Has anyone contemplated what it would be like to have an AI algorithm loaded with boundary laws, case laws and other rules, supplement it with field findings and try to sort it all out?
Are we there yet?
Futurism AI Boundaries?
-
pls5528
- Posts: 241
- Joined: Sat Jan 27, 2007 5:42 pm
Re: Futurism AI Boundaries?
Closest thing I have found was a compilation of Laws (Ca.) prepared and authored by Roy Minnick. It used to be called "Bits and Pieces", and later changed to "Laws for the California Surveyor". Of course, in text, but the best go-to I have ever found.
-
Mike Mueller
- Posts: 328
- Joined: Mon Nov 26, 2012 6:53 am
Re: Futurism AI Boundaries?
Cbarrett:
It is the true end of all professions in my opinion. If you accept the general premise that human brains can handle something like 5 variables at once and most professional situations require the evaluation and weighting of more than 5 variables..... it just makes sense that we will be replaced by an AI. The folks and posts at https://www.lesswrong.com/ goes into the issues with the human brain and what most people consider thinking. Long story short most of us are horrible at thinking, and need lots of practice to get even a little logical, let alone constantly adjust and re-consider what was previously assumed and upon which your life's thinking is predicated on. Considering the basic idea of Max Planck's comment on paradigm shifts IE no one changes their mind, I think an AI would be awesome, since it would be able to actually change its mind :)
Will it happen right away? According to my sister, no.
Will prolly need a new type of AI, as the current ones require ALOT of digitized training material. Not that my sister's opinion is fullproof, but she and her husband have been involved in AI research/programming since roughly 2006ish. The current "off the shelf" AIs need enough training data, like in the order of 10's of thousands of examples (lots more if the data is noisy) for it to work. Considering the current ability of surveyors to agree on a boundary resolution, and the resultant confusion in maps, I think we will need something new/different to come along first.
Or it would have to be one of the "heavily trained by one person" type AIs like AARON (https://en.wikipedia.org/wiki/AARON) but then it would prolly just inherit all the biases from its creator? Would it be locally biased? LOL I can just hear future script kiddies "welp its wrong because you used the So Cal AI in Nor Cal...."
Mikey Mueller. PLS
Sonoma County
PS if you like the stuff at Less Wrong, and have children in your life that like Harry Potter, check out "Harry Potter and the Methods of Rationality"
at http://www.hpmor.com/ inceptions kids into thinking better :)
It is the true end of all professions in my opinion. If you accept the general premise that human brains can handle something like 5 variables at once and most professional situations require the evaluation and weighting of more than 5 variables..... it just makes sense that we will be replaced by an AI. The folks and posts at https://www.lesswrong.com/ goes into the issues with the human brain and what most people consider thinking. Long story short most of us are horrible at thinking, and need lots of practice to get even a little logical, let alone constantly adjust and re-consider what was previously assumed and upon which your life's thinking is predicated on. Considering the basic idea of Max Planck's comment on paradigm shifts IE no one changes their mind, I think an AI would be awesome, since it would be able to actually change its mind :)
Will it happen right away? According to my sister, no.
Will prolly need a new type of AI, as the current ones require ALOT of digitized training material. Not that my sister's opinion is fullproof, but she and her husband have been involved in AI research/programming since roughly 2006ish. The current "off the shelf" AIs need enough training data, like in the order of 10's of thousands of examples (lots more if the data is noisy) for it to work. Considering the current ability of surveyors to agree on a boundary resolution, and the resultant confusion in maps, I think we will need something new/different to come along first.
Or it would have to be one of the "heavily trained by one person" type AIs like AARON (https://en.wikipedia.org/wiki/AARON) but then it would prolly just inherit all the biases from its creator? Would it be locally biased? LOL I can just hear future script kiddies "welp its wrong because you used the So Cal AI in Nor Cal...."
Mikey Mueller. PLS
Sonoma County
PS if you like the stuff at Less Wrong, and have children in your life that like Harry Potter, check out "Harry Potter and the Methods of Rationality"
at http://www.hpmor.com/ inceptions kids into thinking better :)
-
CBarrett
- Posts: 769
- Joined: Thu Dec 16, 2021 12:55 pm
Re: Futurism AI Boundaries?
I am loving the lesswrong website.
The way our brains work is fascinating. When I had extra time on my hands during recession I did a lot of additional digging and studying about psychology and philosophy and related concepts. Rather fascinating. It gets very interesting when I see some of those at work in discussions with our fellow professionals. Little frustrating at times when You realize that that they are not aware of some concepts which the other science would consider simple basics.
That's where a good chunk of my exasperation with professionalism and emotional intelligence comes from. I start expecting that seasonal professionals would have a hefty does of self awareness - and this is often not the case.
For example, I make a significant distinction between words 'think' and 'feel'. They are two different processes. Then you see a technical professional say "I FEEL that this law says .. (fill in the blanks).... "and I immediately wonder if I am going to get a rational thought or a knee jerk reaction, or some sort of a jaggedly awkward mix of the two.
The way our brains work is fascinating. When I had extra time on my hands during recession I did a lot of additional digging and studying about psychology and philosophy and related concepts. Rather fascinating. It gets very interesting when I see some of those at work in discussions with our fellow professionals. Little frustrating at times when You realize that that they are not aware of some concepts which the other science would consider simple basics.
That's where a good chunk of my exasperation with professionalism and emotional intelligence comes from. I start expecting that seasonal professionals would have a hefty does of self awareness - and this is often not the case.
For example, I make a significant distinction between words 'think' and 'feel'. They are two different processes. Then you see a technical professional say "I FEEL that this law says .. (fill in the blanks).... "and I immediately wonder if I am going to get a rational thought or a knee jerk reaction, or some sort of a jaggedly awkward mix of the two.
-
Mike Mueller
- Posts: 328
- Joined: Mon Nov 26, 2012 6:53 am
Re: Futurism AI Boundaries?
CBarrett:
Not sure if it is properly selection bias (https://en.wikipedia.org/wiki/Selection_bias) but I think this forum is not too full of AI folk :)
Oddly enough, this topic in some form has been going on a long time if you consider Ludd's issues of replacement as similiar to AI replacement. Anytime a topic or "thing" has been going on a while, I generally assume it will keep going simply because its rare to be present for a sudden shift in the baseline.
I think Ludd was right for his small world, however I am glad he lost. I think all industries that get mechanized/AIzed will likley follow the same progression of turmoil, fights, and ultimately efficiency will win and everyone not in the effected industry will be happier with cheaper better (on average) work. I would be worried if AI replaced too many things at once, so that there was too much turmoil.
The only thing different about AI replacement is I think enough people realize that we need to keep people busy and working, since it is cheaper than welfare supporting large masses of discontent unemployed folk (think France's youth). I expect it will be like the fishing industry of New England area, or farmers with the fallow field business models. , where government comes in and regulates innovation to ensure our old outdated modes still work. Uber is the most recent example I can think of. That company is losing money hand over fist, buying market share and hoping to survive long enough for AI to replace the human drivers. I do not think they will succeed because government will come in and "protect" people from a "dangerous" robot driver. At least long enough for enough of the taxi drivers and long haul truckers to find a new job so that they don't get pitchforks out. Or when they do there aren't enough to cause too big a problem at the voting booth. Which means Uber will run out of money before they can become profitable... At least thats my bet :)
Of course all the doom-n-gloomers could be right and there really will be a singularity (https://en.wikipedia.org/wiki/Technological_singularity) in the 2030-2040 timeframe which allows for a Matrix or Terminator style takeover. I only use those as they have cultural awareness, not because I think that is the actual path of human replacement that will/could occur. In which case, ohh well, it was a fun ride :) (que Dr Strangelove)
Mikey Mueller, PLS
Sonoma County
PS If you ever get a chance to read the Unabombers Manifesto, he was worried about this too, albeit not specifically AI, but he was writing in the 80's/90's (https://en.wikipedia.org/wiki/Unabomber_Manifesto)
Not sure if it is properly selection bias (https://en.wikipedia.org/wiki/Selection_bias) but I think this forum is not too full of AI folk :)
Oddly enough, this topic in some form has been going on a long time if you consider Ludd's issues of replacement as similiar to AI replacement. Anytime a topic or "thing" has been going on a while, I generally assume it will keep going simply because its rare to be present for a sudden shift in the baseline.
I think Ludd was right for his small world, however I am glad he lost. I think all industries that get mechanized/AIzed will likley follow the same progression of turmoil, fights, and ultimately efficiency will win and everyone not in the effected industry will be happier with cheaper better (on average) work. I would be worried if AI replaced too many things at once, so that there was too much turmoil.
The only thing different about AI replacement is I think enough people realize that we need to keep people busy and working, since it is cheaper than welfare supporting large masses of discontent unemployed folk (think France's youth). I expect it will be like the fishing industry of New England area, or farmers with the fallow field business models. , where government comes in and regulates innovation to ensure our old outdated modes still work. Uber is the most recent example I can think of. That company is losing money hand over fist, buying market share and hoping to survive long enough for AI to replace the human drivers. I do not think they will succeed because government will come in and "protect" people from a "dangerous" robot driver. At least long enough for enough of the taxi drivers and long haul truckers to find a new job so that they don't get pitchforks out. Or when they do there aren't enough to cause too big a problem at the voting booth. Which means Uber will run out of money before they can become profitable... At least thats my bet :)
Of course all the doom-n-gloomers could be right and there really will be a singularity (https://en.wikipedia.org/wiki/Technological_singularity) in the 2030-2040 timeframe which allows for a Matrix or Terminator style takeover. I only use those as they have cultural awareness, not because I think that is the actual path of human replacement that will/could occur. In which case, ohh well, it was a fun ride :) (que Dr Strangelove)
Mikey Mueller, PLS
Sonoma County
PS If you ever get a chance to read the Unabombers Manifesto, he was worried about this too, albeit not specifically AI, but he was writing in the 80's/90's (https://en.wikipedia.org/wiki/Unabomber_Manifesto)
-
hellsangle
- Posts: 710
- Joined: Tue Mar 06, 2007 8:31 am
- Location: Sonoma, CA
- Contact:
Re: Futurism AI Boundaries?
Ahhhh let the computer do the analysis, huh?
https://www.npr.org/2017/08/18/54397600 ... n-s-murder
Not that our opinions area life or death situation. Or are they?
https://abcnews.go.com/US/simple-disput ... d=24695871
https://www.npr.org/2017/08/18/54397600 ... n-s-murder
Not that our opinions area life or death situation. Or are they?
https://abcnews.go.com/US/simple-disput ... d=24695871
-
Mike Mueller
- Posts: 328
- Joined: Mon Nov 26, 2012 6:53 am
Re: Futurism AI Boundaries?
"The head of the non-profit that runs this assessment tool for the San Francisco sheriff's office, called the Pretrial Diversion Project, tells NPR it appears her staff mistakenly entered incorrect data about suspect Mims. The project's CEO Nancy Rubin says someone miscalculated the number of days Mims had previously spent in jail, which created what she calls an incorrect score for the judge to review."
Quote from the article about the bail Algorithm. Human error. Ever had a calculator give you a wrong answer? Not to say i support the replacement of humans, but I think that the case against AI should be rooted in the idea that humans need something to do, like border collies :) Efficacy arguments will either need ever changing goals ( read lost debate) or be conceded when thresholds are met.
"The term "computer", in use from the early 17th century (the first known written reference dates from 1613),[1] meant "one who computes": a person performing mathematical calculations, before electronic computers became commercially available." quote from https://en.wikipedia.org/wiki/Computer_(occupation) bolding mine. Change happens and new things to do will (hopefully) keep changing too. Its why I agree with the idea that all people that want to be surveyors in 30 years need to be able work in the office.
Who on this forum actually believes that youtube influencer is a real job? I didn't, and still don't really, but PewDiePie proves me wrong...https://en.wikipedia.org/wiki/PewDiePie
Mikey Mueller
Quote from the article about the bail Algorithm. Human error. Ever had a calculator give you a wrong answer? Not to say i support the replacement of humans, but I think that the case against AI should be rooted in the idea that humans need something to do, like border collies :) Efficacy arguments will either need ever changing goals ( read lost debate) or be conceded when thresholds are met.
"The term "computer", in use from the early 17th century (the first known written reference dates from 1613),[1] meant "one who computes": a person performing mathematical calculations, before electronic computers became commercially available." quote from https://en.wikipedia.org/wiki/Computer_(occupation) bolding mine. Change happens and new things to do will (hopefully) keep changing too. Its why I agree with the idea that all people that want to be surveyors in 30 years need to be able work in the office.
Who on this forum actually believes that youtube influencer is a real job? I didn't, and still don't really, but PewDiePie proves me wrong...https://en.wikipedia.org/wiki/PewDiePie
Mikey Mueller
-
Peter Ehlert
- Posts: 717
- Joined: Sun Oct 26, 2003 2:40 pm
- Location: N31°43', W116°39'
- Contact:
Re: Futurism AI Boundaries?
"Ever had a calculator give you a wrong answer?"
Yes.
whenever a new version of Acad or the Add-on DCA came out we checked to see if it Still gave us wrong answers in several areas.
there was another CAD program (not Acad based) that was mandated at one of my employers that we had to fight constantly.
AI is software, created by fallible Humans. Errors are made, and some are not detected before AI is used to create code that has even More errors....
===
Closed Source Software is only reviewed and edited and patched by a select few, under direction and control of ______
known bugs live and thrive for Decades
Yes.
whenever a new version of Acad or the Add-on DCA came out we checked to see if it Still gave us wrong answers in several areas.
there was another CAD program (not Acad based) that was mandated at one of my employers that we had to fight constantly.
AI is software, created by fallible Humans. Errors are made, and some are not detected before AI is used to create code that has even More errors....
===
Closed Source Software is only reviewed and edited and patched by a select few, under direction and control of ______
known bugs live and thrive for Decades
Peter Ehlert PLS 5116 (retired)
-
CBarrett
- Posts: 769
- Joined: Thu Dec 16, 2021 12:55 pm
Re: Futurism AI Boundaries?
Errors will always be present, what causes them and how they manifest themselves may change.
Civil 3D still can't do closure calcs which match it's annotation 100%. I wish I knew what kind of math they are using to pull that off, alas it is not published. so much wasted time proving that your cad calcs close, and we still can't submit closed polyline printout ....
Civil 3D still can't do closure calcs which match it's annotation 100%. I wish I knew what kind of math they are using to pull that off, alas it is not published. so much wasted time proving that your cad calcs close, and we still can't submit closed polyline printout ....
-
Jim Frame
- Posts: 1599
- Joined: Thu Oct 17, 2002 8:52 pm
- Location: Davis, CA
- Contact:
Re: Futurism AI Boundaries?
That's pretty amazing -- running closures must be a high-demand task for multidisciplinary shops that churn out subdivisions, the sort of customers that C3D is aimed at. I would think that fixing that bug would be high on the priority list.Civil 3D still can't do closure calcs which match it's annotation 100%.
-
hellsangle
- Posts: 710
- Joined: Tue Mar 06, 2007 8:31 am
- Location: Sonoma, CA
- Contact:
Re: Futurism AI Boundaries?
Don't ya just love the folks at AUTOCAD, Jim?
They always put the the consumer first over profits. (Yeah, when pigs fly, huh?)
Crazy Phil
They always put the the consumer first over profits. (Yeah, when pigs fly, huh?)
Crazy Phil
-
CBarrett
- Posts: 769
- Joined: Thu Dec 16, 2021 12:55 pm
Re: Futurism AI Boundaries?
Oh, don't get me started! We don't run closures in Civil 3D any more, switched to smartdraft... although smartdraft has it's own set of glitches.Jim Frame wrote: Mon Sep 19, 2022 1:10 pmThat's pretty amazing -- running closures must be a high-demand task for multidisciplinary shops that churn out subdivisions, the sort of customers that C3D is aimed at. I would think that fixing that bug would be high on the priority list.Civil 3D still can't do closure calcs which match it's annotation 100%.
I keep threatening I will write a closure program myself (in my all spare time.)
It can be a time sink, and it shouldn't be, chasing a damn rounding error around and around.
-
Mike Mueller
- Posts: 328
- Joined: Mon Nov 26, 2012 6:53 am
Re: Futurism AI Boundaries?
I haven't bothered to check C3D as thoroughly as I did LDD, so I am not sure if its still a "feature", but LDD had issues within rounding up the decimal seconds, and it always rounded down. As you said, it only really mattered to the closures, or when I was rotating for the BoB (Basis of Bearings) and forgot to switch the displayed units to 2 decimals of seconds prior to the edit point/rotate command.
Mikey Mueller, PLS
Sonoma County
Mikey Mueller, PLS
Sonoma County