Brad Porter’s Post

THE THREE LAWS OF ROBOTICS -- Rodney Brooks is the Panasonic Professor of Robotics (emeritus) at MIT, where he was director of the AI Lab and then CSAIL. He has been cofounder of iRobot, Rethink Robotics, and Robust AI, where he is currently CTO. In honor of Isaac Asimov and Arthur C. Clarke, Rodney's two boyhood go-to science fiction writers, he's names the central pillars of his learning his 'Three Laws of Robotics': 1: The visual appearance of a robot makes a promise about what it can do and how smart it is. It needs to deliver or slightly overdeliver on that promise or it will not be accepted. 2: When robots and people coexist in the same spaces, the robots must not take away from people’s agency, particularly when the robots are failing, as inevitably they will at times. 3: Technologies for robots need 10+ years of steady improvement beyond lab demos of the target tasks to mature to low cost and to have their limitations characterized well enough that they can deliver 99.9 percent of the time. Every 10 more years gets another 9 in reliability. (Note: that these laws are written from the point of view of making robots work in the real world, where people pay for them, and where people want return on their investment. Which is very different from demonstrating robots or robot technologies in the laboratory.) 🚀

"AI Will Transform Manufacturing Robotics—Eventually" is my latest Forbes article arguing that the great advances in LLMs have not helped robotics much. But some use cases will benefit in the coming years. #AI #manufacturing https://lnkd.in/gi8eJXua

Revisiting my roots with Forbes… Before founding Plus One Robotics, I spent over two decades with Yaskawa Motoman identifying emerging tech for robotic automation applications. At the time, robotics capabilities were comparably limited. Manufacturing robots couldn’t see, feel or “perceive” their environment like they can today. Meaning they couldn’t capitalize on AI models, even if they’d been architected to do so — which they weren’t. Then, the 2007-2008 financial crisis took place, and like any smart company would do during an economic disruption, Yaskawa assessed alternate end markets for our products. We evaluated aerospace, clinical lab automation, and a whole host of other verticals; cut to the chase, the whole effort led to logistics and warehouse automation. The only missing piece for warehouse robots remained “perception” — aka 3D vision technology enabling robots to gain situational awareness of their environment and make contextualized decisions about picking, placing, etc. There was an opportunity to improve robotic automation outcomes for manufacturing AND the supply chain this way. But it was clear that the logistics industry had the more pressing need for 3D vision at the time. Warehouse robots couldn’t operate solely on repetition because the products to be handled were so varied and changing weekly. Fast forward to today. Plus One robots are performing these tasks daily within the four walls of distribution and logistics because they’re purpose-built for the work. The robot arms themselves are the same as those used in the manufacturing industry, but the software, the computer vision and the AI models are entirely new. Moreover, supply chain demand is STILL spiking, making efficiency on the warehouse floor more critical than ever. But I'm keen to watch as AI tools become increasingly applicable for manufacturing processes. Recent developments have led some experts to suggest that generalized models may now be relevant. I don't think they will, BUT general-purpose functionalities will absolutely bring value to traditional industrial robots. (In fact, that’s what GrayMatter Robotics is doing so successfully.) Does that mean we’re approaching generalized AI models akin to LLMs for industrial robots? Well, if you’ve been following me for a while, you probably know my answer. Not so fast. To return to the article: “Robotics will continue to grow in areas with a large number of repetitive operations. The cost of the robots will decline while the cost of human labor will continue to rise.” ^ This is the future many AI generalists are reaching toward. We’re cheering them on here at Plus One. But we’ll stick with what pays dividends on day one: adaptable AI model robots empowered by humans in the loop and 3D vision software. Thanks, Bill Conerly, for our stimulating discussion. I really enjoyed the call.

Rodney Brooks, the Panasonic Professor of Robotics Emeritus at MIT and co-founder of Rethink Robotics, iRobot, and Robust.ai, brings a critical perspective on generative AI. 💬 “𝐈’𝐦 𝐧𝐨𝐭 𝐬𝐚𝐲𝐢𝐧𝐠 𝐋𝐋𝐌𝐬 𝐚𝐫𝐞 𝐧𝐨𝐭 𝐢𝐦𝐩𝐨𝐫𝐭𝐚𝐧𝐭, 𝐛𝐮𝐭 𝐰𝐞 𝐡𝐚𝐯𝐞 𝐭𝐨 𝐛𝐞 𝐜𝐚𝐫𝐞𝐟𝐮𝐥 [𝐰𝐢𝐭𝐡] 𝐡𝐨𝐰 𝐰𝐞 𝐞𝐯𝐚𝐥𝐮𝐚𝐭𝐞 𝐭𝐡𝐞𝐦.” He cautions against overestimating its capabilities, emphasizing that while impressive, generative AI has limitations compared to human intelligence. In his view, applying language models to warehouse robotics, for example, could hinder efficiency rather than enhance it. 💬 “𝐖𝐡𝐞𝐧 𝐚 𝐡𝐮𝐦𝐚𝐧 𝐬𝐞𝐞𝐬 𝐚𝐧 𝐀𝐈 𝐬𝐲𝐬𝐭𝐞𝐦 𝐩𝐞𝐫𝐟𝐨𝐫𝐦 𝐚 𝐭𝐚𝐬𝐤, 𝐭𝐡𝐞𝐲 𝐢𝐦𝐦𝐞𝐝𝐢𝐚𝐭𝐞𝐥𝐲 𝐠𝐞𝐧𝐞𝐫𝐚𝐥𝐢𝐳𝐞 𝐢𝐭 𝐭𝐨 𝐭𝐡𝐢𝐧𝐠𝐬 𝐭𝐡𝐚𝐭 𝐚𝐫𝐞 𝐬𝐢𝐦𝐢𝐥𝐚𝐫 𝐚𝐧𝐝 𝐦𝐚𝐤𝐞 𝐚𝐧 𝐞𝐬𝐭𝐢𝐦𝐚𝐭𝐞 𝐨𝐟 𝐭𝐡𝐞 𝐜𝐨𝐦𝐩𝐞𝐭𝐞𝐧𝐜𝐞 𝐨𝐟 𝐭𝐡𝐞 𝐀𝐈 𝐬𝐲𝐬𝐭𝐞𝐦; 𝐧𝐨𝐭 𝐣𝐮𝐬𝐭 𝐭𝐡𝐞 𝐩𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 𝐨𝐧 𝐭𝐡𝐚𝐭, 𝐛𝐮𝐭 𝐭𝐡𝐞 𝐜𝐨𝐦𝐩𝐞𝐭𝐞𝐧𝐜𝐞 𝐚𝐫𝐨𝐮𝐧𝐝 𝐭𝐡𝐚𝐭 𝐚𝐧𝐝 𝐭𝐡𝐞𝐲’𝐫𝐞 𝐮𝐬𝐮𝐚𝐥𝐥𝐲 𝐯𝐞𝐫𝐲 𝐨𝐯𝐞𝐫-𝐨𝐩𝐭𝐢𝐦𝐢𝐬𝐭𝐢𝐜, 𝐚𝐧𝐝 𝐭𝐡𝐚𝐭’𝐬 𝐛𝐞𝐜𝐚𝐮𝐬𝐞 𝐭𝐡𝐞𝐲 𝐮𝐬𝐞 𝐚 𝐦𝐨𝐝𝐞𝐥 𝐨𝐟 𝐚 𝐩𝐞𝐫𝐬𝐨𝐧’𝐬 𝐩𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞 𝐨𝐧 𝐚 𝐭𝐚𝐬𝐤.” Brooks advocates for practical, purpose-built automation in controlled environments like warehouses, highlighting the importance of human-robot collaboration and scalability. His insights challenge the notion of exponential technological growth, urging a balanced approach grounded in real-world application and strategic development. Read more: https://lnkd.in/e-BQy6AQ

While AI is now rapidly transforming digital workflows, the next frontier is transforming physical workflows such as factory floors through scale intelligent robotics. While better data integration, AI/ML integration, and cost of robotics (such as the recent moves to electric motors for movement) etc will make the technology widely applicable, the immediate use of AI to improve current robotics is through simulating environments and generating synthetic data for better training. Sharing a good article below. #robotics #ai

Rodney Brooks, co-founder of Rethink Robotics, iRobot and Robust.ai and the Panasonic Professor of Robotics Emeritus at MIT says many “humans tend to overestimate [generative AI]’s capabilities.” “When a human sees an #AI system perform a task, they immediately generalize it to things that are similar and make an estimate of the competence of the AI system; not just the performance on that, but the competence around that. And they’re usually very over-optimistic, and that’s because they use a model of a person’s performance on a task.” He says people see it as so capable they even want to use it for applications that don’t make sense such as in his latest company, Robust.ai. “Someone suggested to him recently that it would be cool and efficient to tell his warehouse #robots where to go by building an LLM for his system. In his estimation, however, this is not a reasonable use case for generative AI and would actually slow things down. It’s instead much simpler to connect the robots to a stream of data coming from the warehouse management software.” “When you have 10,000 orders that just came in that you have to ship in two hours, you have to optimize for that. Language is not gonna help; it’s just going to slow things down,” he said. “We have massive data processing and massive AI #optimization techniques and planning. And that’s how we get the orders completed fast.” Another lesson is that you can’t try to do too much. “We need to automate in places where things have already been cleaned up. So the example of my company is we’re doing pretty well in warehouses, and warehouses are actually pretty constrained. The lighting doesn’t change with those big buildings. There’s not stuff lying around on the floor because the people pushing carts would run into that. There’s no floating plastic bags going around. And largely it’s not in the interest of the people who work there to be malicious to the robot." Even if sufficient cleanup was done, "Brooks says we have to accept that there are always going to be hard-to-solve outlier cases when it comes to AI, that could take decades to solve. Without carefully boxing in how an AI system is deployed, there is always a long tail of special cases that take decades to discover and fix.” Brooks also says "that there’s this mistaken belief, mostly thanks to Moore’s law, that there will always be exponential growth when it comes to technology — the idea that if ChatGPT 4 is this good, imagine what ChatGPT 5, 6 and 7 will be like.” “People say, ‘Oh, the large language models are gonna make robots be able to do things they couldn’t do. That’s not where the problem is. The problem with being able to do stuff is about control theory and all sorts of other hardcore math optimization.” #technology #innovation #startups #hype https://lnkd.in/gjPZ5sbH

Yes, we overestimate. In fact, I think the biggest issue is personification. So many excellent people are now diving deep in the human machine teaming, focusing on this specific challenge of GAI. Excellent summary of a good article. Thanks, Dr. Jeffrey Funk. #ArchAITecture #GAI. #GenA #HMT #Trust

2024 feels like a tipping point for Robotics. Acumino is one of the few companies leading this shift, using high-quality real-world data to bring robotics into new, meaningful applications once thought impossible. For more insights on our thesis in the sector check out Nick's post to dive deeper into why the time for robotics is now. 🤖

🎉 Big Pi Point of View Series: #8 🤖 Why Robotics is Set to Shape Our Future Robotics is at an exciting turning point, similar to where AI was just a few years ago. By combining robotics with real-world data and AI, we’re seeing machines evolve beyond repetitive tasks to take on complex, real-world challenges. In our latest Point of View, our Partner, Nick Kalliagkopoulos, explains why we believe the time for robotics is now and offers more insights behind our investment in Acumino, a global leader in robotics and AI-driver automation. We’d love to hear your suggestions for future topics. Drop them in the comments below!👇 #BigPiVentures #PointOfView #BigPiBlog

5 robotics trends to expect in 2024