Is Artificial Intelligence About to Decimate White-Collar Work?

Zager and Evans had this to say . . .

"Jobs are more than earning a living. They give people some purpose in life and a feeling of self-worth and accomplishment. We shouldn't give that away."

There are other things that give purpose in life and a feeling of accomplishment: volunteering, for one, hobbies, for another.

If we find ourselves in a world where we don't need to work for a boss to earn money to survive, we can work for ourselves, or work for the community.

People would be free to devote their lives to whatever their passion is, be it woodworking, painting, composing songs, tending a garden, teaching the youth, or even just hiking from one end of this country to the other, meeting as many of the people who make up this nation as possible.

You may call me a foolish dreamer, but that is the very world that Gene Roddenberry envisioned for Star Trek. Scarcity of life's necessities has been eliminated on Earth, everyone there is free to choose their own path in life, some choose to go to other wolds, to collect the resources needed to maintain an interstellar society, or perhaps just for the challenge of colonizing a brand new frontier.

Remember, humans today are part of an unbroken chain of survivors going back 4 and a quarter billion years. If we've survived all that history, we can survive a little change and social upheaval.

Jobs are more than earning a living. They give people some purpose in life and a feeling of self-worth and accomplishment. We shouldn't give that away.

But having the 'norms' change in ever so subtle ways,... before you know it, you've become a staunch registered democrat... ahhhh,.. person you so despise today.

"They give people some purpose in life and a feeling of self-worth and accomplishment."

Maybe, "They give SOME people purpose...". What came first, the purpose then the job, or the job then the purpose? It could go either way.

Maybe, "They give SOME people purpose...". What came first, the purpose then the job, or the job then the purpose? It could go either way.

That question should be asked at the end of the shift/day. And asked everyday. And yes,... each day the answer can change.

The job can be an end in itself, or can be the means to a greater end, such as supporting a family. Let's get our priorities right, gentlemen. Without families and support of families, we would not be here to debate these issues.

So blame it all on the famned damilies.

We can pick a job, but we don't get to pick our families, just our mate, then the kids pop out, and you wonder "did I make that"?

This has to be wrong on its face, because the white collars run everything, and will not permit it.

In addition, I submit that other more primitive societies that do not care one iota about "the rules" will take over before the AI can be a threat.

Where do you find such rubbish?

Sorry, James. Anymore white collars run nothing. It's all in the policy manual.

There's that pesky manual again.

Makes one yearn for the days of whaling (or something better) on the high seas.

Good article from the NY Times.

I spent some time last week with a mix of robotics researchers and companies that are looking at new ways to use robots. I heard a lot of the usual--get to robots to do the jobs that are dangerous/boring/unpleasant; robots won't eliminate jobs, they'll create more jobs; jobs will come back to the US from overseas if robots can drive costs of production down.

The conference I was at was in Pittsburgh, where coal miners and steelworkers saw their jobs go away (the death throes were well underway when I lived in Pgh in the 1980s) without similar jobs to replace them. The new jobs--mostly generated by the Univ. of Pittsburgh medical center and Carnegie Mellon's computer science/robotics complex--had nothing in common with the displaced workers' skill sets, or interests, for that matter.

I have not yet come across a reassuring prognostication about the future of people who don't have the aptitude for the jobs of the future, or who are near retirement and (quite rightly) don't want to have to go back to school. As another commenter said, I'm glad I'll be pushing up daisies before this revolution plays out.

He is not alone. I understand employment on the other side, is guaranteed, equal wages, and wonderful working conditions. Not to mention the streets paved with purest gold, so pure it is in fact transparent.

I sense the beginnings of an experiment here on Earth: Can we make gold so pure, that it is transparent similarly to glass?

Would the knowledge we have about SPP (electron cloud above the metal that affects how it reflects light) electrons rule transparency out?

How many "9's" would it take to make transparent gold? I understand there are gold films a few atoms thick that do transmit (dimly) some light, but are mainly for reflecting most of the light and heat of a fire.

they'll create more jobs;

Exactly, it may not create a '1 job addition to 1 job lost due to automation', but the job created will be higher educated and paid, at least in the beginning.

When ROBOTS become easier to "train" than HUMANS, *that* is when humans become irrelevant...and, ultimately, expendable.

For example, DOD wants ROBOTS to fight their wars, so that the only CO$T becomes the ROBOTS cost and not LIVES.

Of course, the possibility of SKYNET will always lurk when HUMANS relinquish 'command & control' over to ROBOTS...can we "trust" the ROBOT to always 'roboticly' obey HUMAN orders when HUMANS don't?

One programs a robot, one trains a human being (to think and act rationally), but also to follow instructions (orders) from commanding officer.

If a robot has no commanding officer and it totally autonomous with only one mission (leave that to your imaginings), then it is clear this could be a very dangerous adversary when one ends up on the wrong side of the cross-hairs.

I prefer soldiers whose sworn duty is to serve, protect, defend, and restore order when chaos is taking over. I do not prefer a mindless robot.

If the automated soldier could be made into a cyborg, or even an autonomous rational (self-aware) being, it is still highly dangerous to release these without a "kill" switch, or cease action relay. I prefer soldiers with souls, even if that soul if somewhat flawed and imperfect.

No one can ever entirely predict what mental disease will do to devastate the soul of a soldier to the point that rational decisions are out the window, and that is why the military has a code of laws that regulate behaviors. At the same time, neither civil authorities, nor military enforcement should ever fail to report crimes to each other, as this is part of keeping a healthy fighting force, and maintenance of the safety of the civilian population.

The main reason that this will likely not occur anytime in the foreseeable future is that war is, by its nature, convoluted, unstructured, and unpredictable.

Typically, COAs are made once you have 70% of the information. The platoon leader makes the necessary COA changes mid-engagement based on a combination of situational awareness, status of assets, and overall understanding of the main command intent (the big picture) which is often subjective and intangible - especially in today's counter insurgency landscape.

Robots can augment, but never replace a ground pounder.

On a limited basis, they could make a very effective area denial method, like a minefield that goes fetch in a defined area or with effective IFF.

Certainly, their situational usage benefit is undeniable as a force multiplier, but they will never replace the force outright.

True.

"The main reason that this will likely not occur anytime in the foreseeable future is that war is, by its nature, convoluted, unstructured, and unpredictable."

Don't say that to a quantum physicist. That's based almost entirely on the Heisenberg's Uncertainty Principle. Unless "in the foreseeable future" are the key words here. In which case, if quantum physics can solve that problem, then maybe an unconvoluted, structured, and predictable physical-world/political-world could be established. One can only hope.

Lets not get carried away. Jetpacks, flying cars and living on the moon have been promised as the next big thing over the years that will totally change our way of life. This has not happened, not even slightly even thou the technology is there in some way shape or form.

It will be a slow and gradual transformation giving plenty of time for society to adjust, not some massive technological and commercial jump which makes unskilled labourers obsolete in a year. Even then there will always be a use for unskilled labour.

This has led to economic devastation and a general sense of political unrest.

Has it, or has America (for example) just f$cked itself fallen behind other countries by losing its manufacturing and commercial edge through poor decisions, outsourcing rather than training and keeping up to date (losing irreplaceable knowledge and experience), focusing on its own market while the rest of the world ran straight past them and now they have to play catchup, etc.

A little harsh, but come on, America used to produce most of the worlds products. Now they cannot compete in general. There are reasons behind this, and it shouldn't be blamed on one thing like automation, cheap overseas labour or the "specture" of AI.

No one country stays on top for ever regardless of the choices and decisions they make (England ruled most of the world once) so people need to take ownership and adapt, not complain that things aren't as easy as they used to be or that change is looming and they don't like it.

Tax structure has a huge role in this, and don't forget it was America that went into Japan, S. Korea, Hong Kong, and China (not exactly the same as the Marshall Plan) and put factories down on the ground there, due to cheap labor, lax regulations with loopholes, and lower taxation by virtue of being abroad.

Later on, the watch changed, the holdings sold off to native holders, or other corporate holdings within each country led to a large shift in manufacturing.

With the right stimulus, money will flow back into the United States, in spite of your comments, and the world economy will be much better off for it.

We all need to spend less time raining on the other's parade, and more time joining the parade. We Americans really are for making all countries flourish, not just our own.

Perhaps we should start referring to robot/AI labor as "othersourcing".

Pete and Repete were standing on a bridge. Pete fell off, who was left?

Tell that joke to an AI, and watch "it" melt down.

I went to the 1964 World's Fair where IBM had quite a presence. One of the demos was Optical Character Recognition; I still remember thinking, "How can a computer figure out the nature of my character by looking at me??"

Another demo was language translation; you typed in a phrase on a KSR-35 and the 1620 typed out the translation. I typed in "out of sight, out of mind", then took the translation and asked the machine to translate the foreign phrase back to English. The answer came back "blind and insane".

It's a brave new world out there...

Lovely results.

Heh, you were playing 'Babelfish babble' before there WAS a Bablefish site.

Of course, that old site is gone, and we use Google or Yahoo Translate to get our funny results, such as "The spirit is willing, but the flesh is weak" into Russian and back ==> "The vodka is good, but the meat is putrid."

Like Spock telling the computer to compute Pi to the last decimal.

Repete was a moron. He didn't fall off because he was a little moron.

Its going to happen, not if but when. Mankind will gradually fade away dying off due to being useless.

Then the remaining few poor examples of mankind can worship our true overlords.

Apparently, you never made it to the end of the movie.

We will fight, because we have our rights, to Life, Liberty, and the Pursuit of Happiness.

It makes me unhappy to think I could be Hireling or Slave to a machine. Oh snap, here I am serving the keyboard. Never mind that for the moment.

The Hireling or Slave part comes from the 3rd verse of The American National Anthem.

Do not touch my rights, my flag, my anthem, or my property, or it will be war.

British political philosopher John Locke wrote originally that governments are instituted to secure people's rights to "life, liberty, and property."

Jefferson substituted "happiness" when he wrote the Declaration of Independence in 1776.

I suggest that today, "property" is more precious than "happiness" to many people.

So it would appear, lamentably. Who wants to be the king of a vast domain, and not rule his own heart. Thanks, Lyn, good point.

I think that "happiness" is more universal than "property". "Property" is just one manifestation of "happiness". Just like "value" is more universal than "money". "Money" is just one manifestation of "value". Etc.... (meaning/logic; purpose/reason or reason/purpose, depending on which came first, the chicken or the egg).

My view is perhaps more parochial in that I am only considering the USA and the never ending thirst for evermore wealth/property exhibited by most truly wealthy people and institutions.

I do not see AI as leveling that playing field, at all.

Yeah, I get it. Some pursue property/wealth to the detriment of their larger happiness. But, sometimes people in general get so busy, or in a routine, that the life they started on takes on a life of its own, and is no longer enjoyed/valued by the very people living that life. You know, the old "your property owns you" thing. Sometimes it sneaks up on you while you're not looking.

"Sometimes it sneaks up on you while you're not looking."

Just like getting old, a subject with which I am becoming very familiar. It sneaks up on you one day at a time and when you wake up and look in the mirror each morning, you look exactly the same as you did yesterday.

[Edit]

Well that response below ⇓ brings a furrow to my brow.

I don't know about that last statement. I can't possibly look as bad in the mirror today as I did Wednesday. We had a memorial service for one of my nieces, and it really saddened me that last few years to watch her waste away to nothing, almost a relief now her suffering is done.

The foggy days are upon us here, and we (finally) caught the first frost of the season yesterday, and today has been fog again, and gloomy clouds expected all day.

I want to go home and sleep.

I know what you mean and what it feels like. Getting old is not for the faint hearted, but there are enough rewards to make it well worth while.

I like foggy days as they give one some good private space to reflect. I learned something a few years ago from a negotiation instructor. If it's not going to literally kill you in the next 300 seconds, it's probably not worth worrying over at that moment. Occasionally I will be going somewhere or doing something, get to a certain location and just feel miserable, but then thinking back on that lesson, I realize that it was the history leading up to that moment in time that makes me feel the way I do and that I can feel any way I want to. Sometimes that helps.

Get your sleep and feel better.

Thanks buddy. Today is sort of like that old whaler song: "Rollin' down to old Maui".

I challenge any AI to come to terms with such sentiments.

It is warm fuzzy time!

I feel happy for you.

At the rate things has been taking place lately, better days are welcome.

When it comes down to an economic POV a robot will always win. White or blue collar workers will be replaced. AGAIN, not if but when.

The ARPI (Autonomous Robotic Pipe Inspection) system that we will be demonstrating after the first of the year can make > 8K measurements on a 40' long pipe in 2.5 minutes. It encompasses all required dimensional requirements per API 5L. That is impossible for a human inspector or inspectors to accomplish.

Yes I am sorry for the loss of jobs in inspection in the pipe industry but it will save the producers of the pipe $. As well as providing the insurance carriers greater reliability and reduced cost to the buyers as the longevity of the line can be greater as most of the critical dimension measurements are flat out not being done because there is no efficient method of doing the measurements.

Who will be analyzing all this data the ARPI spits out? A computer program? Can you override the automatic data inspection, and review further in any section of pipe you deem necessary?

Who will be analyzing all this data the ARPI spits out? A computer program?

Yes

Can you override the automatic data inspection, and review further in any section of pipe you deem necessary

Yes.

However any anomalies detected will be returned to in the area of the pipe where detected an a tighter inspection will occur.

This is API recommended method for determining out of round of the middle of a pipe.

I have been in the industry for > 50 years, never seen it done.

The time and manpower to accomplish this on every pipe at 24" increments on a 40' pipe is not cost effective for a human inspector.

If "determining the out of round of the middle of a pipe", a 40' pipe shouldn't require a measurement ever 24". That would be an order of magnitude greater than necessary.

No. API states at any point along the length of the pipe. I have seen OUT RND within 24" lengths.

The whole reason for the tolerance is to prevent turbulent flows which causes preferential corrosion.

It goes beyond just OUT RND as pipe can swell at sections along the length due to weak spots during hydro test and over expansion during process. We have incorporated two methods for detection. The second is an output gyro (accurate to 2 sec./arc) which also measures pipe bow.

At present all that is being done is random PI taping done on every 10th pipe , this can detect swells IF the inspector is in the right spot on the swell. Not OUT RND.

The ends of the pipe are measured for OUT RND but that is primarily for fit up in the field. The angular deviation of the bevel cut as related to the central axis of pipe is also not being measured, but is a requirement.

We have incorporated a laser reflection method which can detect the above.

We are not trying to rain on anybody's parade but the present requirements for pipe dimensions can not be done in a cost effective manner by a human inspector/inspectors.

Initially the acc/rej status will still be done by 3rd party(white collar) based on the scan reports but I believe over time that this can change.

This has not been an easy task and we have been working on this concept for > 5 years. Budget cutting, deaths in the family( had my father, brother, mother die during this development) lay offs, health problems, and change in management have slowed down development.

But I believe in what we are doing and nothing can stop us unless its our own death.

God speed to the completion of the project, and long life, old friend.

I believe you. I also think you have a truly smart machine there.

This sounds like the perfect case for an automated system to do the inspections while traveling down the inside of the pipe⁽¹⁾. The Pipe Inspection Gauge⁽²⁾ could take measurements as it traveled, providing an accurate picture at almost any linear resolution desired, even when product is flowing down the pipe with it.

Notes:

1. I'm going to assume that if the inside is not OUT RND, that the outside is also not OUT RND. I'm also assuming that if there *IS* a flaw on the outside that takes the pipe OUT RND without affecting the inside at the moment, it is only a matter of time before the inside starts showing signs, and this system could potentially spot the flaws before they become large enough to break up the laminar flow within the pipe, allowing repair work to happen BEFORE the flaw impacts production.
2. From an old (now removed) exhibit on petroleum at the Museum of Science and Industry in Chicago, I learned about 'Pigs' and the backronym developed from the name once they evolved from "industrial strength beach balls" into sophisticated sensor drones.

They call those pipe pigs. They are not used during fabrication of a pipeline, only after construction is completed, and it is the cleaning first done to make sure line is ready for service. Periodic pig runs are made to test the pipe integrity, and scour off any undesirable crap on the inside diameter of the pipe.

That depends on the book keeping system. If the amortized cost of the robot is rolled into the operating expense, rather than capital budget, then human labor can still be competitive.

I ran into a situation at one manufacturer where they injection molded two parts, moved them to inventory by forklift, transacted them out of inventory and moved them by forklift to assembly, put the two pieces together, transacted the subassembly into inventory and forklifted the subassembly to the warehouse, transacted the parts out of inventory and forklifted them to assembly where they were thrown into a final assembly box, transacted into inventory and forklifted back to the warehouse. The parts travelled about 4 miles to get from the injection molding machine to the final box and several inventory system transactions, not to mention intermediate subassembly part numbers based on level of assembly. I asked why they didn't assemble them just off the injection molders and get rid of the bulk of the handling and computer transactions. I was told that it wasn't cost effective.

Computer transactions and computer labor was absorbed by administrative overhead as were forklifting expenses, warehousing and inventory handling and were free of charge to the production department. The production department didn't have any incentives to reduce administrative overhead expenses.

A lot of the issue is with the accounting system.

Sounds like a money maker if they just kept the same two parts moving around in a great circle from one department to another, but the shakey part is getting them back to square one, for another go around.

I expect someone should have some 'splainin' to do at some point to a Congressional oversight committee, or some such, as surely this has to be government contract.

Nope. Consumer string trimmers. "weed whackers".

"I asked why they didn't assemble them just off the injection molders and get rid of the bulk of the handling and computer transactions. I was told that it wasn't cost effective."

Sounds like there's a lot of back-and-forth transport of the parts. Perhaps they should meet my friend, Henry F.

Stupid auto-complete, I meant THIS Henry F.

They are much like the endangered Damogran Frond Crested Eagle that covers the eggs in it's nest with three feet of paper mache. They have heard of survival of the fittest, but don't believe in it.

I will bite. Where do they get all this papier mache'? From France?

I knew there was a hitch to that. You sure went a long way to get that joke in.,

The emerging AI & IOT somehow remind me of the old movie "Fahrenheit 451". I think the driving mechanism for evolving AI is cost reduction & profit optimisation, hence humanistic aspects of its impacts on future generations is not considered within the equation. World uncontrolled population increase, global warming, draught & dehydration, migrations disputes & cultural differences, fast urbanisation & automation culture are pursuing corporations to invest in AI rather than human resourcing for future. The problem is that Future Regulations on demarcation between human resource versus AI- Robotics is likely to be very grey and subject to interpretations of the investment entrepreneurs behind the AI evolution. Industry investment providers does not find significant profit yield in challenging subjects such as Cancer Cure, Global Warming remedies , etc. Hence high profit margin areas like AI automation technology for e.g driving, labour work, nursing, farming, manufacturing, service delivery, food supply, etc. have priority and it is only a matter of time before AI will take over not only blue colour jobs but also white colour ones like banking, trading, surgeries, lawyers offices, defence systems etc. The 1% of society who owns 90% of assets worldwide is unlikely to care about white colour job loss, humans sitting home without human to human communication, gradual loss of human emotions and perhaps new emerging mental syndromes which might attract more research investments by pharma industry. Sorry for being realistic rather than optimistic !

As scientists, I hope you don't dismiss this as Mumbo-jumbo. I think this is more of a philosophical question. Closely related to another familiar question is, "Can Man create an intelligence that even He cannot out-think?", which needs to be asked. I would like to submit the basic rules of Existentialism, which includes infinity. Oh, wait.... How can anything include infinity?? Doesn't infinity include everything else? Ready for it? 1). No system can fit into one of its own sub-systems. 2). If you want a system to include all infinities, then you need a bigger system. 3). If you ever encounter something that threatens your existence, go to rule #1. These three rules describe a dual open/closed system, that is both finite and infinite. How well you can function "within" it determines your chances of existence. A violation of these rules results in a value outside of existence.

Yes. And as the machines continue to evolve and displace human workers, what is to become of the workers? Yes, a lucky few will have the ability to collaborate with the machines, and we will need human programmers and robot repairmen, at least until the machines can do those tasks better and cheaper. But long before then, we will have millions of obsolete workers. Can we retrain all of them to some other occupation that still needs humans? Show business? Writers, musicians, dancers, poets? The future for much of humanity looks to be empty, without purpose or a paycheck.

I have an unwritten musical about the subject. Art becomes the only "legitimate" form of currency. And the machine judges and pays based on criteria. People become highly stylized stereotypes living well... If unwilling to contribute with "passion" .. exile of course.

No.. I'm not writing musicals for God's sake.. my clone's I don't have should be busy at work doing so along with all the other things I can't be bothered with,

I spent part of the day watching my winter grass grow, the birds play and eat, smoking a fine Cuban cigar while sitting on my park bench in my front yard.

I don't need no stinking AI!

There you go! You alone are holding back this deluge of so-called "progress".

Being that decimate means to reduce by 1/10th then; yes, AI will likely decimate white collar jobs.

...and it is very likely that workers displaced from maturing industries will shift to previously unknown industries.

That always is disconcerting because, 'how can one have faith in the idea that industries that don't yet exist will absorb all the job losses that can clearly be seen heading this way?'.

So there will continue to be predictions of innovation bringing the demise of the economy. This has not happened over more than the short term. It has always done the opposite over more than a very limited period.

Until there is good evidence that this time really will be different than all the other times that were supposed to be different, I will continue to believe things will behave as they have; that innovation may cause disruption in the short term, but will have positive effects on prosperity over longer terms.

Not if this will happen in the next 50 years, it will happen. Some of us saw this coming and had the good sense to get old and retire before it happened to us. Now all we have to figure out is how to keep drawing our pensions and healthcare when there are no employed people contributing to the pot. I have spent the last 40 years automating people out of jobs. There are more people employed now than when I started, so either I am a total failure or new jobs are being found to replace the ones I eliminated.

Process industries have already eliminated most of the manual jobs, manufacturing industries are well on the way and the next big target is agriculture. So the trend in blue collar and unskilled jobs is set to continue over the next 50 years.

The change is that new AI programs can learn by experience. Software already exists to check mammograms for breast cancer. Is a medical technician white collar or blue collar? Borderline I would say. It does in 10 seconds what the human eye takes 15 minutes to do and is much more accurate in identifying cancerous cells at an early stage. It is the same technology that will be applied to drive cars and trucks. In the case of cancer the program it was given a few million old mammograms and asked which ones turned out to develop cancer. It learned without risk to any current patients. If cameras were fitted to the front of cars and the video footage of 500 million driving miles was given to the AI programs so that they could learn in the same safe way I might be in favour but we all know that will not happen. It didn't happen when automated trading computers sent the stock market into a nose dive a few years back but those trading computers still exist in essentially the same form as back then. I really hate to say this but maybe we still need bankers and fund managers ...yet. Part of the new jobs for old jobs solution is to employ people to monitor and regulate what the AI programs are doing.

I suspect that rather than paying people to sit idle they will be paid to do jobs that are not at present judged to be economic, like bringing low grade land into cultivation, cleaning up the environment or building homes for the homeless.

"... Software already exists to check mammograms for breast cancer. Is a medical technician white collar or blue collar? Borderline I would say. It does in 10 seconds what the human eye takes 15 minutes to do and is much more accurate in identifying cancerous cells at an early stage. ..."

.

Definitely faster, but I haven't seen studies indicating AI review is better than a second human review.

.

https://www.google.com/url?sa=t&source=web&rct=j&url=https://www.ncbi.nlm.nih.gov/pubmed/17954651&ved=0ahUKEwiwyuOTxLHXAhXqg1QKHTBZD1wQFgggMAA&usg=AOvVaw19Cnw0HDF6Z1PZ09PFcnqThttps://www.google.com/url?sa=t&source=web&rct=j&url=https://www.ncbi.nlm.nih.gov/pubmed/17954651&ved=0ahUKEwiwyuOTxLHXAhXqg1QKHTBZD1wQFgggMAA&usg=AOvVaw19Cnw0HDF6Z1PZ09PFcnqT

Information was taken from a BBC television documentary aired in the UK about four months ago on various aspects of AI. It referred to a US study where ten mammogram scanning technicians with the best statistical ratios in USA for spotting tumours : false positives had their averages compared to the AI program. The program scored better on both identifying tumours and had less false positives. The documentary did not give full details of where in the US the study was conducted or by whom but quick search on line suggests it may be the research done by the Houston Methodist Research Institute in Texas. There are many similar research studies all across the US and all showing similar results especially in skin and bowel cancers.

I'd much rather have a lower false negative than lower false positive, especially when multiple reads are standard.

The study I linked above doesn't suggest better detection on whole. Perhaps you can link one or the several studies to which you made reference?

This reminds me of the human vs computer in a chess game argument. It is now impossible to win a game of chess against the well written program that can analyze all of the possible moves. But not all programs are created equally. To do so, requires a complete understanding of the game which is based on rules. AI happens to be based on rules as well. So, in order for AI to be able to do everything better, it will have to build upon the initial rules that are provided by some human understanding. Then after it learns, by trial and error, it will eventually become better at humans with some given task. That does not equate to the ability to shift between tasks and become proficient at all tasks. All of the rules and various tasks take a lot of memory and coming up with new methods for completing the same tasks are not likely to be a big part of the overall system. Consequently, I'd think we are still a long way from being threatened, except for science fiction which has no problem with violating physics and possibly getting stuck with circular reasoning.

There is also the problem with eventual component failure. Dealing with all of the possibilities of such a failure will take a very substantial amount of diagnostic capability. In order to have infinite capability, it will need infinite memory. We have problems with very large numbers. Because of concept of infinity, I don't think we have that much to worry about.

While my computer can beat me in a game of chess, it just didn't do so good at kickball.