> I am sure that engineers, with some effort, could program a machine that could take the human out of the tree-cutting process.
If that engineering is a mental task, that AIs are already ahead of us on, then the only limiting factor on them replacing the humans is that using human labor is relatively cheap, and the AI has better things to do. Humans are just legacy hardware that hasn't been replaced yet.
Probably a year or few years to get first prototypes working, but perhaps an entire decade to engineer a profitable product that takes significant market share from existing products. The autonomous car being an example with some evidence behind it. One of the big problems with robots is that their optical-electrical-mechanical-software systems are vulnerable to moisture, dust, and wear. Engineers can make it work for a while, but once in the hands of the customer will it keep working as intended for a reasonable cost? Probably not. A product has to be able to last 5-10 years with little maintenance at a price similar to existing products. There aren’t enough sufficiently well-trained engineering teams to build these systems at reasonable costs. Hence autonomous flying cars, robotic weeders, dishwasher emptiers, robotic window cleaners, autonomous fishing boats, have not arrived. Still we live in a world of abundance where American children do relatively few chores. Perhaps we should consider having our children do more chores rather than less?
Lastly, the places where these systems would likely be engineered are places where woke ideology tends to thrive, Silicon Valley, Boston, Seattle, Los Angeles, Boulder, etc. This further reduces the number of engineers available for creating products that could help us.
Arnold, you need to get out from behind the screen and get into the world where folks make stuff. I'm not going to list all the stuff that 8 axis robotic arms have been trained to do that only a few years ago I would say "no way".
Of course that stuff isn't here today in an economically viable form. All the commentators talk about today...it can't do this or that, etc., etc...
Wait a couple more years. Taking down a tree? No biggie.
Yup. And of course, there will be dissenters. They will be old guys proven wrong.
I live most of the year in Wuhan. I've seen what their agricultural robotics have already attained. Where they're at now is mind boggling. Where they will be in a few years will blow folks minds.
Pre-pandemic I interviewed at Resonon in Bozeman, Montana. Really cool hyper-spectral cameras. https://resonon.com. Apparently most are purchased by Chinese companies, likely in the agricultural and manufacturing sectors. So I'm not surprised that their homegrown agricultural machines are taking off.
In graduate school, my advisor began a partnership with an institute in China that was building space-based optical systems likely for military purposes. He noted amazing progress in the years between his visits. What was he going to get out of the partnership? He didn't know, but he was going with them with Arizona tax payers funding a portion of it. I didn't feel right about that especially because I was working on a defense related optical system and was being asked to send slides to China summarizing progress on my work. This is the type of situation that sets off a red flag during DOD secret clearance interviews after graduation. Let's just say that the FBI interviewed me a number of times.
I know how to design and build, but due to politics in Silicon Valley I quit and we “fled to the wilderness.” Family is safe now, so I could get back into it.
Yes, there’s so much distrust between our governments. We could be doing a lot more to build a better world, but incentives are not there.
"There are many computers that can outperform the average chess-playing human when it comes to strategy. But how many of them can pick up a piece on the board and move it?"
I think an important question is what jobs are protected either by unionization and occupational licensing to such a degree that they will continue to be performed by humans long after an AI can outperform them.
I'm not a lawyer but my observation is that Gemini and other LLMs can already process large swathes of information and compile it into a cohesive, well-argued case in a fraction of the time that it would take a crack team of legal scholars. However, I seriously doubt the implications of that are anything other than lawyers will start using AI to make their jobs significantly less tedious.
Do you not think that some jobs are harder than ppl seem to imagine, for computers? Like self driving cars. 2.5 people on average seem to work behind the scenes in real time to drive a "self driving" car, according to a blog post by Rodney Brooks (an MIT CS prof). My friend working at a leading autonomous car company confirmed this for me.
"There are many computers that can outperform the average chess-playing human when it comes to strategy. But how many of them can pick up a piece on the board and move it?"
Yes, and Herb Simon pointed out as much in ~1960.
https://hunchbox.substack.com/p/predictions-from-a-throw-away-herbert
> I am sure that engineers, with some effort, could program a machine that could take the human out of the tree-cutting process.
If that engineering is a mental task, that AIs are already ahead of us on, then the only limiting factor on them replacing the humans is that using human labor is relatively cheap, and the AI has better things to do. Humans are just legacy hardware that hasn't been replaced yet.
Probably a year or few years to get first prototypes working, but perhaps an entire decade to engineer a profitable product that takes significant market share from existing products. The autonomous car being an example with some evidence behind it. One of the big problems with robots is that their optical-electrical-mechanical-software systems are vulnerable to moisture, dust, and wear. Engineers can make it work for a while, but once in the hands of the customer will it keep working as intended for a reasonable cost? Probably not. A product has to be able to last 5-10 years with little maintenance at a price similar to existing products. There aren’t enough sufficiently well-trained engineering teams to build these systems at reasonable costs. Hence autonomous flying cars, robotic weeders, dishwasher emptiers, robotic window cleaners, autonomous fishing boats, have not arrived. Still we live in a world of abundance where American children do relatively few chores. Perhaps we should consider having our children do more chores rather than less?
Lastly, the places where these systems would likely be engineered are places where woke ideology tends to thrive, Silicon Valley, Boston, Seattle, Los Angeles, Boulder, etc. This further reduces the number of engineers available for creating products that could help us.
Arnold, you need to get out from behind the screen and get into the world where folks make stuff. I'm not going to list all the stuff that 8 axis robotic arms have been trained to do that only a few years ago I would say "no way".
Of course that stuff isn't here today in an economically viable form. All the commentators talk about today...it can't do this or that, etc., etc...
Wait a couple more years. Taking down a tree? No biggie.
Plus a little longer for engineers to discuss these topics here.
Yup. And of course, there will be dissenters. They will be old guys proven wrong.
I live most of the year in Wuhan. I've seen what their agricultural robotics have already attained. Where they're at now is mind boggling. Where they will be in a few years will blow folks minds.
Pre-pandemic I interviewed at Resonon in Bozeman, Montana. Really cool hyper-spectral cameras. https://resonon.com. Apparently most are purchased by Chinese companies, likely in the agricultural and manufacturing sectors. So I'm not surprised that their homegrown agricultural machines are taking off.
In graduate school, my advisor began a partnership with an institute in China that was building space-based optical systems likely for military purposes. He noted amazing progress in the years between his visits. What was he going to get out of the partnership? He didn't know, but he was going with them with Arizona tax payers funding a portion of it. I didn't feel right about that especially because I was working on a defense related optical system and was being asked to send slides to China summarizing progress on my work. This is the type of situation that sets off a red flag during DOD secret clearance interviews after graduation. Let's just say that the FBI interviewed me a number of times.
So, you know stuff. Real stuff.
I live there (here). It's shameful that our relationship with China is so ludicrously stupid. We could be solving the problems of the world.
I know how to design and build, but due to politics in Silicon Valley I quit and we “fled to the wilderness.” Family is safe now, so I could get back into it.
Yes, there’s so much distrust between our governments. We could be doing a lot more to build a better world, but incentives are not there.
"There are many computers that can outperform the average chess-playing human when it comes to strategy. But how many of them can pick up a piece on the board and move it?"
https://youtu.be/RNA4byDFIWo?si=E_tVSr-cdsWGt5nR
I think you grossly underestimate the ability of robotics to do physical tasks. Adding AI will surely enhance that immensely.
I think an important question is what jobs are protected either by unionization and occupational licensing to such a degree that they will continue to be performed by humans long after an AI can outperform them.
I'm not a lawyer but my observation is that Gemini and other LLMs can already process large swathes of information and compile it into a cohesive, well-argued case in a fraction of the time that it would take a crack team of legal scholars. However, I seriously doubt the implications of that are anything other than lawyers will start using AI to make their jobs significantly less tedious.
Do you not think that some jobs are harder than ppl seem to imagine, for computers? Like self driving cars. 2.5 people on average seem to work behind the scenes in real time to drive a "self driving" car, according to a blog post by Rodney Brooks (an MIT CS prof). My friend working at a leading autonomous car company confirmed this for me.
What do you mean "work behind the scenes"? Haven't we had self driving taxis for which this is not the case for a few years now?
A remote indian guy still takes over if things get too complicated. Or so I've heard.
Not if Rodney Brooks is to be believed. I know, shocking.
"There are many computers that can outperform the average chess-playing human when it comes to strategy. But how many of them can pick up a piece on the board and move it?"
Not many, but here is one (fun to watch):
https://www.youtube.com/shorts/US7xz_fw65w