Advertisement
Help Keep Boards Alive. Support us by going ad free today. See here: https://subscriptions.boards.ie/.
https://www.boards.ie/group/1878-subscribers-forum

Private Group for paid up members of Boards.ie. Join the club.
Hi all, please see this major site announcement: https://www.boards.ie/discussion/2058427594/boards-ie-2026

Will AI take your job?

1111214161723

Comments

  • Registered Users, Registered Users 2 Posts: 3,036 ✭✭✭rogber


    One of the more useful and instructive posts here. People ignore this stuff at their peril



  • Registered Users, Registered Users 2, Paid Member Posts: 8,827 ✭✭✭plodder


    I guess this journalist is in the camp where - "AI is just making me more productive … "

    ai.jpg

    “The opposite of 'good' is 'good intentions'”



  • Registered Users, Registered Users 2 Posts: 346 ✭✭purplefields


    I'll give you an example of what I recently did, using standard ChatGPT.

    I had a medical problem so I went to the Doctor last April. He said it was migrane related. I put my symptoms into ChatGPT, which suggested I get an MRI. I went back to my doctor, and they agreed for me to get the MRI.

    I got a link to the scans emailed to me, so I took some screenshots and uploaded those to ChatGPT. It analysed the images and told me exactly what the problem was. A week later, by own GP came back with exactly the same results from the radiologist, but much less detailed. I got an appointment to see a neurologist in but in one year's time!

    Not wanting to wait that long for a condition that can deteriorate, I am now following a ChatGPT treatment program, and am getting better with it.

    So, if plain old out of the box ChatGPT can do all that, imagine what a custom RAG will do? I can't see how legal advice will be much of a challenge for it.



  • Registered Users, Registered Users 2 Posts: 346 ✭✭purplefields


    It's a probabilistic large language model. It's finding the most likely sequence of words to get an answer. This is how it works.

    But, technology is very quickly moving forward. Even over a matter of months I can see huge improvements.

    Also:

    The Battle of Stalingrad took place during World War II from:

    23 August 1942 – 2 February 1943

    It was one of the largest and deadliest battles in history and marked a major turning point on the Eastern Front, with the Soviet Union defeating Nazi Germany’s 6th Army.

    If you want, I can give a short summary or key facts about the battle.

    Wikipedia says 17 July 1942 it started.



  • Registered Users, Registered Users 2 Posts: 1,435 ✭✭✭Merrion


    We have replaced a junior developer (who left to get another job) with Claude [AI] and no meaningful difference in quality.

    Realistically I (senior developer) will be replaced with AI within the next 3 years. It works very well in the constrained environment of a programming language with a huge training set…all the open source code ever developed.



  • Advertisement
  • Registered Users, Registered Users 2 Posts: 4,553 ✭✭✭arctictree


    If you think about it, AI has the experience of thousands of volumes of research. Are we at the stage where a consultation with AI is better than an in person doctor visit?



  • Registered Users, Registered Users 2 Posts: 927 ✭✭✭littlefeet


    Possibly: That's missing a huge amount of what some individuals attend doctors for, vague symptoms, validation for adopting the sick role, or mental health issues, looking for a medical prescription for modern maladies of feeling of loneliness, unhappiness, and dissatisfaction with life. It's not as simple as saying here is a list of symptoms to an online AI doctor.



  • Registered Users, Registered Users 2 Posts: 6,558 ✭✭✭silliussoddius


    Research or scraping the internet, doesn't mean they can deal with ad-hoc on the spot issues.



  • Registered Users, Registered Users 2, Paid Member Posts: 7,775 ✭✭✭Allinall


    It also doesn’t mean that the results are trustworthy.

    Can AI identify fake news, of which the internet is now overloaded?



  • Registered Users, Registered Users 2, Paid Member Posts: 14,231 ✭✭✭✭Cluedo Monopoly


    That makes no sense.

    AI is used to create fake news. It's quite simple - AI will do what you want it to do within the boundaries that you specify. It is reliable and adds value if given specific instructions and knowledge sources.

    This thread is about AI taking jobs and it absolutely will.

    What are they doing in the Hyacinth house?



  • Advertisement
  • Moderators, Music Moderators Posts: 11,430 Mod ✭✭✭✭lordgoat


    if you're using an ai for medical advice you deserve everything that comes with it.



  • Registered Users, Registered Users 2 Posts: 27,019 ✭✭✭✭Mrs OBumble


    So in this poster's case, they've got a treatment approach now, instead of in a years time when a specialist can see them. Most people would see that as an improvement.

    Seems to me that medicine is one area where judicious use of AI could have very large benefits and allow specialists to use their time better.



  • Registered Users, Registered Users 2 Posts: 346 ✭✭purplefields


    Absolutely.

    I also did go to the GP first, as my first port of call. I was mis-diagnosed. I could have spent the money on a few months of ChatGPT plus instead!



  • Registered Users, Registered Users 2, Paid Member Posts: 14,231 ✭✭✭✭Cluedo Monopoly


    That's very bad minded. AI will absolutely transform healthcare. It's one of its best use cases. It is already making an impact.

    What are they doing in the Hyacinth house?



  • Moderators, Music Moderators Posts: 11,430 Mod ✭✭✭✭lordgoat


    people want convenience and to hear the right thing. ChatGPT and AI are great for this like I said if you're using an ai for medical advice you deserve everything that comes with it.

    AI is already changing healthcare. Using chatgpt is not healthcare. Unless you're an idiot.



  • Registered Users, Registered Users 2, Paid Member Posts: 8,827 ✭✭✭plodder


    Anyone who wants to find out what AI is about should listen to this. Brendan O'Connor interviewing the godfather of AI, Geoffrey Hinton last Sunday.

    https://www.rte.ie/radio/radio1/clips/22559818/

    “The opposite of 'good' is 'good intentions'”



  • Registered Users, Registered Users 2, Paid Member Posts: 14,231 ✭✭✭✭Cluedo Monopoly


    Employment in the tech sector here is down 20,000 from it's peak last year.

    Post edited by Cluedo Monopoly on

    What are they doing in the Hyacinth house?



  • Registered Users, Registered Users 2, Paid Member Posts: 2,034 ✭✭✭Packrat


    Thats modern horror...

    How long before we have reasonable rational people (rather than crusty cranks) disconnecting themselves from tech of all sorts in an effort to keep their thoughts amd lives private and un-managed by AI.

    Maybe the Amish have the right idea....



  • Registered Users, Registered Users 2 Posts: 4,308 ✭✭✭AidoEirE


    He's literally detailed how its worked for him for the better in his health.

    Not using it as a doctor but feedback is nothing wrong with it



  • Registered Users, Registered Users 2 Posts: 14,878 ✭✭✭✭mrcheez


    I'm actively looking at roles where I train AI … I'm a dev for 25 years but all for AI taking over and us humans just fine-tuning their output. The speed that stuff can be done is insane, just quality and more importantly consistency is lacking.

    As much as I love coding, I prefer AI doing much of the grunt work that always happens when starting a new project.



  • Advertisement
  • Registered Users, Registered Users 2 Posts: 1,076 ✭✭✭JPup




  • Registered Users, Registered Users 2, Paid Member Posts: 8,827 ✭✭✭plodder


    What I meant to say was that Hinton gave a very good explanation of how AI works. What he says about the future with AI is certainly thought provoking and credible, and could be described as "modern horror", but he's fairly clear that none of it is inevitable, except maybe the job losses.

    “The opposite of 'good' is 'good intentions'”



  • Registered Users, Registered Users 2, Paid Member Posts: 2,034 ✭✭✭Packrat


    His one hope for it not destroying us is that it can be taught to behave like a mother to a child with us... I'd say the bad stuff is pretty inevitable.



  • Registered Users, Registered Users 2 Posts: 3,036 ✭✭✭rogber


    Old men have a tendency to be apocalyptic and conservative and think everything is going to hell. Just because this guy is the godfather of ai really doesn't make him a profound authority on the future (likewise we should be similarly sceptical of the utopian claims that sam altmann etc make)



  • Registered Users, Registered Users 2, Paid Member Posts: 8,827 ✭✭✭plodder


    Bit of an unfair generalisation there. Though I agree his obvious expertise doesn't necessarily make him a profound authority on the future. My feeling is that he is a very skilled communicator. The dog/cat analogy is the single best explanation for the basis of AI that I've heard. So, I think the mother and child idea for how we might live with AI, was similarly chosen for its communicative impact (which it certainly has), as much as anything else.

    “The opposite of 'good' is 'good intentions'”



  • Registered Users, Registered Users 2, Paid Member Posts: 14,231 ✭✭✭✭Cluedo Monopoly


    Can I ask you - what is the impact of all this increased productivity? I mean due to AI churning out code at the touch of a button there must be more and more features getting into live applications faster and faster. Is it having support impacts? Are product owners expecting much faster deployment of requirements? Are the users coping with increasing change? Etc.

    I think the real money will be in Security. AI makes it easier for hackers to hack. They can let AI try and access systems 24/7. It's hard to see how systems can be fully protected in the future.

    What are they doing in the Hyacinth house?



  • Registered Users, Registered Users 2 Posts: 14,878 ✭✭✭✭mrcheez


    Spend less time writing code and in particular writing tests which is always a pain.

    The key is being explicit about what you want in as much detail as possible and then the results usually turn out as expected.

    Think of it as AI laying out the ingredients for a cake, mixing the ingredients together while being observed by the chef, then the chef puts it in the oven and adds the toppings.

    It frees up more time for testing and bug fixing which would generally be done by humans.



  • Registered Users, Registered Users 2, Paid Member Posts: 14,231 ✭✭✭✭Cluedo Monopoly


    I absolutely get you and I have heard other developers saying that they essentially coax the AI to get the code they need and then review it for accuracy and fine tune it etc. I suppose I am asking about the downstream effects of all this extra productivity and features being delivered faster. See my previous post for context.

    What are they doing in the Hyacinth house?



  • Registered Users, Registered Users 2 Posts: 14,878 ✭✭✭✭mrcheez


    As in do projects get done faster? Well yeah 😀



  • Advertisement
  • Registered Users, Registered Users 2, Paid Member Posts: 14,231 ✭✭✭✭Cluedo Monopoly


    I think you missed my questions again!

    Here they are....Re increased productivity and increased content in software releases - Is it having production support impacts? Are product owners subsequently expecting much faster deployment of complex requirements? Are the application users coping with the increasing change? Etc. Also has hiring for programmers stalled?

    What are they doing in the Hyacinth house?



Advertisement
Advertisement