AI - Artificial Intelligence

There are 29 replies in this Thread which has previously been viewed 525 times. The latest Post (December 21, 2024 at 9:13 AM) was by UK Man.

    I have the feeling that AI is as revolutionary as the internet itself, but I admit I don’t get it entirely. I am dedicating December to educate myself on the subject.

    Right now, I subscribed to the paid version of ChatGPT and I am learning to create and customize GPTs for my work.
    So far, I have been just using it to research terminology and rephrasing convoluted sentences, checking the grammar and creating glossaries based on online resources.


    I have also added a GPT on plants and I am testing it for plant caring advice in the southern hemisphere.

    Are you ready for AI? Do you use it?

    I recall when AI was in the early stages of gaining traction through Chat GPT that I decided I wanted to be an early user of it. I also recall thinking that if I did not I’d end up regretting it.

    I’m regretting it.

    The only thing I consider myself to have a high level of knowledge or expertise in is bankruptcy law in England and Wales. So on occasion I’ve asked Chat GPT to produce something or other to do with that. Slowly but surely it has got extremely good at answering some very challenging questions.


    It’s never too late of course to learn how to best utilise what we have available to us now and as serafina said above it’s as revolutionary as the internet itself.

    But…

    Like the internet itself it is incumbent upon the user to learn how to use it properly to get the best out of it.

    Addendum.

    A few weeks ago I watched a video of a school teacher ask some AI tool or other to create a podcast between 2 people discussing a document that details a particular examination syllabus. She input a document detailing it and let AI do its thing.

    The result was nothing short of astonishing.

    I wonder what it would come up with if asked to create a podcast between a British expat and a local Argentine discussing Milei’s first year in office!

  • Post by Bombonera (December 8, 2024 at 1:51 PM).

    This post was deleted by the author themselves (December 23, 2024 at 1:57 AM).

    serafina that’s when anti-Natalism will kick in as a bigger force than it is now. Though traction is gaining slowly with that already.

    @Bombonera , perhaps you can answer this question: I know there is more than one AI engine out these. What is not clear to me, is if ChatGPT is the most advanced one, or if it was just the first to be launched and this is why it is much more famous than the others. I see that my Chrome browser uses Gemini by default. Whenever I am googling something, it will first return a short reply from Gemini and it is usually spot on.


    Also, one thing I am failing to grasp is the following: I am bombarded with advertisements from various apps that rely on AI for specific tasks. Some are for home staging, others to "try" the best haircut based on your picture, or will show how you will look like when you are older, or what will the babies of a couple will look like based on the couple's picture, or creating engaging posts for social media. Are they all GPTs that have been customized to a high level for that task or are they actual softwares that have been coded?

    When I asked ChatGPT what haircut would look best for my face, its answer was borderline condescending, telling me that the best haircut is the one that makes me feel happy and that I shouldn't be rely on other people's judgment to decide on a haircut. :facepalm:

    That’s a very salient question serafina

    Because I know there are other AI apps or engines or whatever they are. Some paid and unpaid that stand outside of Chat GPT.

    So I am faced with the same or similar question.

    I recall a friend referring me to one most useful for academic research for example. Though I forget now what it was.

    Now there’s another question. How is academia dealing with this threat to academic integrity these days?

    There are websites tailored to spot AI-generated text, although from my limited research they work with English texts, only. However, you can always copy-paste a text in ChatGPT and ask it if they wrote it.
    I tried it a few time and ChatGPT was lying. :) It tend to take too much credit.

    There is a lot of discussion around AI in the translation community. A colleague run her own translation of an architectural brochure in ChatGPT and ChatGPT said it was AI-generated (it was not). I tried the same with an essay I had written on my own, and it also took credit for it (it was not).

    A friend of mine teaches in a bilingual high school and told me he stopped assigning essays as homeworks as they are all done by ChatGPT. He was getting essays that sounded written by a native adult speaker, but his students were unable to communicate that fluently orally and definitely don't have that much vocabulary.


    For me, personally, I find ChatGPT completely off for legal translations, as it misses the mark on the legalese terminology and false friends. However, it does a fine job rephrasing my own translations. I can't use AI extensively due to NDA and the sensitive nature of most of my work, but it is able to point me in the right direction. However, when rephrasing it tends also to add stuff that surely sound it more professional, but it is not appropriate to add for embellishment in the case of a translation, as it would stray too much from the source.

    serafina it's disappointing that AI engines are over-zealous in identifying Original work from AI-produced work. It suggests the challenge remains ongoing in that regard.

    Though for you I guess it’s a tool of assistance rather than replacement and as you can see its flaws you can adjust your work accordingly.
    It’s a fascinating area but I cannot anecdotally discuss it that much due to my own limited use.

    I will run through a 15000 word insolvency work I produced myself and see what AI makes of it as to originality.

    Kudos to you, serafina , for dedicating this month to studying AI. For me, it is still as elusive as bitcoin, and I know I need to get with it. Am just too lazy at the moment to take the plunge.

    😵‍💫

    I think I developed a negative attitude about AI when it was used to produce videos and voice messages that were essentially indistinguishable from the real people. It doesn’t take a genius to imagine the damage that can be done.

    Now that it is part of our daily lives, it’s time to look for beneficial uses of AI. Medical diagnostics and precise translation are the first two that jump to mind.

    Here’s a more esoteric example of what AI can do for us.

    Hundreds More Nazca Lines Emerge in Peru’s Desert (Gift Article)
    With drones and A.I., researchers managed to double the number of mysterious geoglyphs in a matter of months.
    www.nytimes.com

    No idea what the rules/laws are regarding putting false AI stuff on the internet. Perhaps - if it's possible - the site hosts should be forced to display whether the content is genuine or not?

    In my experience of site hosting companies they can range from “don’t give a flying fuck” to being over-zealous.

    I uploaded a domain to my hosting that has the word “bankruptcy” in the domain name. The host wouldn’t let me upload it because the name contained the word “bank”.

    I kid you not.

    I did resolve the matter obviously.

    Quote from Bombonera

    In my experience of site hosting companies they can range from “don’t give a flying fuck” to being over-zealous.

    There are probably too many of the latter, but there are DEFINITELY too many of the former.

    The internet is still the Wild, Wild West.

    Quote from serafina

    perhaps you can answer this question: I know there is more than one AI engine out these. What is not clear to me, is if ChatGPT is the most advanced one, or if it was just the first to be launched and this is why it is much more famous than the others. I see that my Chrome browser uses Gemini by default. Whenever I am googling something, it will first return a short reply from Gemini and it is usually spot on

    As much as I hate giving any feedback to zuckers Meta, their LLM called LLama is another that is talked about. Most interesting to me since it is OpenSource allowing everyone to research the internal code. Perhaps they needed to do this since they were late to the Large Language Model party. I keep meaning to download a version, or install on a more powerful cloud virtual machine than I have at home, to better educate myself

    I am optimistic that they all will, if not already, make good assistants. Helping us to produce things that were not possible because of time constraints, or just too boring. Lots of pros, cons, and “dragons”.

    Are we on the verge of a self-improving AI explosion?
    An AI that makes better AI could be “the last invention that man need ever make.”…
    arstechnica.com
    Meta unveils a new large language model that can run on a single GPU [Updated]
    LLaMA-13B reportedly outperforms ChatGPT-like tech despite being 10x smaller.
    arstechnica.com

    My FB page where I offer my professional services was banned from FB ads for over a year because - I suppose, as no official explanation was given - I was using images of passports to advertise my translation of official records for naturalization purposes. Their algorithm likely understood that I was selling fake passports and forged vital records. I am pretty clear in my wording, so I believe it just based the flag on the imagery I used.

    No appeal was available, and I don't really advertise that much. Then one day I received a notice that they had reinstated my account for advertising on FB. By then, I had moved my marketing efforts to Instagram, anyway.