239. The Saucies: 2025 Recap & TNB Turns Five!

Manufacturing

An MIT team orchestrated existing tools to translate voice commands into 3D designs and robotic toolpaths, enabling the automated manufacturing of reconfigurable furniture pieces.

“they fused, orchestrated all these existing tools, fused them into one stock together to where you can basically be like, Jarvis, make me a water bottle. And then that voice is translated into text, then LM can understand that LM turns your text into a CAD, a 3D design, and then it also talks, it creates that, returns that 3D design into a tool path for robotic manufacturing…”

Engineering & simulationdeep learning protein design tool

A team used deep learning to determine the best combination of components to neutralize snake venom toxins, significantly reducing the research workload required to create antidotes.

“what this team was able to do was leverage deep learning for the protein design tool that then determines the best combination to neutralize the toxin found in the snake venom. And then that offloads a significant amount of the work from the team so that all they have to do. is just create that venom. Antidote and then verified in the lab.”