back to article Will LLMs take your job? Only if you let them

Let's face it. You're a smart cookie. You work in tech, you create or manage or help others in some of the most complex machinery in human history. So is it true that LLMs are gunning for your job, as they are for those of others working in similarly arcane areas of knowledge work, like the law, finance, medicine and the …

  1. Filippo Silver badge

    A good article. It's good to remember that any jobs eliminated by automation pale in comparison to those destroyed by mismanagement and poor leadership.

    Mismanagement and poor leadership, part of which includes rejecting, misunderstanding or misapplying automation.

    In the case of LLMs, though, I wouldn't fault any business leaders for being highly suspicious right now. It's a tech worth keeping an eye on, but it's not at all clear it will actually be that useful in practice.

    1. Michael Wojcik Silver badge

      I agree with the last point. Certainly it will vary quite heavily by the user. I've read a number of recent research papers about LLMs, and reasonably well-informed and intelligent articles from a variety of perspectives, and I've yet to see an LLM do anything better than I can, in my areas. Or, indeed, do anything that would make one worth my time.

      Natural language is a poor search interface and a poor user interface for the vast majority of use cases. LLM code completion is a trap: learned helplessness coupled with a failure to understand the proposed solution, and a concomitant one to explore the solution space and potentially learn something. Leaning on an LLM at a minimum costs the user the opportunities for skill development and serendipitous discovery.

      This is always true of information technology, of course. The printing press cost a number of scribes the opportunity to incidentally learn things from the books they copied. But the trade-off for the printing press was clearly profitable: a small opportunity cost to a few people, which could be recouped by using some of their returned time to simply read, in exchange for a huge benefit to many people. So far the demonstrated "benefits" of LLMs are much, much less, and the cost to users much higher.

  2. b0llchit Silver badge
    Facepalm

    Overly optimistic

    ...prompt engineering...

    That is a euphemism for prompt slave.

    They [LLMs] deal in knowledge but have no idea about truth.

    The LLMs have no concept of truth. LLMs have no concept of anything whatsoever. LLMs cannot conceptualize and therefore can not reason.

    Replacing people with LLMs is like telling bacteria to come up with the theory of gravity and space-time. You can do so, but it will fail to take you to the sought after results.

    1. Doctor Syntax Silver badge

      Re: Overly optimistic

      Aa far as I can see they're the ultimate GIGO machines with an added twist. If you put in something that's no garbage you might or not get somebody else's garbage out and if you knew enough to tell the difference you wouldn't need the machine anyway.

      1. Yet Another Anonymous coward Silver badge

        Re: Overly optimistic

        All the programmers are going to be made redundant by compilers which allow anybody to write simple clear prompts and have the machine automatically translate that into software.

        Now that we have COBOL, management can instruct the machines directly

        1. A.P. Veening Silver badge

          Re: Overly optimistic

          Now that we have COBOL, management can instruct the machines directly

          The same was once said about Report Program Generator, TLA'd to RPG, by now the original name is scrapped and it is just the computer language RPG.

          1. Michael Wojcik Silver badge

            Re: Overly optimistic

            And 4GLs, etc. I expect the previous poster and most other Reg readers are well familiar with this particular aspiration in the industry.

        2. An_Old_Dog Silver badge

          Re: Overly optimistic

          One thing which the widespread use-experiences of COBOL taught us was, "Managers (who are not programmers shoved up into managment as a reward for good programming performance) do not read programs -- even if they are able to do so."

          1. Rich 11

            Re: Overly optimistic

            I've had managers who struggle to use full stops correctly in normal written English, so clearly they'd not get very far with COBOL.

            (Bloody hell, when was the last time I gave any thought to COBOL? Thirty years ago?)

    2. This post has been deleted by its author

  3. Anonymous Coward
    Anonymous Coward

    "If you're smart enough to do your job better with LLMs or smart enough to know why that's not happening, you're smart enough to keep it."

    Yeah but managers are dumb enough to just arbitrarily pick the staff they want to keep and which ones to bin, when AI starts shrinking workloads, it's the weirdos that get the bullet first, which in an industry full of indispensable weirdos is a bad thing. Managers and higher ups don't sack people that let them win at golf. Know what I mean?

    Doing a good job and making a higher ranking member of staff look good are not usually the same thing.

    As techies, we need to be finding ways to use AI and LLMs to replace these bellends before they try and replace us. Project management is a big juicy target at this stage. There is absolutely no reason we can't use AI to automate that. Same thing applies to CTOs. Why not make the "technical" part literally true...it can't be worse than the sheer number of human CTOs out there that have never deployed a single box in their life / written a line of code since 1987.

    I swear, if I spent less time "bringing them up to speed", I'd probably get a productivity jump in line with using an AI to assist me.

    As a person that has worked as a CTO, project manager and general management shadow on and off for many years (which is essentially the same sort of thing as prompting a beta version of an AI trained on a dataset built in the 80s), I'd dearly love it if I could be the shadow for an AI. I'd be doing the same job, but I'd get much better results.

    Quite a few of you out there will be working as a shadow and not know it yet...some pointers to help you figure it out.

    1) There is an important executive meeting coming to discuss some sort of technical strategy. You aren't invited, but each of the attendees comes to have a chat with you on the sly individually to ask your opinion on things that have very little to do with any of the projects you're working on.

    2) You hear the phrase "off the record" a lot.

    3) A weirdly high number of suggestions that you make "off the record" seem to get implemented, but never accredited to you.

    4) You're low ranking, but basically running the business.

    5) You know exactly what was discussed in a meeting without having been there and you start to control these meetings by providing one person with one set of opinions, mostly ones that you placed there to make the person look smart, but a couple that you agree with, and you arm another person in the meeting with counter arguments to cancel out of the fluff you gave the other person, thus controlling the outcome...without setting foot in the room.

    6) If these people are being too cagey with information and you can't quite execute a full plan, and you need more details...you engineer the information that you give these people (basically control the meeting to come to know particular conclusion) to lead to them requesting you join the meeting half way through at which point, you gather a little more information and tip the balance where you see fit.

    7) People start sweating when you try and book a 2 week holiday.

    I swear, I sometimes feel like the BlackAdder of tech...I've always got a young Baldrick that I'm training up and I'm always deploying cunning plans...it's never been limited to just one company either...I've managed this at several businesses. Mostly because the Baldricks tend to move on and eventually become managers, which opens the doors elsewhere for more cunning plans. They don't always work, but they are usually, so cunning you could brush your teeth with them.

    As for AI and how it would help in these situations, I'm not sure it would per se...but past a certain level in most tech firms, you'll find that it's already automated to a degree by someone like me.

    Remember kids, puppets are controlled by strings from above, but muppets are controlled with sticks from below. Just ask Jim Henson.

    1. An_Old_Dog Silver badge
      Devil

      Controlling Muppets

      Items numbers 5 and 6 look like BOFH/Evil Overlord territory.

      (Icon for BOFH/Evil Overlord)

  4. Anonymous Coward
    Anonymous Coward

    LLMs respond best to concise, well-defined, and well-constrained problems. If anybody working in the technology industry has a job solving these kinds of problems please raise your hand.

    1. Anonymous Coward
      Anonymous Coward

      Yup!!

      After 20 years, installing a server and watching the users gather around it is still like that scene from 2001 Space Odyssey.

  5. katrinab Silver badge
    Meh

    Take the finance sector as an example:

    Before computers, if you wanted to calculate the redemption yield of a bond, you needed a room full of humans to calculate it for you, and it would take several hours. A spreadsheet could do that in about a second.

    Did the number of people employed in finance fall in the 1980s compared with the 1970s and earlier?

    No. It increased. A lot.

    I don't think ChatGPT is going to be as transformative as Lotus 123 was back in the day, but even if it is, I'm not worried.

    Also, anyone thinking about investing in OpenAI. That could go about as well as an investment in Lotus or whoever made Multiplan.

    1. itzumee

      Multiplan was made by Microsoft

      https://en.wikipedia.org/wiki/Multiplan

    2. unimaginative

      A room full of people over several hours? That is hyperbole. It is just an IIR and people doing it by hand would do it by trial and error, starting from an intelligent estimate. It would not take that long - maybe 10 mins for a long maturity bond. A pocket calculator would speed that up dramatically too.

      Bond pricing also used simplified assumptions (like all years are the same length) to make calculations easier.

    3. Michael Wojcik Silver badge

      Counterpoint: Putting a PC on every middle-manager's desk destroyed the typing pool. Typist was a skilled trade, and replacing the typing pool with a PC and word-processing software cost those jobs, and it cost managers time because dictation to a skilled human is faster than hunt-and-peck typing into Microsoft Word or the like, and it reduced the quality of business prose because it was no longer trained professional writers producing the final copy.

      There have been a number of studies which suggest the "PC revolution" was actually fairly expensive in terms of productivity.

      On a similar note, giving spreadsheet software to bookkeepers and others who understood how to use paper spreadsheets was productive. Giving them to people with no idea how to use them correctly? Quite possibly not.

      A CACM article on the 20th anniversary of Powerpoint (which presumably was published around 16 years ago, but I'm not going to go look for it) noted that in the '80s, similar presentations were generally either B&W overhead transparencies1 or carefully-orchestrated multimedia presentations with synchronized slide projectors and tape decks that took many hours to create. Now Powerpoint Rangers generate zillions of fancy presentations every day with graphics! and animation! and mind-numbing stupidity! – which, yes, is a lot more output, but is it more value?

      And I recall a Byte article from many years ago (obviously) about the "Fat Bits" option in Mac Paint (or whatever it was called): the zoom function, basically. The author suggested that having a zoom function, and being able to do pixel-by-pixel editing, led to people wasting a vast amount of time fiddling with details that no audience member was likely to notice, and thus offered essentially no return on investment.

      Information technology has severe revenge effects, especially when it attracts a lot of attention2 and triggers obsessive behavior in users.

      1Or "foils", if you worked at IBM, the Land of Our Own Damn Nomenclature, Live With It.

      2One of the great ironies of the current LLM fervor is that it was touched off by a paper titled "Attention is All You Need". The use of "attention" as a term of art in transformer algorithms is an accidental gesture toward the greatest problem they currently cause.

      1. katrinab Silver badge
        Meh

        Typing a letter in Word or WordPerfect and pressing print isn't any quicker than typing it on a typewriter. That is true.

        The productivity improvements came from being able to edit the document and press print again. So you can have things like template letters where you just need to edit the relevant details, much quicker than typing the whole thing out again. Also, if you are doing negotiations on some big legal document / procedure manual / etc where there are multiple drafts before you agree on the final version, or you need to create a new version for whatever reason, then a word processor makes that process a lot quicker.

  6. Anonymous Coward
    Anonymous Coward

    In The Long Run.......

    ......technology has NEVER reduced the number of jobs in the economy.....or reduced the wealth in the economy.....

    In 1600 AD England had a population of around five million, and most of them were in agricultural jobs close to subsistence....

    In 2023 AD England has a population of around fifty million......and "subsistence" today means -- smartphone, 60 inch TV, a five year old car, and maybe some benefits paid by other citizens.......

    So....take a deep breath......technology (eventually) makes life better for everyone........

    1. Giles C Silver badge

      Re: In The Long Run.......

      Well that means I am under subsistence level then.

      Only got a 50” tv (8 years old) do have a smartphone and the newest car is 10 years old and the other two are 23 and 34 years old and don’t claim any benefits…..

      Oh well

      1. Stork Silver badge

        Re: In The Long Run.......

        Pay, cushy! Our newest car is 18 this month, our tv no more than 40”. And I have to be pool boy myself!

        1. Anonymous Coward
          Anonymous Coward

          Re: In The Long Run.......

          You youngsters have got it easy. I have a 15" portable that I watch through binoculars (because it is up the street in the security booth of a car park, and it only has one channel...the ticket machine) and my car is a horse!

          1. SonofRojBlake

            Re: In The Long Run.......

            There were undred and fifty of us livin in't shoebox in't middle o't road.

            1. Anonymous Coward
              Anonymous Coward

              Re: In The Long Run.......

              I bet your shoe box had 4 sides an a lid an all. You flash bastard.

              We ad to sleep wit pit ponies in't box an't old man made us shoe't ponies at 3 in't mornin.

              1. SonofRojBlake

                Re: In The Long Run.......

                I think you meant...

                "Cardboard box? You were lucky. We lived in a paper bag in a septic tank."

    2. unimaginative
      Unhappy

      Re: In The Long Run.......

      Subsistence in 1600 would have been a lot more than a few centuries earlier, and a whole lot better than "subsistence" a 1500 years early in the heavily slave based Roman economy where agricultural slaves would have had an absolute minimum "subsistence" - just enough to get the most useful work out of them.

      The problem is that technology makes life better, but there is no guarantee it will make life better for everyone. Consider the US is 1800. It was far in advance of any economy in 1600, but it still had a substantial population of agricultural slaves who lived on a similar minimum subsistence to Roman slaves.

      In the UK there was no slavery, but a substantial part of the population became a lot worse off after the enclosures of the agricultural revolution - the very thing that created economic growth.

      Even if you are right, "eventually" can be a very long time.

      1. Rich 11

        Re: In The Long Run.......

        In the UK there was no slavery

        There was slavery in the UK in 1800, but in these islands it was mostly confined to the role of personal servants to the wealthy. The vast majority of British slavery was offshored to the Caribbean, just as exploitative near-slavery conditions were imposed across East Africa, West Africa, India and SE Asia: sugar, coffee, tea, cotton, hemp, indigo and rubber all flowed back to feed the factories in the UK.

    3. martinusher Silver badge

      Re: In The Long Run.......

      >and "subsistence" today means -- smartphone, 60 inch TV, a five year old car, and maybe some benefits paid by other citizens.

      Not from what I've been hearing. Finding an affordable place to live and enough food to eat seems to be at the top of many peoples' agendas. Some kind of phone is essential for many jobs/lives even though they're rather expensive compared to the landlines of yesteryear, an oversized TV might be nice if the living accommodation wasn't so small, a car might be essential for a job (here in the US they're now averaging 12 years of age but I think the laws have been tweaked in the UK to ensure that older vehicles are scrapped even if they're serviceable) and as for living on the government's dime -- that's a lot of work for very little payback.

      They're obviously exceptions to the rule but the point I'm making is that a lot of people don't really understand poverty until it hits them square between the eyes -- they never think its going to happen to them until it does so are easy prey to the small voices that constantly tell them "its all the fault of the workshy/parasites/undeserving".

  7. riparian zone

    Pretotyping, anyone?

    Good article. Reminds me of the IBM post war experiment to determine the need for speech to text services in the 1950's I think it was. After inviting everyone on their address book to come to an demonstration, not everyone turned up. This number then shrunk after the demonstration, which was really an audio typist in another room producing text after executives spoke into a funnel. Of course the execs didn't know this. The number of companies wanting the service would not justify the expense, development and resulting cost of the tech to fulfill the purpose, so got shelved for many years. Seems that the need is sort of there, but now a company has to understand what it does and who does what to get rid of people before taking on the lucid dreaming language monster that magnifies the crap in crap out adage. it ain't likely, but more likely a successful AI take up will happen when companies' management structures understand what people do first?

  8. Doctor Syntax Silver badge

    "admitting that she's got no real idea what's going on"

    Degree in business studies, worked in marketing. It's almost like cause and effect.

  9. Yet Another Anonymous coward Silver badge

    Great quote

    The robots taking over creating poetry and paintings while the humans pick fruit and clean bathrooms isn't the future we expected

    1. xyz Silver badge

      Re: Great quote

      Stop it, you'll make Sue Ellen Braverman all wet.

    2. Filippo Silver badge

      Re: Great quote

      We spent a hundred years writing fiction about evil superintelligent AIs that are unbeatable at logical thought, but eventually get defeated by exploiting their inability to think creatively.

      It's really ironic that, in reality, we've figured out how to make computer programs that can create pretty decent paintings and fiction, but are utterly unreliable at logic and facts.

    3. Dinanziame Silver badge
      Trollface

      Re: Great quote

      Yeah, I read so many SciFi stories about people losing their jobs to a robot, but curiously not a single one of them was about a SciFi writer losing their job to a chatbot

  10. stiine Silver badge
    Unhappy

    "Expose students to the technology, not to replace teachers or encourage dependency, but to be skilled in critical analysis of what it does and how it does it."

    You have more faith in humanity and its educational systems than I. The ones I know can't do that sans technology.

  11. Paul Cooper

    I am an LLM!

    The abbreviation LLM has other, well-established uses - I am one (Licensed Lay Minister!). Can we please adhere to the convention of expanding initializations on first use and in titles?

    .

    1. Mike 137 Silver badge

      Re: I am an LLM!

      Also the postnominal for those with a Masters degree in law (very common). I don't think any of them would like to be confused with a glorified text autocomplete engine.

    2. Simon Rockman

      Re: I am an LLM!

      I'm rather pleased that we've grown out of people commenting on what else AI stands for.

  12. Anonymous Coward
    Anonymous Coward

    Large Language Model Deployment Fallout

    I know someone who played with a LLM. She works in HR for a very large microprocessor design and production company, and said, "This thing is coming for my job." (Note she is not involved in filtering resumes.)

    I think we'll see a lot of the wrong people being fired, a few companies going down in flames, and some of the other companies realizing they aren't using their people and tech effectively, that they need to reorganize which tasks are done in which jobs, which jobs need to be eliminated, and which jobs should be created.

    Whether those companies will do so effectively, or fall into rearranging deck chairs on the Titanic is anyone's guess.

  13. HMcG

    I'm old enough to remember when VisiCalc for the Apple II was going to put 90% of accountancy & finance workers out of a job.

    Then WordPerfect was going to put 90% of secretarial and administrative workers out of a job.

    And still unemployment hovers around the same level it did back in the 1970's.

  14. Mike 137 Silver badge

    "[LLMs] deal in knowledge but have no idea about truth"

    Rather then knowledge, they really deal in words but only in a marginally more sophisticated sense than a parrot does. They haven't the faintest concept of meaning (or, in fact any concept of anything at all). That's the fundamental distinction between the output of an LLM and human mentation, and it's not going away any time soon (if ever).

  15. martinusher Silver badge

    Complacency is a bad mindset

    Some would say that people who work in the airline industry -- for example -- also "work with some of the most complex machines and systems that humanity has ever devised". Despite this there are plenty of lower wage jobs to be had and there is constant downward pressure on wages with the products and procedures being constantly revised to squeeze more from fewer workers. This wasn't always the case -- its not that long ago that working in this industry was a prestige job with great wages and envious working conditions.

    So, yes, working with LLMs might garner prestige and a well paid job in the short term (so you'd obviously be dumb not to milk it for all its worth) but judging by the history of all other jobs its not going to last. What's going to happen is that after the honeymoon period there's going to be a shakeout with the workforce stratifying into a relative handful of highly paid professionals and an army of relatively low paid 'droids (assuming they're still employed). Unlike many other industries which do need a minimum level of hands on (humans being a whole lot more cost effective than humanoid robots for most applications) pure knowledge work is not only really easy to automate but also really easy to relocate.

  16. Brian Miller

    Mediocrity destroys more than AI

    A while back I left a position because, despite re-implementing an API for over 126 times faster performance, my manager's manager said, "We don't need the speed."

    While I was working with the old code, I liked to imagine that it had, in fact, been written by a LLM neural network. Data format was verified with regular expressions. Binary data was converted to a string via badly implemented functions. The string of zeros and ones was then subjected to bitwise operation implemented with if-then statements. Finally it was converted from a string to a number. There was a function named, "RandomHexPrependLastByte". I never quite figured out what that did. Yes, there was random data. The "blessed" code ran at 10 requests per second on an Amazon AWS C5 Large instance. Of course my manager lept to the defense of the "software engineer" in question.

    So yeah, I want to see LLM take over. It should take over and destroy the maximum number of jobs that can be possibly replaced by AI. I am absolutely confident that AI can't write code at my level. I am confident that all of the mediocre "software engineers" will be out of a job.

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like