create the #Autocado machine that Chipotle Mexican Grill uses to peel, pit and half its avocados. “People don’t see this as taking jobs; they see it as assisting.” I'm looking forward to this conversation continuing on stage in October at #MURTEC when Rich Hull CEO of Miso Robotics takes the stage with El Pollo Loco, Inc. CIO Clark Matthews! FULL NYT Article here (worth the read): https://lnkd.in/eSn8hW_U

35

TLDR: If you're building a robot/Raas product you will make more money if you build an optimized, tailored, and scalable control system opposed to building the cheapest solution possible. (better control system = more work done = more work done = more revenue & happy customers) There's really two routes when building a computer/control system for RaaS products (Robots as a service). Autonomous/semi autonomous material handling etc. #1- (Short term solution) Build the control system using a generic off-the-shelf computer. Throw it in a box with the other power, networking, plc, cabling hardware, etc. #2- (Long term solution) Build the control system around a custom chassis that is optimized for your exact application. Design for scalability, flexibility, and reliability. Have a killer supply chain. Here's what I've seen. Even though #2 has a higher purchase price, there is higher ROI in the long run and your customers will be pleased with a purpose built solution. Let's chat and see if your application is something we can help with!

3

InstantDrag introduces a novel drag-based image editing framework that enhances both speed and interactivity by avoiding traditional optimization steps. Key insights: FlowGen: A GAN-based model that generates dense optical flow from sparse user inputs (drag instructions). It decouples motion generation from the actual image editing, allowing for fast processing. FlowDiffusion: A diffusion-based model that leverages the generated optical flow to perform high-quality, motion-conditioned edits. It eliminates the need for text prompts or masks, reducing computational overhead. Efficiency: InstantDrag achieves edits in ~1 second, offering up to 75x faster performance and consuming 5x less GPU memory compared to prior methods like DragGAN. No Inversion Needed: Unlike traditional methods requiring complex latent optimization and DDIM inversion, InstantDrag provides real-time, photo-realistic results by directly conditioning on motion dynamics. This approach is a step forward in interactive, real-time editing, making it suitable for a wide range of applications. Link to the paper : https://lnkd.in/g9FusJU7 #AI #ComputerVision #GAN #DiffusionModels #InteractiveEditing #OpticalFlow #InstantDrag

5

1 Comment

I have mad respect for Nvidia and all who work there. But as the company that more or less rose from the ashes of Silicon Graphics, it isn't potentially making the same core mistake? Its chips get ever more powerful AND power hungry, akin to building massively centralized supercomputers. We recall the thing that killed SGI was missing the move to PCs and effective decentralization and commodification. Today that means "mobile"--as in low-power on-device AI/neural processing built into custom silicon by companies like Apple and lately Microsoft.

25

4 Comments

I’ve previously mentioned there are two questions I’m always asking my teams: “What problem are you solving?” and “Who are you solving it for?” They’re two separate questions, but they should really be answered together. The problem you’re solving can be strongly influenced by who you are solving it for. Are you solving a problem for a homeowner or a business owner? It can make a difference. Take sweeping, for example. A homeowner who needs to clean their house might struggle with brooms that pet hair clings to, lines of dust that don’t make it to the dustpan, dust that eludes sweeping, and floor surfaces that can’t be vacuumed. So, the problem is inefficient sweeping—and to solve it, product developers can work with the homeowner, interview other broom users, and come up with a solution that addresses those issues. Most likely it will be an automated or mechanical broom. But a parking deck owner might struggle with the expense of using contractors to sweep and clean their large spaces. Yet this sweeping can reduce interior contamination that could damage HVAC systems, filtration systems, and more. Gregg Ratanaphanyarat and Andre Ding noticed that outdoor maintenance was a year-round problem for both consumers and commercial property managers and decided to apply robotics to the problem. The result was RUNO sweeping modules that can charge their own batteries, empty their own debris hoppers, and sweep up to seven days a week. #problemsolving #solvingforusers #productdevelopment #productstrategy

13

Did you hear Wind River Linux and NVIDIA Jetson Orin power the new Yaskawa MOTOMAN NEXT? Yaskawa says these systems will help the robot “realize new levels of intelligence and autonomy.” https://ow.ly/QVvS30sBZ89 #AI #WeAreWindRiver

29