chfritz

How horrendously incomplete! Just four physical robots? You don't even have Roombas and similar cleaning robots on it. And why do we need another such list? I think https://robotsguide.com/ is pretty good. When we were working on Lumin Robotics we collected a list of just the UVC robots: https://luminrobotics.com/companies/

So now Elon makes not only boring products, but also others that just suck?

What I'm trying to say: robotics is a very broad field and your current skills are needed just as well. And while working on those aspects of the robotics stack you can learn about the rest if you want. But you don't even have to to progress in your career in the robotics industry.

Definitely 1! As I commented before here, https://www.linkedin.com/advice/0/heres-how-you-can-stay-ahe...:

Many roboticists overlook the need for JavaScript in their skill set and tech stack. This is because of the evolution of robotics projects and companies: first you build a prototype and only then do you deploy and scale your fleet. The first part is hard and doesn't require any JS. But the second part is even harder and requires a ton of JS, because without web-tooling you won't be able to operate efficiently. This offers a great opportunity for roboticists to distinguish themselves from others be mastering JS for both the backend (rosnodejs, rclnodejs) and the front-end (React, Transitive Robotics). This sets you apart from both "regular" roboticists and from "pure" web developers.

Learn ROS. It has tutorials using just turtle sim. Also https://www.theconstruct.ai/ is a place designed for people who want to learn robotics (based on ROS). https://www.youtube.com/@mikelikesrobots also has some nice beginner videos.

This is similar to ROS Board, but with some significant differences. Let me know if you care to hear about that.

Sorry to hear, Kyle! But keep in mind: good ideas always survive, you just need to find the right business model for them. Maybe Basis didn't work as a VC funded business from day one. But take Foxglove again: they came out of the open-source release of webviz from Cruise. That's how they "bootstrapped". Would love to see what would happen if you worked inside another company that wants to build their own middle-ware but also knows that they do not need to own it and would benefit from more users (because more "test coverage", more developers, more extensions). Then, after a couple of year, it might be ready enough where other companies can try it and get value very quickly (and much faster than they could now).

It's Moravac's paradox: Many of the things that are easy for humans (like doing dishes) are still very hard for computers/robots. That is why so far no one has been able to build a robot that can do the things people want help with at home: dishes, laundry, taking out the trash, cleaning. And by cleaning I mean cleaning everything, incl. toilets, window shades, and the shower not just vacuuming and mopping the floor. It's hard enough to build a robot that can do one thing really, really well and 99.9% autonomously, which is why we still see innovation in things like robot vacuuming (see Matic).

https://en.wikipedia.org/wiki/Moravec%27s_paradox

Duplicate of https://news.ycombinator.com/item?id=42457213

> to do it all over the internet, but adds the associated serialization and transport latency.

I wrote this blog post on that topic a while back after having seen various approaches robotics companies take and their shortcomings: https://transitiverobotics.com/blog/streaming-video-from-rob...

I've heard about you guys before. Nice application! What are you using for teleop/remote-monitoring? Did you build that yourself?

WebRTC is end-to-end encrypted, too, and unlike OpenMLS is designed for video streaming, so it's great for low-latency live viewing, even over terrible network connections. We are using it in robotic and built a component you can just embed in any web page/app (not open-source, sorry): https://transitiverobotics.com/caps/transitive-robotics/webr... On the device side it supports hardware acceleration on Nvidia, RockChip, and Intel devices with VA-API support (e.g., NUKs).

Funny, just yesterday I was complaining about OpenAI's gptbot's stupidity costing site operators too much: https://x.com/chfritz/status/1863689365740020012

Not new, but still very cool!

No pictures of the mobile robots?

> We find that robot adoption is accompanied by an increase in employment and retention and the relationship is strongest for nonregular care workers and monitoring robots. The share of specific tasks performed by robots increases with the adoption of the respective type of robot, leading to reallocation of care worker effort to “human touch” tasks that support quality care.

Nice!

Interesting question! I think you should run an experiment! You could remote control a telepresence robot, call an Uber for it, ask the driver to load and unload the robot, and then continue remote controlling when you arrive wherever you wanted to go. There there will be more obstacles, especially doors, elevator buttons to be pushed, and non-ADA places, I assume. You could ask by-standers for help, or just tailgate people (esp. into elevators). I'd love to read the result of such an experiment.

Were you involved in the research of the video from Stanford you are showing?

> These QR codes are not reliably scannable!

Isn't the whole point of QR codes to reliably transfer information from the physical world into the digital one without specialized equipment or software? They look pretty, sure, but so did DataGlyphs (https://www.microglyphs.com/english/html/dataglyphs.shtml) and those could be scanned reliably, just not by your average phone.

Cool. And I noticed that a surprisingly high number of songs are in Spanish. So I'll venture to hypothesize that this project will identify a correlation between musical taste and preference for how loud it is played, rather than accurately capturing the "musical taste of the neighborhood". Any thoughts on that? Have you tested how loud a song needs to be played in order to be picked up?

Re. infrastructure integration: it's always a cost-benefit analysis. I've worked at a robotics company where we integrated with doors and elevators. Doors was really easy, cost almost nothing, and didn't come with any regulations. Elevators, on the other hand, was a length process, required certified elevator technicians, and cost a lot of money. On the other hand, adding a manipulator to open manual doors is very difficult and costly (per robot), but adding a button-pusher for elevator buttons is not.

Humanoids are not a threat to anyone. Not from China and not from anywhere. They'll fall flat on their face, all by themselves.

> this essay looks at how AI and automation are flattening the Phillips curve

Sounds like a causal relationship between a very narrowly focused observation (more AI + robots) and a very general phenomenon (Phillips curve flattening). This is a futile hypothesis to begin with.

> There is an entire industry in trying to understand the changing curve, and I will just lay out the main arguments made.

And then you just go and dismiss them, without providing evidence, just to claim (not prove or support) that your own theory is a better explanation, i.e., AI and Robots are to blame.

> Why is bargaining power changing? In a word, automation.

Automation has existed for 100+ years. So you are actually providing evidence against your own argument here, because, as you noted, the graph did not flatten until, allegedly, 2009 -- and even then I'm with Terr. BTW, what significant event happened in 2009? If AI will revolutionize labor then that's only starting now (maybe as of 2022). And the robotics market has not had any significant break-through event either. It has grown very steadily, but so has, say, the number of smartphones.

How is your approach different from Peanut's?

OK, so now we have: Basis Robotics (C++), Copper Robotics (Rust), and NComm (Rust) all trying to replace ROS. Shouldn't you guys be working together instead? Also, can't we just all instead improve ROS instead? I think Zenoh is leading the way there, introducing a new ROS 2 middleware, rather than trying to replace it wholesale.

Also, in case you didn't know, some much larger companies tried to replace ROS and eventually saw the light and gave up: NVidia (Isaac, yes, Isaac was at first a new framework for robotics that was meant to replace ROS, they've since reused that name to mean something else), and Viam (Go; now a ROS sponsor).

Why limit this to arxiv papers? Why not any paper published online, e.g., via https://bibbase.org? btw, very cool that you seem to have overcome the initial inertia of getting something like this going. The idea is not new, but it's a marketplace dynamic that is hard to bootstrap.

Very cool. Thanks for starting this!

Nice use case! Can you elaborate a bit more on robotics piece? What role does the robot play? I assume it's required to turn the part around for inspection. If so, how do you (automatically?) compute the grasping pointing? Also feel free to find me on LinkedIn if you want to chat more about growing a robotics businesses and/or geometric reasoning for manufacturing.

It doesn't sound like a robot at all, because it doesn't "plan". A robot is an embodied cognitive agent. Embodied = has a body and is acting in physical space. Cognitive agent = it uses a sense-plan-act loop. This thing may sense and act, but its controls are too primitive to call it "planning". In particular it doesn't plan multiple actions into the future. It's doing something folks in the AI literature would call "reactive control".

It's not robotics, because it doesn't "plan". A robot is an embodied cognitive agent. Embodies = has a body and is acting in physical space. Cognitive agent = it uses a sense-plan-act loop. This thing may sense and act, but its controls are too primitive to call it "planning". In particular it doesn't plan multiple actions into the future. It's doing something folks in the AI literature would call "reactive control".