LESSWRONG
LW

All of frontiersummit's Comments + Replies

How AI Takeover Might Happen in 2 Years
frontiersummit2-2

. The main bottleneck to robotics was brains not limbs

 

I feel that it is an overly optimistic take on the state of robotics, particularly the degree of human maintenance that equipment like this requires, how little of it is produced today, and how long it takes to bring new manufacturing on-line considering the deep supply chain needed. The robots described are not the robots of 2027, or even 2127. They are the robots from a long time ago in a galaxy far, far away.

Reply
2Noosphere89
I know you had discourse on this before, and short form, even assuming the critics were correct and there were no mitigating factors, this still gives us an estimate of 250 million robots in 5 years (Though I agree it invalidates the story told), and also robots don't have to be humanoid, which makes the task a lot easier: https://www.lesswrong.com/posts/6Jo4oCzPuXYgmB45q/#WwEvZ7zdPpJppvmhc https://www.lesswrong.com/posts/6Jo4oCzPuXYgmB45q/#nRMkDCYmDCKhXEAJM I agree that we can't get the story of cool robots by 2027, but it's nowhere near as impossible as you portrayed it.
How quickly could robots scale up?
frontiersummit52

The average American drives 45,000 miles in three years, but a car operated 20/7 (like your robot) would accumulate about a million miles in that timeframe. Probably it would go through 2 engines and 3 transmissions if it could even be kept on the road. All things being equal it would need 22x as much maintenance than the average of the US fleet, so probably more like 220% of the capital cost.

A really nice printer/photocopier-combo costs about $10,000 like your robot, and is built from of motors, cameras, and computers just like your robot. While it's mature technology and built to generally high quality standards, if you try running copies 24/7 you will quickly be on a first-name basis with local Kyocera guy.

Reply
1Benjamin_Todd
That's helpful! Makes me think the all in hardware costs could be off by a factor of 2x. 
How quickly could robots scale up?
frontiersummit3-2

Fly into Monterrey Mexico sometime and notice who is on the flight with you: Nearly everyone will be a technician-looking guy in his 30s or 40s. The times I've done it, probably 80% of the flight (including me and my colleagues) fit that criteria. The eastern side of that city is packed wall-to-wall with shiny new manufacturing plants, each filled to the brim with the latest and greatest in industrial automation, robots not excluded. Many of those foreign technician-looking guys are wearing polo shirts emblazoned with the logos of the companies which manuf... (read more)

Reply
3Benjamin_Todd
I did wonder about maintenance costs, but I figured they wouldn't change the picture too much because I only assume an avg 3 year lifetime for the robot, and figured they wouldn't need a huge amount of maintenance to make it to that point. Moreover, if there's worthwhile maintenance that extends the lifetime further, then the hardware costs could end up cheaper than my per year estimate.  I'm also envisioning the costs after a big scale up, and there would be robot repair shops as numerous as car repair, rather than needing to fly in specialists. That said, I agree it would be interesting to look at how much is spent on car maintenance per year on a car vs. capital costs. (I expect it would be under 10%?)