Hacker News new | comments | ask | show | jobs | submit login
Ironies of Automation (1983) [pdf] (ncsu.edu)
80 points by kiyanwang 7 days ago | hide | past | web | favorite | 27 comments





Some newer papers looking back at this and it's continuing relevance:

2017: Ironies of Automation: Still Unresolved After All These Years https://doi.org/10.1109/THMS.2017.2732506 (couldn't find an openly accessible link :/)

2012: The ironies of automation... still going strong at 30? - http://johnrooksby.org/papers/ECCE2012_baxter_ironies.pdf



I used to work in ETL sometime back on a popular tool daily.

It used to take a LOT of time for doing code reviews.

30min per object. Decent project has 300 such objects.

Turns out the tool can export in XML so I wrote a Python script to do the review.

What happened?

I could review basic manual things like coding standards in 15milli seconds irrespective of the object number.

That's automation. It is usable in other projects , and nobody other than my team used it ever for review. Why? Because it requires efforts to understand & a basic understanding of XML &Python and they don't have it because it's not their field. I am a nerd so I kept exploring.

So the script is now ancient history.

And coming to my point. These people say automation is demise of everything. AI is taking over all jobs. Won't happen for a long time.

For starters, scientists in the cold war had said "We will have live voice translate from Russian to English in 5yrs"

Cold war has ended for how many years now? And in 2010s MS finally was able to bring this tech. Imagine the time gap.

secondly, if such an 'automation' thing is sold by a company, then who will create & sell it? And will that firm have 1000000000 of such small scripts for automation? Highly unlikely. Whatever automation I did in past five years was something that relieved my work load.

It is highly unlikely that such automation packages will be bought by companies and the much harder thing is for automation companies to sell such tools.

Companies like mine have spent millions to buy that tool, will they really spend more money on buying a bunch of automation tools?

Note that the core aspect of the tool's usage can't be automated . You still need a living thinking human being to use it, and make objects using the non existing requirement documents. Scripts need 100% perfect everything which is not possible in real life


Yeah, unfortunately it's fairly easy to rebut this paper. I only had to read the abstract and the first paragraph. The last line of the first paragraph:

> ... the more advanced a control system is, so the more crucial may be the contribution of the human operator.

This is absolutely true. The human operator working with an automation system needs to be way more skilled, more educated, and smarter compared to a human working as a cog in a sweatshop.

But this still doesn't support the thesis that it expands the problems associated with human operator, for the very simple reason:

You are replacing 1000 low-skilled sweatshop workers with 5-10 (or less) highly-skilled engineers and experienced technicians. And at the same time increasing the throughput of your production system by an order of magnitude or more. It's a complete win.

And I didn't even need to invoke the role and/or potential of modern AI (i.e., one based on ML/DL/DRL/CV etc, etc), which is going to take the automation to the next level.

Highly recommended watch: CGPGrey 'Humans Need Not Apply'

https://www.youtube.com/watch?v=7Pq-S557XQU

(in the video just look at the human labor chart regarding agriculture, at 15 second mark, and you'll understand what I'm saying).


> this still doesn't support the thesis that it expands the problems associated with human operator

In some circumstances it might. Look at Air France Flight 447 and the role of the co-pilot in stalling the aircraft.

"Neither weather nor malfunction doomed AF447, nor a complex chain of error, but a simple but persistent mistake on the part of one of the pilots" [0]

An aircraft like that has so many highly automated systems it is impossible for the human operators to fully understand them.

[0] https://www.popularmechanics.com/flight/a3115/what-really-ha...


I'm not sure what's your point: the human was in charge because the automated system failed. Yes, one pilot made an error which crashed the plane but IMHO the designers of the "airplane UI" are guilty too: one pilot is pushing, the other is pulling: why isn't there some big warning displayed? Both pilots are supposed to do the same action..

> An aircraft like that has so many highly automated systems it is impossible for the human operators to fully understand them.

The aircraft was stalling because the pilot kept the flight stick pulled all the way back. Nothing automated or complex about it - a wood-and-canvas biplane would have had exactly the same problem.


If I remember correctly: Due to sensor issues the plane dropped out of autopilot and switched into a different control mode ("Alternate law").

The pilot had to suddenly take over and had a wrong mental model of the state and configuration the plane was in, and likely didn't realize the consequences of the different control mode, among which is that it doesn't provide the same level of automated stall protection as the normal case. Since automation normally works, the pilot wasn't trained or experienced in flying in this mode.

While in the end the operator made the wrong decision, these automation ironies fit very well contribute to why he made those decisions or wasn't prepared to make the right ones.


> While in the end the operator made the wrong decision, these automation ironies fit very well contribute to why he made those decisions or wasn't prepared to make the right ones.

Not at all convinced. Hauling back on the stick and just keeping it there would never have been the right decision under any circumstances. At the same time it's a reasonably common panic reaction, and was long before any kind of cockpit automation.

A pilot who had been in tense situations before would have been more likely to make the right decision, sure, and modern automated systems may well mean that the first truly deadly situation occurs later in a pilot's career than it might otherwise. But by definition there's no safe way to test how a pilot will handle genuine danger.


>> Hauling back on the stick and just keeping it there would never have been the right decision under any circumstances.

From the BEA report [0]:

"The horizontal bar then indicated a slight nose-up order compared with the aeroplane symbol."

and

"Nevertheless, the PF was also confronted with the stall warning, which conflicted with his impression of an overspeed. The transient activations of the warning after the autopilot disconnection may have caused the crew to doubt its credibility.

Furthermore, the fact that the flight director was advising a nose-up attitude may have confirmed the PF’s belief that the stall warning was not relevant"

(my highlight)

[0] https://www.bea.aero/docspa/2009/f-cp090601.en/pdf/f-cp09060...


> You are replacing 1000 low-skilled sweatshop workers

> It's a complete win

Complete win for who exactly ?

How do you reconcile automation and the current economy of consumption ? As far as I can tell they don't work hand in hand very well. These 1000 workers will still want their part of the cake, to do so you'll have to get them new jobs. I wouldn't be surprised if every time we go through a cycle of automation we also go one layer deeper in the "bullshit jobs" creation.

It's all fun and games until you have to take into consideration the soci-economical aspect of automation.

"Automation, the most advanced sector of modern industry as well as the model which perfectly sums up its practice, drives the commodity world toward the following contradiction: the technical equipment which objectively eliminates labor must at the same time preserve labor as a commodity and as the only source of the commodity. If the social labor (time) engaged by the society is not to diminish because of automation (or any other less extreme form of increasing the productivity of labor), then new jobs have to be created. Services, the tertiary sector, swell the ranks of the army of distribution and are a eulogy to the current commodities; the additional forces which are mobilized just happen to be suitable for the organization of redundant labor required by the artificial needs for such commodities." - The society of spectacle.


I'm sorry. Allow me to qualify my claim and restate it:

It's a complete win for everyone except for the morons in charge, and their so-called expert economic advisers, who think the functioning of the society has to depend upon everyone needing to work 9 to 5, Monday through Friday in order to have food, shelter, clothing, health, education, justice. (Even if it means creating more and more bullshit jobs).

If your pathetic obsolete-jobs-are-always-replaced-with-new-kinds-of-jobs economists cannot update their worldview and their economic policy in light of technological progress, that's not a problem with technological progress. (Only a handful of economists, Erik Brynjolfsson, Andrew McAfee, Robin Hanson, to name a few, have a realistic view of what's coming in terms of socioecononmic impact of tech progress).


I genuinely can't tell if you're passive aggressive and referencing the author of the quote as "pathetic economists", I'll give you the benefit of the doubt.

Either way, look at the past 40 years of so of technical progress, it's mostly driven by personal power/wealth quests and has nothing to do with bettering society. People are still working as much as before (actually, society as a whole is working more), but instead of working for survival they're doing it for an enhanced form of survival which mostly consist of buying (gadgets) and consuming ("culture").

The need of buying an iphone n+1 keeps you going day in, day out while the new netflix show distracts you just enough to keep you sane until your next shift. IG / FB reminds you that "if you keep going you'll make it, eventually, maybe. Oh and btw don't forget to buy these new shoes that _influencer_xyz_ talked about". Amazon, Facebook, Apple, Netflix, Instagram, are enabler / distributer of mindless consumption, be it physical or "cultural". They're not bettering the world in any way.

Meanwhile people in the US, one of the most powerful country/economy don't have a proper public health system and have horrendous working conditions. Keep in mind that as tech workers we're relatively immune to most of these things, we're generally above the daily struggle of the _true_ working class (flexible work hours, chill offices, good salaries, job safety)

The entire world economy is based on work<>consume, if you remove half of the equation thinking that the system is going to self-regulate I doubt you fully grasp the issue.

But hey, everything that's not full blow neoliberalism has to be communism, right.


> I genuinely can't tell if you're passive aggressive and referencing the author of the quote as "pathetic economists", I'll give you the benefit of the doubt.

I'm not. I'm referring to today's mainstream economists who have their heads in the sand when it comes to analyzing, assessing, diagnosing, and advising on the effects of tech progress.

> Either way, look at the past 40 years of so of technical progress, it's mostly driven by personal power/wealth quests and has nothing to do with bettering society. People are still working as much as before (actually, society as a whole is working more), but instead of working for survival they're doing it for an enhanced form of survival which mostly consist of buying (gadgets) and consuming ("culture").

That's a problem of politicians and economists, not technology. You should understand something about technology. It's a force of nature. There's a factor of inevitability in play here.

It is not the case that every year a bunch of greedy, cigar smoking, capitalist billionaires gather in a lavish, dimly lit, oak-wood room and decide how they're going to advance technology and screw over the population.

That's not how it works.

Technology and innovation comes from the garages, from the workshops, from the basements, whether they belong to big corporations or to the parents of a bunch of geeks. It doesn't matter. Technology happens because people are creative. And they're constantly thinking, and they're coming up with new ideas. What are you going to do to stop this deluge of creativity and innovation? Have some kind of thought control. Ban innovative technological ideas? Even if you are able to do it, what makes you think it won't come from outside. If US doesn't innovate, China will. If Chine doesn't, Russia will. If Russia doesn't, the rest of the world will.

When the time for agricultural revolution had come, there wasn't much hunter gatherers could do about it.

When the time for industrial revolution had arrived, there wasn't much the kings, queens, peasants, and luddites smashing the looms could do about it.

When it's time for artificially intelligent automation and complete elimination of human labor, there won't be much you or I or anyone else will be able to do about it.

The best we can do is adapt. And we won't adapt if we don't understand the world around us for what it is.


>Complete win for who exactly ?

The sweat shop workers, the company, the customers.

Sweat shop workers are losing their ability to do mindless labor, next up, they will be using their brains to make low skill decisions where automation cannot. Think Amazon Turk.

Customers get lower cost, likely higher quality products.

Company gets more reliable workforce.

>It's all fun and games until you have to take into consideration the soci-economical aspect of automation.

This has already happened in history. You COULD make it illegal, but then your countries technology will be out performed by another country.

---

Its easy to disagree, its harder to come up with your own solutions.


What would be the end goal of such a system ?

Looks to me like a rat race.


Higher standards of living. Less starvation, better environment, etc...

That's an honourable goal but automation was supposed to deliver* us in the 60s, 70s, 80s, 90s, 00s etc ....

It never happened and will never happen with the current world economy. These things are mutually exclusive.

https://hbr.org/2016/02/todays-automation-anxiety-was-alive-...

https://scholar.lib.vt.edu/ejournals/SPT/v7n3/hong.html

https://timeline.com/robots-have-been-about-to-take-all-the-...

* or destroy, depending on who you listen to


Only if there is a taxation function that takes money from the top and distributes it, which gets derided as full on communism on HN. One of the big issues with automation is it tends to concentrate wealth with a few big players

Never underestimate the individual and organizational ability to cut corners. If your system depends on supervision of a highly-skilled and highly-trained human, better be sure it has a solid process around it that enforces the supervising human is actually skilled and trained.

It'd be helpful if you argued against the contents of the paper and not some idea you got from the first few lines of what you think it is about. You seem to be mostly talking past it, and it contains counters to your "very simple reason", even if we assume all systems fit that example (which they clearly don't, e.g. which 1000 unskilled workers does a plane autopilot replace?)

Agree. Which is also I think a key argument against "AI will replace everything". You will still need operators to create and maintain the algorithms, and these will have to be pretty skilled and costly. There will be many applications where the cost is simply higher than a few blue collar workers (or white collars!).

In essence it's not new, software has replaced many tasks, but there is a huge amount of manual tasks that could already be automated righ now with non-AI code but where it is simply uneconomical to maintain a highly skilled dev team.

Also there are many domains which are so obscure and complex that really only specialists can write software for it. For some professions it does happen (engineering, finance) some others it doesn't (accounting, law).


> Which is also I think a key argument against "AI will replace everything".

Thank you for agreeing with me, but I'm sorry I did not make an argument against "AI will replace everything".

When 990 of the 1000 previously fully employed workers are now all of a sudden put out of work, yes "AI/automation does replace pretty much everything". It creates a massive disruption in society.

Again, in the video I linked, during the whole of 20th century tens of millions of people in the US were put out of work, and only one new kind of job was created (computer programmers) that resulted in more than a million new jobs. Even then it's 30th in rank (in terms of total jobs), after 29 types of jobs that were there even in the 19th century.

I don't understand how you can see a comment that links to CGPGrey's 'Humans Need Not Apply' and conclude that the commenter is making a point against 'AI will replace everything'


Should add (1983).

Is there some LaTeX library to have your PDF show up like its been through the printer/scanner at least twice?

On a serious note, why do academics do this? It doesn't help me with readability.


Is it really that surprising that they don't have a crisp source PDF for a paper from 1983?

My mistake. Didn't notice that.



Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: