The Next Byte Podcast · 7 Jan 2025 · 36m
The hosts of The Next Byte Podcast present their 2024 'Saucies' awards, reviewing ten standout episodes from the past year. They highlight technologies ranging from AI-generated 3D design by Raymond Whitekamp and neuroprosthetics from ETH Zurich to Dyson’s automated strawberry farm and Stanford’s mobile Aloha robot. The discussion also covers practical applications like Audi’s use of 3D printed jigs, MIT’s AI doctor co-pilots, and smart traffic systems from the University of Michigan, concluding with an interview on AI safety with Karsten Roscher from the Fraunhofer Institute.
Where this conversation describes AI being used to build or deliver a product. Part of the AI in practice tracker.
Engineering & simulationAny3D
An AI system used ChatGPT to generate code that controlled CAD software to automatically design and create 3D objects from text descriptions.
“The 1st part was chat GPT, essentially. You have an LLM that would take whatever input you gave … And then it would feed it into code that was hooked up to some sort of, you know, CAT system … that used code to generate that object.”
Speaker 1 · Hear it at 4:16
ManufacturingAudidemand for jigs went up to a couple 100
Audi used generative design software to allow non-designers to create custom 3D-printed jigs for the assembly line, increasing the number of jigs produced from a few to hundreds.
“there was a software that we discussed, which allowed someone that might not even know how to do mechanical design, to just kind of explain what they want to do, and it would kind of help you generate your idea and get it printed pretty easily.”
Speaker 1 · Hear it at 20:34
Engineering & simulationStanford University / Toyotacomplete over 850 common household tasks
Researchers used imitation learning to train a robot to perform complex tasks like frying shrimp by observing human control, enabling it to complete over 850 household tasks faster than traditional programming.
“they created a robotic platform that was capable of learning tasks through imitation rather than programming. So they used artificial intelligence and machine learning to help a robot learn how to do something, as it was being controlled by a human…”
Speaker 2 · Hear it at 13:39
Running the companyUniversity of Michigansave the extent of about 3 days of person's life every year
Researchers used AI to control smart stoplights in response to GPS data from connected cars, reducing commute times and saving approximately three days of a person's life per year.
“They were able to, they tested this in a small city, um, by connecting GPS to connected cars and then having smart stoplights that can be controlled. Um, essentially in response to traffic data. They were able to think, or they could save the extent of about 3 days of person's life every year…”
Speaker 2 · Hear it at 29:09
makers and hobbyists
Other shows on the same ground, closest first.
Host Steve Crowe and the Robot Report editorial team (Mike Oitzman, Eugene Demaitre, Brianna Wessling) discuss the highlights of the 2024 RBR50 Innovation Awards. They highlight Agility Robotics' Digit as Robot of the Year, Bright Pick's Auto Picker as Application of the Year, and Electric Sheep as Startup of the Year, while also reviewing standout winners like Glidance, Harvard’s Parkinson’s exoskeleton, and Flip Robotics.
Hackaday editors Mike Szczys and Elliot Williams discuss new AI-on-the-edge hardware from BeagleBone and Google, the reverse engineering of a keytar firmware, and a critical analysis of Tesla's self-driving safety data. They also cover a custom smart lamp build, flexible PCB contests, and Elliot provides an audio tour of the Make Munich maker fair.
Steve Crowe, Mike Oitsman, and Dan Kara of Robot Report and WTWH Media review the 2022 RBR50 Innovation Awards, highlighting nine standout companies including Cruise, Nvidia, Savio Robotics, Labrador Systems, Soul Robotics, Insight Track, Swiss Mile, and Verizon. The hosts discuss specific products such as Cruise's Level 4 robotaxi, Nvidia's Omniverse Replicator for synthetic data, and Savio Robotics' Relay Plus elevator interface. They analyze the engineering and business decisions behind these innovations, from autonomous vehicle safety to agricultural sanitation and hybrid mobility platforms. The episode emphasizes how these solutions address complex challenges in robotics, from scaling production to solving niche market problems.
Hackaday editors Elliot Williams and Kristina Panos discuss a range of hardware hacks and industry news, including the debate over AI-generated art winning a Colorado State Fair competition and Japan's move to outlaw floppy disks. They review specific projects such as Matt Thomas's study on making 3D-printed objects food-safe, Fletcher Hasler's AI lie-detecting app, and Revox Dinah's active foot cooling prototype. The episode also covers the decline of self-hosted email due to Google's spam filtering, the history of the Nintendo U-Force, and Wesley Treat's heat-shielding solution for hot shops.
Hackaday editors Elliot Williams and Tom Nardi discuss a range of hardware hacks, including an AI-driven surgical robot from Johns Hopkins University that performs ex vivo pig bladder surgery, and budget robotic dog actuators inspired by MIT's Mini Cheetah. They also cover the restoration of a FAA-certified flight simulator, the history of the Blu-ray vs. HD DVD format war, and a clandestine machine shop built by POWs in WWII. The episode concludes with an update on the struggling Lunar Trailblazer spacecraft and entries from the One Hertz Challenge.
The summary, chapters and topics are generated from the episode’s audio by this directory, not written by The Next Byte Podcast. The episode itself belongs to its publisher.