🎙 We held the first meeting of the new season of Lecture Club - Princeton ROB 345: Introduction to Robotics.
Lectures 1–2 (Anirudha Majumdar): why robotics is important now and how the task of "getting from A to B" is turned into a graph search problem. Founders, engineers, a PR specialist, and a psychologist attended - almost all with the same question: when will the GPT moment happen in robotics?
Some thoughts from the discussion:
• Nikita - robots only seem simple: at a hackathon, a manipulator programmed with Claude Code didn't know about the table it was attached to and hit it with full force. What I liked most was how the task "from A to B" is turned into a graph where "it's easier for a drone to fly forward than sideways" - it's just different edge weights.
• Elena Chernysheva (co-founder of an AI dubbing startup) - wants to understand where the bottlenecks in robotics are and why there are still so few robots around. She analyzed the second lecture with Claude Code: he drew everything, and it became clear.
• Maksim (IT founder, in ML since 2018) - the GPT moment will come unevenly, first in narrow cases. The main insight is the uncertainty, which even those building robotics businesses underestimate. Home is the hardest case: even with 99% accuracy, errors accumulate over the day.
• Yaroslav Verinchuk (PR agency for startups and VC) - in six months, a noticeable part of clients are in robotics, wants to understand the topic deeper than narratives for TechCrunch. Suggested participants from Europe do practical assignments on quadcopters together.
• Nikolay Bogatyrev - if the memory problem is solved, a robot will not have to build a world model from scratch but can load a ready "copy" from other robots.
• Anatoly Stansler (product engineer) - agency is a scale, and a dishwasher is also a robot. What caught my attention most was the configuration space: it shows in advance where a robot can be and where it cannot.
• Konstantin Trunin (founder of an AI control platform) - the main unresolved task is to teach robots Asimov's first law and understand where to embed "that very button" of control.
• Ivan Khomyakov (robot programmer, Symbotic) - knows control and planning "from the smoking room" and wants to sort it out. Dreams of a "Lego" robot made of smart blocks: attach an arm - and it works.
• Sam Lapko (business in drone shows) - the GPT moment has already happened in drones. Doubts the economics: hardware will be made cheaper in China, and selling software for hardware is difficult - so why so much money in humanoids?
📺 Meeting recording: https://youtu.be/cx_-FRV53Nc
👍 I'm just starting to develop my YouTube channel - if the recording is helpful, please like and subscribe, it really helps.
🎥 For the next meeting - Lecture 3 · Randomized planning (RRTs): https://youtu.be/KDIe1094FMo
📆 I will announce the date of the next call separately.
Discuss in the chat: @rvnikita_blog_chat
#robotics@rvnikita_blog #education@rvnikita_blog #ai@rvnikita_blog #princeton@rvnikita_blog
