The Consumer Robotics Innovator's Dilemma, Apple, and Vehicles Revisited

The Consumer Robotics Innovator's Dilemma, Apple, and Vehicles Revisited

The Tragedies of Reality in Robotics and Rodney Brook's Three Laws of Robotics


Leaky abstractions and reality not hardware as friction. Google Research Scientist, Alex Irpan, recently wrote a post ‘The Tragedies of Reality’ that is doing the rounds on the heels of Benjie Holson’s ‘Mythical Non-Roboticist’ and it’s very good at explaining why robotics is already working on all the things that currently make LLMs tricky. Basically, reality.

  • The reason robot learning progress is slower is because it’s very hard to do anything without tackling the hard problems.
  • The hard problems of robotics are not unique to robotics.

There’s a saying that “all robot demos lie”, and people are discovering all LLM demos lie too. I think this is fundamentally impossible to avoid, because of the limitations of human attention. What’s important is evaluating the type, size, and importance of the lie. Did they show how it could generalize? Did they mention how cherry-picked the examples were? These questions become more complicated once you connect reality into the mix. Sure, Messi’s looked like a good player so far, but “can he do it on a cold rainy night in Stoke”?

Since this post, Alex Irpan has moved from Robotics into AI Safety, which makes a lot of sense.


Rodney Brooks released his Three Laws of Robotics , which generally speaking I completely agree with. These Three Laws are designed for robot builders. The 5 Laws of Robotics that I’ve proposed cover some general principles not covered here, if you want a more abstracted position.

Here are some of the things I’ve learned about robotics after working in the field for almost five decades. In honor of Isaac Asimov and Arthur C. Clarke, my two boyhood go-to science fiction writers, I’m calling them my three laws of robotics.

  1. The visual appearance of a robot makes a promise about what it can do and how smart it is. It needs to deliver or slightly overdeliver on that promise or it will not be accepted.
  2. When robots and people coexist in the same spaces, the robots must not take away from people’s agency, particularly when the robots are failing, as inevitably they will at times.
  3. Technologies for robots need 10+ years of steady improvement beyond lab demos of the target tasks to mature to low cost and to have their limitations characterized well enough that they can deliver 99.9 percent of the time. Every 10 more years gets another 9 in reliability.

Below I explain each of these laws in more detail. But in a related post here are my three laws of artificial intelligence .


Thanks for reading Robots & Startups! This post is public so feel free to share it.

Share


Robot Expo Opportunities through SVR

  • 2024 SVR Robotics Investment Forum - Oct/Nov (date & venue coming this week!) - startup showcase and 1:1 investor matchmaking
  • 2025 & 2026 CES Eureka Park pavilion - for startups <5yrs - interested?
  • 2025 Turkey Poland tour for manufacturing startups with DOC


Want a byte-sized version of Hacker News? Try TLDR’s free daily newsletter.

TLDR covers the most interesting tech, science, and coding news in just 5 minutes.

Subscribe for free!


Autonomous Vehicles - Cars

Another great post is ‘Whymo’ Vincent Vanhoucke’s description of his move from Google to Waymo.

This comes at an interesting time as the Federal Government decides to rethink Vehicle to Vehicle Communications with the release of a V2X Roadmap . (Here’s an overview of V2V )

In robotics, autonomous driving included, the long tail of difficulty is increasingly more about common sense reasoning than about low-level planning and control. It is about a deep understanding of the situation and reasoning not just about geometry, but about the semantics of a scene, which in a multi-agent scenario includes other agents’ actions. Much of it is also about scaling: every order of magnitude you grow your usage, every new context you bring your autonomous system into, brings you to the very edge of your generalization capabilities. In that sense, my favorite grand challenges in perception, semantic understanding and reasoning have not aged a bit. But what is different today is that with large multimodal models we have new tools at our disposal to address them.


Keep going! The rest of this post... and there's a lot more robot news and events.... is over at Substack https://robotsandstartups.substack.com

Share

要查看或添加评论,请登录

社区洞察

其他会员也浏览了