"the funding all these startups are getting should allow them to scale their methods 10x-100x.." .. "Therefore, we might soon see a ChatGPT moment in robotics" -- I don't think so and no, the second statement is NOT entailed by the first. Why would it? Because 100 is a big number? Do you have any idea how much more data LLM needed to be trained for a GPT3 level compared to the data available for robot training right now, and how low dimensional the space is in which LLMs operate compared to robots?
"My intuition is there's 40% chance we will see it this year" -- again, why? Don't you realize that people have been working in robotics for 65 years, and these people don't live under a rock either. They knew about GPT3 because 2023. So why is it NOW less then 10 month you think that this breakthrough will happen?
I think it’s a combination of simulation, YouTube videos, and specially recorded training footage. The last one is expensive, but given the funding these startups receive, I’m pretty sure they can scale their RL methods at least 10x.
How can you seriously state that as a positive? Are you trying to exemplify the new AI-patent-generation-mill? No one in their right mind can expect a person to make 10 meaningful inventions in 4 days.
Working with nine different robotics companies over the course of 10 years has taught us a thing or two about designing robotic full-stack architectures. All this experience went into the design of Transitive, the open-source framework for full-stack robotics. In this new mini-series of blog posts we dive into the three core concepts of the framework.
We think that once you've experienced the magic of full-stack encapsulation, you'll not want to go back to the old ways of writing distributed code, esp. in robotics.
Go through the ROS tutorials, using a simulated robot at first, then buy a tiny AMR kit and do the same thing IRL. Once you have those basics, you can ask yourself into which direction to go next, the obvious choices being 1) deeper into the research of the new-fashioned (bi-)manual manipulation (arms), or 2) more into business and actually build a real-world application for your mobile robot (which will involve a lot of tinkering with hardware). And +1 on what brudgers said. It's a hobby, so have fun.
If someone wants to get into robotics as a hobby for the first time, and the #1 thing you tell them is "start with learning ROS", one questions whether you are trying to help them or sabotage them.
Sure, let's instead encourage them to reinvent the wheel (TF, simulation, architecture, navigation, localization, SLAM, 2D map representation, etc.) that seems much smarter. ROS might be like democracy ("the worst form of government except for all the others we've tried"), but it's still the only stake in the ground around which we've manage to corral a community.
For this question specifically: there are many ways to approach robotics, incl. from the hardware side. But approaching it from the software (sim) side is not invalid, so I don't think down-voting my comment was warranted.
Reverse engineering is illegal in many cases. Aren't you afraid you might be automating the process for your users to get into (legal) trouble? Will your tool warn the user if they are about to violate laws?
Doesn't that depend on what a failure looks like? Failure to achieve the task (with no other side effects) is very different from accidentally killing the patient. Maybe also send a link to that study/article you are referring to?
Awesome, congrats on the launch! Can you say more about the software stack? Is it open-source? I see you are using ROS on the robot, but what about the full-stack (robot + cloud + web)? Did you build all that yourself? At Transitive Robotics we are developing the Transitive open-source framework for full-stack robotics (complementing ROS, not replacing it). On top of Transitive it is easy to build capabilities and we already offer many such modules, e.g., for webrtc video-streaming, teleop, mission dispatch, etc. This means you don't need to build all that yourself and once Transitive is installed on a robot, your end-users can install more capabilities now or in the future. Do you think this could be of interest to your users?
