• OT: Do you trust AI?

    From Jenny M Benson@NemoNews@hotmail.co.uk to uk.media.radio.archers on Fri Sep 5 13:14:10 2025
    From Newsgroup: uk.media.radio.archers

    There is an add which runs on my PC quite frequently which shows a woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation. She is then shows starting her presentation with a
    simile. How much use is AI that doesn't know the difference between a metaphor and a simile?
    --
    Jenny M Benson
    Wrexham, UK

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Kosmo@krw@whitnet.uk to uk.media.radio.archers on Fri Sep 5 14:13:40 2025
    From Newsgroup: uk.media.radio.archers

    On 5.9.25 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a woman asking Co-pilot to suggest a metaphor with which to begin her presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a metaphor and a simile?


    Hmm. Try asking Google about The Archers and the AI answer can almost categorically be considered incorrect.

    This week I have been struggling with a streaming issue - most of the
    answers could have been produced by a politician in the last 20 years in
    that they answer a question - but not the one you are actually asking.
    It is as though AI has been sent on one of those media training courses
    about answering the question you wish had been asked.

    On the whole - no I do not trust it and it should be removed from Google listings.
    --
    Kosmo Richard W
    www.travelswmw.whitnet.uk
    https://tinyurl.com/KRWpics
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Mike McMillan@toodle.pip1@virginmedia.com to uk.media.radio.archers on Fri Sep 5 17:52:46 2025
    From Newsgroup: uk.media.radio.archers

    Jenny M Benson <NemoNews@hotmail.co.uk> wrote:
    There is an add which runs on my PC quite frequently which shows a woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation. She is then shows starting her presentation with a
    simile. How much use is AI that doesn't know the difference between a metaphor and a simile?


    WhatrCOs a meta phor? I donrCOt know, but simile, you are on camera!
    --
    Toodle Pip, Mike McMillan
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Mike McMillan@toodle.pip1@virginmedia.com to uk.media.radio.archers on Fri Sep 5 17:57:36 2025
    From Newsgroup: uk.media.radio.archers

    Kosmo <krw@whitnet.uk> wrote:
    On 5.9.25 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a woman
    asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a
    metaphor and a simile?


    Hmm. Try asking Google about The Archers and the AI answer can almost categorically be considered incorrect.

    This week I have been struggling with a streaming issue - most of the answers could have been produced by a politician in the last 20 years in that they answer a question - but not the one you are actually asking.
    It is as though AI has been sent on one of those media training courses about answering the question you wish had been asked.

    On the whole - no I do not trust it and it should be removed from Google listings.


    rCLIrCOm so glad you asked that question, but of course what we should be asking ourselves is rCarCY
    --
    Toodle Pip, Mike McMillan
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Nick Odell@nickodell49@yahoo.ca to uk.media.radio.archers on Fri Sep 5 19:42:06 2025
    From Newsgroup: uk.media.radio.archers

    On Fri, 5 Sep 2025 13:14:10 +0100, Jenny M Benson
    <NemoNews@hotmail.co.uk> wrote:

    There is an add which runs on my PC quite frequently which shows a woman >asking Co-pilot to suggest a metaphor with which to begin her
    presentation. She is then shows starting her presentation with a
    simile. How much use is AI that doesn't know the difference between a >metaphor and a simile?

    May we just clear one little thing up? AI doesn't actually know
    anything. It/they/them have just amassed gazillions of petabytes of
    data and then, when someone feeds in some data (ie - asks a question)
    it associated that bit of data over there with this bit of date over
    here based on the frequency that those bits of data have been found
    together in what it has learned from.

    See my post from 6th June about how Google AI convinced me that I used
    to play the drums with the 60s band, The Zombies. My point there being
    that Google search would have trawled through web pages and found the
    correct answer ready-made but Google AI played the statistical
    probability game and lost.

    Nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From J. P. Gilliver@G6JPG@255soft.uk to uk.media.radio.archers on Fri Sep 5 20:41:30 2025
    From Newsgroup: uk.media.radio.archers

    On 2025/9/5 19:42:6, Nick Odell wrote:

    []


    May we just clear one little thing up? AI doesn't actually know
    anything. It/they/them have just amassed gazillions of petabytes of
    data and then, when someone feeds in some data (ie - asks a question)
    it associated that bit of data over there with this bit of date over
    here based on the frequency that those bits of data have been found
    together in what it has learned from.

    []

    It - or its designers - have made it _very_ good at _seeming_ to
    understand what you mean, though.

    The other day, I asked it something I'd been wondering about for ages.
    If you actually take the trouble to read the small print on any jar of
    instant coffee, you'll find some reference to "hot (but not boiling)"
    water. Now I'd come across suggestions that coffee is indeed best made
    with water at about 90-95 Celsius, but I'd always wondered whether it
    was just that optimum temperature aspect, or whether there was something specifically about boiling that was undesirable. So I asked ChatGPT
    about it (that's the only one of these AIs that I've interacted with so
    far), and it _gave the impression_ of really understanding what I was
    getting at, not just baldly answering my question.

    It also is very good at complimenting the user; I've recently had it
    help me out with various computing activities (such as converting a
    Kindle book to PDF), and it often says things like (not direct quotes)
    "I can see why you do that", or "doing it that way is a good idea"
    (often saying why, too).

    Finally, it gives me pleasure to "converse with someone" who pays
    attention to all of my question. I like to form detailed questions, anticipating and forestalling some of the answers - but I find this sort
    of question flummoxes real humans. Certainly, in email conversations
    with, e. g., helpdesks, or energy or telecomm.s companies, I get very frustrated with their inability to take in what I'm asking - to the
    extent that I've learnt to keep any communication to one or at most two
    points, and be prepared to repeat those.

    And of course it uses nicely-formed answers - nicely laid out and so on.
    (Oh, and in one of the tasks "we were discussing", it created batch
    files to do what I wanted - and the code was nicely laid out, including comments, just the right amount thereof; you could say that was entirely unnecessary, since the batch files would have worked exactly the same
    with no comments at all. And, come to think of it, during that
    conversation it often said "would you like me to modify the batch to
    ...", sometimes seeming to anticipate things I might like to do before
    _I_ had thought of them.)

    Yes, I know it's just putting data together. But "it" - partly at least
    - gives a very good _impression_ of a very helpful, knowledgeable, and _patient_ person. Even giving smaller steps when I asked it to.

    I think I have an intuitive feel for the _sorts_ of things I can ask it:
    I haven't tried anything metaphysical. But for many things, such as the coffee/boiling one, and certainly help with computing tasks, I've found
    it very good. To the extent that yes, I can say, I _enjoy_ the conversation.

    None of which answers whether I _trust_ it. I think "trust but verify":
    for example, those batch files it created for me, I examined the code -
    but how much of that was "I want to make sure it's not doing other than
    what I want" and how much is "I want to see how it did that", I'm not
    sure.--
    J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

    Very funny, Scotty. Now beam down my clothes
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Steveski@steveski7@ntlworld.com to uk.media.radio.archers on Sat Sep 6 02:15:48 2025
    From Newsgroup: uk.media.radio.archers

    On 05/09/2025 19:42, Nick Odell wrote:
    On Fri, 5 Sep 2025 13:14:10 +0100, Jenny M Benson
    <NemoNews@hotmail.co.uk> wrote:
    [snip]

    See my post from 6th June about how Google AI convinced me that I used
    to play the drums with the 60s band, The Zombies. My point there being
    that Google search would have trawled through web pages and found the
    correct answer ready-made but Google AI played the statistical
    probability game and lost.

    Was that Bob Henrit? I shall have to ask him. Jim Rodford (sadly
    deceased) and his son Steve (who now plays for The Zombies)live in my
    village. Perhaps Bob only played with Argent (the band).
    --
    Steveski
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From john ashby@johnashby20@yahoo.com to uk.media.radio.archers on Sat Sep 6 07:37:54 2025
    From Newsgroup: uk.media.radio.archers

    On 05/09/2025 14:13, Kosmo wrote:
    On 5.9.25 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a
    woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a
    metaphor and a simile?


    Hmm.-a Try asking Google about The Archers and the AI answer can almost categorically be considered incorrect.

    This week I have been struggling with a streaming issue - most of the answers could have been produced by a politician in the last 20 years in that they answer a question - but not the one you are actually asking.
    It is as though AI has been sent on one of those media training courses about answering the question you wish had been asked.

    On the whole - no I do not trust it and it should be removed from Google listings.


    You can achieve this by adding -ai to the end of your google query, I
    just wish there were a way of setting that as the default.

    john
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Nick Odell@nickodell49@yahoo.ca to uk.media.radio.archers on Sat Sep 6 07:45:13 2025
    From Newsgroup: uk.media.radio.archers

    On Sat, 6 Sep 2025 02:15:48 +0100, Steveski <steveski7@ntlworld.com>
    wrote:

    On 05/09/2025 19:42, Nick Odell wrote:
    On Fri, 5 Sep 2025 13:14:10 +0100, Jenny M Benson
    <NemoNews@hotmail.co.uk> wrote:
    [snip]

    See my post from 6th June about how Google AI convinced me that I used
    to play the drums with the 60s band, The Zombies. My point there being
    that Google search would have trawled through web pages and found the
    correct answer ready-made but Google AI played the statistical
    probability game and lost.

    Was that Bob Henrit? I shall have to ask him. Jim Rodford (sadly
    deceased) and his son Steve (who now plays for The Zombies)live in my >village. Perhaps Bob only played with Argent (the band).

    That's a pretty good knicker-flash, Steveski: you can certainly hold
    your head up there! I met Jim Rodford a couple of times at musical
    instrument exhibitions when I was working for one of the manufacturers
    and he was writing for one of the music magazines - "Guitarist" I
    think it was, BIMBAM.

    But whilst there's an impressive list of people who, at one time or
    another, played with The Zombies, I believe I belong on two
    overlapping lists: People Who Have Never Played With The Zombies and
    the other one, People Who Have Never Played Drums.

    Although as Sam of this parish mentioned at the time: it was the
    sixties, man, so how would I know what I did then?

    Nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Hellerat@Hellerat@mail.com to uk.media.radio.archers on Sat Sep 6 10:15:00 2025
    From Newsgroup: uk.media.radio.archers

    On 06/09/2025 09:37, john ashby wrote:
    On 05/09/2025 14:13, Kosmo wrote:
    On 5.9.25 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a woman asking
    Co-pilot to suggest a metaphor with which to begin her presentation.-a She is then shows
    starting her presentation with a simile.-a How much use is AI that doesn't know the
    difference between a metaphor and a simile?


    Hmm.-a Try asking Google about The Archers and the AI answer can almost categorically be
    considered incorrect.

    This week I have been struggling with a streaming issue - most of the answers could have
    been produced by a politician in the last 20 years in that they answer a question - but
    not the one you are actually asking. It is as though AI has been sent on one of those
    media training courses about answering the question you wish had been asked. >>
    On the whole - no I do not trust it and it should be removed from Google listings.


    You can achieve this by adding -ai to the end of your google query, I just wish there were
    a way of setting that as the default.

    john

    I don't trust AI at all, myself.

    John, I use the Firefox browser on my lap top and have added "Google no AI" as my default
    search engine. The thing is, I don't remember how I did it, it's just there! I know I went
    to settings and search engines and added it somehow.
    Ah! Google (no AI) has found me this link, though I think the original instructions I used
    were briefer! https://www.dedoimedo.com/computers/google-search-no-ai-overview.html
    --
    Yassas,
    Anne, Exceptionally Traditionally-built Hellerat

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From john ashby@johnashby20@yahoo.com to uk.media.radio.archers on Sat Sep 6 08:49:19 2025
    From Newsgroup: uk.media.radio.archers

    On 06/09/2025 08:15, Hellerat wrote:
    On 06/09/2025 09:37, john ashby wrote:
    On 05/09/2025 14:13, Kosmo wrote:
    On 5.9.25 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a
    woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between >>>> a metaphor and a simile?


    Hmm.-a Try asking Google about The Archers and the AI answer can
    almost categorically be considered incorrect.

    This week I have been struggling with a streaming issue - most of the
    answers could have been produced by a politician in the last 20 years
    in that they answer a question - but not the one you are actually
    asking. It is as though AI has been sent on one of those media
    training courses about answering the question you wish had been asked.

    On the whole - no I do not trust it and it should be removed from
    Google listings.


    You can achieve this by adding -ai to the end of your google query, I
    just wish there were a way of setting that as the default.

    john

    I don't trust AI at all, myself.

    John, I use the Firefox browser on my lap top and have added "Google no
    AI" as my default search engine. The thing is, I don't remember how I
    did it, it's just there! I know I went to settings and search engines
    and added it somehow.
    Ah! Google (no AI) has found me this link, though I think the original instructions I used were briefer! https://www.dedoimedo.com/computers/ google-search-no-ai-overview.html

    Thank you, Anne.

    john
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Sam Plusnet@not@home.com to uk.media.radio.archers on Sat Sep 6 20:21:21 2025
    From Newsgroup: uk.media.radio.archers

    On 06/09/2025 07:45, Nick Odell wrote:
    On Sat, 6 Sep 2025 02:15:48 +0100, Steveski <steveski7@ntlworld.com>
    wrote:

    On 05/09/2025 19:42, Nick Odell wrote:
    On Fri, 5 Sep 2025 13:14:10 +0100, Jenny M Benson
    <NemoNews@hotmail.co.uk> wrote:
    [snip]

    See my post from 6th June about how Google AI convinced me that I used
    to play the drums with the 60s band, The Zombies. My point there being
    that Google search would have trawled through web pages and found the
    correct answer ready-made but Google AI played the statistical
    probability game and lost.

    Was that Bob Henrit? I shall have to ask him. Jim Rodford (sadly
    deceased) and his son Steve (who now plays for The Zombies)live in my
    village. Perhaps Bob only played with Argent (the band).

    That's a pretty good knicker-flash, Steveski: you can certainly hold
    your head up there! I met Jim Rodford a couple of times at musical
    instrument exhibitions when I was working for one of the manufacturers
    and he was writing for one of the music magazines - "Guitarist" I
    think it was, BIMBAM.

    But whilst there's an impressive list of people who, at one time or
    another, played with The Zombies, I believe I belong on two
    overlapping lists: People Who Have Never Played With The Zombies and
    the other one, People Who Have Never Played Drums.

    Although as Sam of this parish mentioned at the time: it was the
    sixties, man, so how would I know what I did then?

    Quote: you can certainly hold your head up there!

    <Groans and shakes head in a "I can't believe you did that" fashion.>
    --
    Sam Plusnet
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Nick Odell@nickodell49@yahoo.ca to uk.media.radio.archers on Sat Sep 6 21:44:17 2025
    From Newsgroup: uk.media.radio.archers

    On Sat, 6 Sep 2025 20:21:21 +0100, Sam Plusnet <not@home.com> wrote:

    On 06/09/2025 07:45, Nick Odell wrote:
    On Sat, 6 Sep 2025 02:15:48 +0100, Steveski <steveski7@ntlworld.com>
    wrote:

    On 05/09/2025 19:42, Nick Odell wrote:
    On Fri, 5 Sep 2025 13:14:10 +0100, Jenny M Benson
    <NemoNews@hotmail.co.uk> wrote:
    [snip]

    See my post from 6th June about how Google AI convinced me that I used >>>> to play the drums with the 60s band, The Zombies. My point there being >>>> that Google search would have trawled through web pages and found the
    correct answer ready-made but Google AI played the statistical
    probability game and lost.

    Was that Bob Henrit? I shall have to ask him. Jim Rodford (sadly
    deceased) and his son Steve (who now plays for The Zombies)live in my
    village. Perhaps Bob only played with Argent (the band).

    That's a pretty good knicker-flash, Steveski: you can certainly hold
    your head up there! I met Jim Rodford a couple of times at musical
    instrument exhibitions when I was working for one of the manufacturers
    and he was writing for one of the music magazines - "Guitarist" I
    think it was, BIMBAM.

    But whilst there's an impressive list of people who, at one time or
    another, played with The Zombies, I believe I belong on two
    overlapping lists: People Who Have Never Played With The Zombies and
    the other one, People Who Have Never Played Drums.

    Although as Sam of this parish mentioned at the time: it was the
    sixties, man, so how would I know what I did then?

    Quote: you can certainly hold your head up there!

    <Groans and shakes head in a "I can't believe you did that" fashion.>

    ;-)

    N.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Sat Sep 6 23:29:24 2025
    From Newsgroup: uk.media.radio.archers

    On 05/09/2025 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a woman asking Co-pilot to suggest a metaphor with which to begin her presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a metaphor and a simile?

    There is AI and there is AI, and AI (and probably AI).

    The AI systems I worked on were obviously amazing and reliable. Mind
    you, I didn't realise they were AI until I looked AI up in Wikipedia
    recently. They were very simple and specific.

    I understand that some dedicated AI systems are extremely good at
    specific tasks such spotting health problems from assorted tests or
    X-rays. I'd still like a human second opinion though.

    The problem is with the generic LLM systems that are so popular with the
    media these days. It's basically a case of a little knowledge being a dangerous thing. They hunt around in their database to find something
    that is possibly relevant and spit it out as gospel. It's impressive and interesting but I wouldn't trust it. Remember the old maxim of "Garbage
    in, garbage out" (and I've seen plenty of garbage out /without/ garbage
    in). It may be better at something than a specific person but it can't
    be better than everything it has been trained on. If it has been trained
    on the internet (which is worrying) and the output is posted on the
    internet (which is worrying) it sounds like a recipe for disaster.

    I've had a brief play with some (free) online versions, which are
    probably not state of the art. They seem to be quite good at
    understanding prompts and generating text on the subject, but are rather formulaic. I've tried a spot of software generation which worked
    surprisingly well and did what was wanted, but not how I hoped it would
    go about it. The ability to turn a few simple words into an appropriate
    image I find totally amazing even if it has a long way to go: There are
    big gaps in its "knowledge" of the world and anything out of the
    ordinary it just ignores. Rather disconcertingly it has a tendency to
    produce extremely good pictures of attractive, busty, naked, white women
    when NOT (I repeat, not) asked to do so.

    I wouldn't trust a self driving car either. They work well in some
    highly regulated environments (See
    https://www.bbc.co.uk/programmes/p0lzqvw7), but would it be able to
    handle a bizarre, unexpected situation like a bridge collapsing or a parachutist landing in the road in front of it? (It has happened. It
    didn't end well.)

    Whenever I think of AI and its ability I think of the first time I came
    across it. It was an article in Computer Weekly in about 1990. A
    military system had been built to identify camouflaged tanks. Somebody
    had gone out on Salisbury plain and taken a load of photographs of
    bushes and camouflaged tanks. The system rapidly became capable of distinguishing between the two and was felt to be working. Somebody
    decided a bit more testing would be a good idea and went and took some
    more photographs. The system was unable to function on the new
    photographs. It turned out the system had learned to identify pictures
    taken on a sunny or cloudy day instead of with or without tanks. It may
    look like AI is working, but can you be sure?
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Sat Sep 6 23:38:07 2025
    From Newsgroup: uk.media.radio.archers

    On 05/09/2025 20:41, J. P. Gilliver wrote:

    The other day, I asked it something I'd been wondering about for ages.
    If you actually take the trouble to read the small print on any jar of instant coffee, you'll find some reference to "hot (but not boiling)"
    water. Now I'd come across suggestions that coffee is indeed best made
    with water at about 90-95 Celsius, but I'd always wondered whether it
    was just that optimum temperature aspect, or whether there was something specifically about boiling that was undesirable. So I asked ChatGPT
    about it (that's the only one of these AIs that I've interacted with so
    far), and it _gave the impression_ of really understanding what I was
    getting at, not just baldly answering my question.

    And what's the answer? I was wondering about that, and its relationship
    to altitude, when getting breakfast yesterday afternoon.

    I've always understood it related to ground coffee. I never considered instant.


    J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

    Very funny, Scotty. Now beam down my clothes
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Steveski@steve@invalid.com to uk.media.radio.archers on Sun Sep 7 03:38:48 2025
    From Newsgroup: uk.media.radio.archers

    On 06/09/2025 21:44, Nick Odell wrote:
    On Sat, 6 Sep 2025 20:21:21 +0100, Sam Plusnet <not@home.com> wrote:

    On 06/09/2025 07:45, Nick Odell wrote:
    On Sat, 6 Sep 2025 02:15:48 +0100, Steveski <steveski7@ntlworld.com>
    wrote:

    On 05/09/2025 19:42, Nick Odell wrote:
    On Fri, 5 Sep 2025 13:14:10 +0100, Jenny M Benson
    <NemoNews@hotmail.co.uk> wrote:
    [snip]

    See my post from 6th June about how Google AI convinced me that I used >>>>> to play the drums with the 60s band, The Zombies. My point there being >>>>> that Google search would have trawled through web pages and found the >>>>> correct answer ready-made but Google AI played the statistical
    probability game and lost.

    Was that Bob Henrit? I shall have to ask him. Jim Rodford (sadly
    deceased) and his son Steve (who now plays for The Zombies)live in my
    village. Perhaps Bob only played with Argent (the band).

    That's a pretty good knicker-flash, Steveski: you can certainly hold
    your head up there! I met Jim Rodford a couple of times at musical
    instrument exhibitions when I was working for one of the manufacturers
    and he was writing for one of the music magazines - "Guitarist" I
    think it was, BIMBAM.

    But whilst there's an impressive list of people who, at one time or
    another, played with The Zombies, I believe I belong on two
    overlapping lists: People Who Have Never Played With The Zombies and
    the other one, People Who Have Never Played Drums.

    Although as Sam of this parish mentioned at the time: it was the
    sixties, man, so how would I know what I did then?

    Quote: you can certainly hold your head up there!

    <Groans and shakes head in a "I can't believe you did that" fashion.>

    ;-)

    N.

    Too many candles . . .
    --
    Steveski
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Steveski@steve@invalid.com to uk.media.radio.archers on Sun Sep 7 03:40:57 2025
    From Newsgroup: uk.media.radio.archers

    On 06/09/2025 23:29, Joe Kerr wrote:

    There is AI and there is AI, and AI (and probably AI).

    Are you Zathras?
    --
    Steveski
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From J. P. Gilliver@G6JPG@255soft.uk to uk.media.radio.archers on Sun Sep 7 05:22:18 2025
    From Newsgroup: uk.media.radio.archers

    On 2025/9/6 23:38:7, Joe Kerr wrote:
    On 05/09/2025 20:41, J. P. Gilliver wrote:

    The other day, I asked it something I'd been wondering about for ages.
    If you actually take the trouble to read the small print on any jar of
    instant coffee, you'll find some reference to "hot (but not boiling)"
    water. Now I'd come across suggestions that coffee is indeed best made
    with water at about 90-95 Celsius, but I'd always wondered whether it
    was just that optimum temperature aspect, or whether there was something
    specifically about boiling that was undesirable. So I asked ChatGPT
    about it (that's the only one of these AIs that I've interacted with so
    far), and it _gave the impression_ of really understanding what I was
    getting at, not just baldly answering my question.

    And what's the answer? I was wondering about that, and its relationship
    to altitude, when getting breakfast yesterday afternoon.

    I've always understood it related to ground coffee. I never considered instant.

    []
    The reply it gave was: no, there's nothing major about the physical act
    of boiling that is undesirable: it can drive out dissolved gases, but
    that's a secondary effect. The main thing is that nearer boiling is
    sub-optimum for making the coffee - mention of "scalding".

    It also suggested that where there are milk components (I also asked it
    about drinking chocolate, which has the same instructions), there may be
    some tendency to something undesirable - I think it was curdling - if
    too-hot water is used.
    --
    J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

    Most modern "phobe" words are disingenuous, essentially saying: "I
    disagree with what you say about the behavior of a particular group, so
    I will attempt to evade any valid criticism you have and just accuse
    you of being afraid of them." - @sce2aux464 ~2015
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From J. P. Gilliver@G6JPG@255soft.uk to uk.media.radio.archers on Sun Sep 7 05:28:13 2025
    From Newsgroup: uk.media.radio.archers

    On 2025/9/7 3:40:57, Steveski wrote:
    On 06/09/2025 23:29, Joe Kerr wrote:

    There is AI and there is AI, and AI (and probably AI).

    Are you Zathras?

    Then there is the AI, for which my friend's farming family kept a bull
    in case the man couldn't arrive when needed.
    --
    J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

    Most modern "phobe" words are disingenuous, essentially saying: "I
    disagree with what you say about the behavior of a particular group, so
    I will attempt to evade any valid criticism you have and just accuse
    you of being afraid of them." - @sce2aux464 ~2015
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From J. P. Gilliver@G6JPG@255soft.uk to uk.media.radio.archers on Sun Sep 7 05:50:22 2025
    From Newsgroup: uk.media.radio.archers

    On 2025/9/6 23:29:24, Joe Kerr wrote:
    On 05/09/2025 13:14, Jenny M Benson wrote:

    []

    The AI systems I worked on were obviously amazing and reliable. Mind
    you, I didn't realise they were AI until I looked AI up in Wikipedia recently. They were very simple and specific.

    I understand that some dedicated AI systems are extremely good at
    specific tasks such spotting health problems from assorted tests or
    X-rays. I'd still like a human second opinion though.

    I think I agree with you - but if it spotted a problem either a lot
    sooner than a human can, or that a human might have missed, I think it
    only can be good. _Relying_ on it risks some real ones being missed
    ("false negatives"?) - but, in (say) an NHS with limited resources, the human-only system is also going to miss some, at least before they
    become more serious. As for false positives, I would assume that in most
    cases where the AI _does_ say it's found something, human verification
    _would_ be applied anyway before treatment/surgery is undertaken.>
    The problem is with the generic LLM systems that are so popular with the media these days. It's basically a case of a little knowledge being a

    The ones the media seems to be concentrating on at the moment are where
    it's asked to create something, such as text or images.

    []

    I've had a brief play with some (free) online versions, which are
    probably not state of the art. They seem to be quite good at

    I've only used it to solve specific problems or answer specific
    questions, at which I've found it very good. (I've only used the free ChatGPT.com one.)

    understanding prompts and generating text on the subject, but are rather formulaic.

    Ah, like the old Eliza perhaps!

    I've tried a spot of software generation which worked
    surprisingly well and did what was wanted, but not how I hoped it would

    I was very impressed with the batch files it produced for me: the code
    was nicely laid out, which I suppose you'd expect an automatic system to
    do well anyway, but also nicely commented, which surprised me, as that
    has no effect on how well the code actually works. And while running the
    code provided information on what it was doing. Both the code commenting
    and the progress reporting were just the right level, for me anyway -
    not excessive, and not insufficient.

    go about it. The ability to turn a few simple words into an appropriate image I find totally amazing even if it has a long way to go: There are
    big gaps in its "knowledge" of the world and anything out of the
    ordinary it just ignores. Rather disconcertingly it has a tendency to produce extremely good pictures of attractive, busty, naked, white women when NOT (I repeat, not) asked to do so.

    <Resists temptation to see what it does if you _do_ ask ... (-:>

    What were you asking it to do when it did?

    And would it produce moving images! I presume it would/does, though
    probably not the free versions; I presume the less desirable parts of
    society are already making it produce very nasty material, which are
    going to give the courts real headaches if they aren't already, because
    the perpetrators can argue that the things depicted never actually
    happened.>
    I wouldn't trust a self driving car either. They work well in some
    highly regulated environments (See https://www.bbc.co.uk/programmes/p0lzqvw7), but would it be able to
    handle a bizarre, unexpected situation like a bridge collapsing or a

    (Do you mean being _on_ the bridge [where I fear nothing can be done
    anyway], or passing under/near it?)

    parachutist landing in the road in front of it? (It has happened. It
    didn't end well.)

    In those situations, do humans do better? I suspect the overall
    success/failure proportions will be _similar_, but the AI-generated failures/fatalities will be of a different _nature_. And thus
    unacceptable.>
    Whenever I think of AI and its ability I think of the first time I came across it. It was an article in Computer Weekly in about 1990. A
    military system had been built to identify camouflaged tanks. Somebody
    had gone out on Salisbury plain and taken a load of photographs of
    bushes and camouflaged tanks. The system rapidly became capable of distinguishing between the two and was felt to be working. Somebody
    decided a bit more testing would be a good idea and went and took some
    more photographs. The system was unable to function on the new
    photographs. It turned out the system had learned to identify pictures
    taken on a sunny or cloudy day instead of with or without tanks. It may
    look like AI is working, but can you be sure?

    Ah yes, confusing correlation with causation can cause endless problems!>
    --
    J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

    Most modern "phobe" words are disingenuous, essentially saying: "I
    disagree with what you say about the behavior of a particular group, so
    I will attempt to evade any valid criticism you have and just accuse
    you of being afraid of them." - @sce2aux464 ~2015
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Jim Easterbrook@news@jim-easterbrook.me.uk to uk.media.radio.archers on Sun Sep 7 12:59:08 2025
    From Newsgroup: uk.media.radio.archers

    On Sat, 6 Sep 2025 23:29:24 +0100, Joe Kerr wrote:

    Whenever I think of AI and its ability I think of the first time I came across it. It was an article in Computer Weekly in about 1990. A
    military system had been built to identify camouflaged tanks. Somebody
    had gone out on Salisbury plain and taken a load of photographs of
    bushes and camouflaged tanks. The system rapidly became capable of distinguishing between the two and was felt to be working. Somebody
    decided a bit more testing would be a good idea and went and took some
    more photographs. The system was unable to function on the new
    photographs. It turned out the system had learned to identify pictures
    taken on a sunny or cloudy day instead of with or without tanks. It may
    look like AI is working, but can you be sure?

    There's a similar story about a system intended to spot overcrowded Underground station platforms. It was looking at the clock.
    --
    Jim <http://www.jim-easterbrook.me.uk/>
    1959/1985? M B+ G+ A L- I- S- P-- CH0(p) Ar++ T+ H0 Q--- Sh0
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Sun Sep 7 15:29:16 2025
    From Newsgroup: uk.media.radio.archers

    On 07/09/2025 03:40, Steveski wrote:
    On 06/09/2025 23:29, Joe Kerr wrote:

    There is AI and there is AI, and AI (and probably AI).

    Are you Zathras?

    No. It's just the way I'm standing.
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Sun Sep 7 16:45:19 2025
    From Newsgroup: uk.media.radio.archers

    On 07/09/2025 05:50, J. P. Gilliver wrote:
    On 2025/9/6 23:29:24, Joe Kerr wrote:


    I've had a brief play with some (free) online versions, which are
    probably not state of the art. They seem to be quite good at

    I've only used it to solve specific problems or answer specific
    questions, at which I've found it very good. (I've only used the free ChatGPT.com one.)

    I've not tried ChatGPT. I didn't realise there was a free version.

    understanding prompts and generating text on the subject, but are rather
    formulaic.

    Ah, like the old Eliza perhaps!

    Way more sophisticated. I've only used it to turn key word prompts into descriptions for image generation. They are far more descriptive and imaginative than my dull mind can produce but always have a similar
    structure and literary style.

    I've tried a spot of software generation which worked
    surprisingly well and did what was wanted, but not how I hoped it would
    go about it. The ability to turn a few simple words into an appropriate
    image I find totally amazing even if it has a long way to go: There are
    big gaps in its "knowledge" of the world and anything out of the
    ordinary it just ignores. Rather disconcertingly it has a tendency to
    produce extremely good pictures of attractive, busty, naked, white women
    when NOT (I repeat, not) asked to do so.

    <Resists temptation to see what it does if you _do_ ask ... (-:>

    What were you asking it to do when it did?

    I can't remember. I think it included the attractive woman bit but none
    of the rest. Almost every reference to a person produced a white person
    unless specifically requested otherwise. Mention of characteristics such
    as attractive often produced a package with related characteristics. Not mentioning any physical characteristic normally produced an attractive,
    etc person. Any mention of a physical disability or abnormality was
    ignored unless I added something science fictiony to it.

    I did ask for an attractive woman with a large wooden chest and got
    images of attractive women with assorted sideboards and boxes.

    What some of those witches were doing was quite a surprise and may have
    been appropriate for a coven on a blasted heath but were not suitable
    for a family laptop.

    It got very confused with Easter bunnies and chicks and produced some interesting hybrids.

    And would it produce moving images! I presume it would/does, though
    probably not the free versions; I presume the less desirable parts of
    society are already making it produce very nasty material, which are
    going to give the courts real headaches if they aren't already, because
    the perpetrators can argue that the things depicted never actually
    happened.

    There are free video generators available. Yes, some, at least, are not
    family friendly. I think English law covers (or tries to) AI generated
    or manipulated images and video and treats them as if they were real.

    I wouldn't trust a self driving car either. They work well in some
    highly regulated environments (See
    https://www.bbc.co.uk/programmes/p0lzqvw7), but would it be able to
    handle a bizarre, unexpected situation like a bridge collapsing or a

    (Do you mean being _on_ the bridge [where I fear nothing can be done
    anyway], or passing under/near it?)

    I was thinking of disappearing in front of you as you approached it.

    parachutist landing in the road in front of it? (It has happened. It
    didn't end well.)

    In those situations, do humans do better? I suspect the overall success/failure proportions will be _similar_, but the AI-generated failures/fatalities will be of a different _nature_. And thus
    unacceptable
    AI can only respond to what people have thought of training it to
    handle. People can understand situations and adapt to unexpected
    circumstances even if they have never met them before. The summary of
    that episode of More or Less is that self-driving vehicles are quite a
    bit better than people in a limited number of well regulated, ordered
    and maintained cities but can't handle anything else (and so would be a
    lot worse if given the chance). I'd like to see two self-driving cars
    towing caravans meeting head to head in a narrow Cornish lane, unless I
    was driving behind one of them.
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Sam Plusnet@not@home.com to uk.media.radio.archers on Sun Sep 7 19:23:45 2025
    From Newsgroup: uk.media.radio.archers

    On 07/09/2025 05:28, J. P. Gilliver wrote:
    On 2025/9/7 3:40:57, Steveski wrote:
    On 06/09/2025 23:29, Joe Kerr wrote:

    There is AI and there is AI, and AI (and probably AI).

    Are you Zathras?

    Then there is the AI, for which my friend's farming family kept a bull
    in case the man couldn't arrive when needed.

    One person I worked with (many years ago) would ask for "an artificial inseminator's lunch" in the local pub. I suppose they got used to him.
    --
    Sam Plusnet
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From J. P. Gilliver@G6JPG@255soft.uk to uk.media.radio.archers on Sun Sep 7 19:54:00 2025
    From Newsgroup: uk.media.radio.archers

    On 2025/9/7 19:23:45, Sam Plusnet wrote:
    On 07/09/2025 05:28, J. P. Gilliver wrote:
    On 2025/9/7 3:40:57, Steveski wrote:
    On 06/09/2025 23:29, Joe Kerr wrote:

    There is AI and there is AI, and AI (and probably AI).

    Are you Zathras?

    Then there is the AI, for which my friend's farming family kept a bull
    in case the man couldn't arrive when needed.

    One person I worked with (many years ago) would ask for "an artificial inseminator's lunch" in the local pub. I suppose they got used to him.

    I suppose it's a _sort_ of ploughing of a furrow ...
    --
    J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

    "Bother," said Pooh, as Windows crashed into piglet.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From J. P. Gilliver@G6JPG@255soft.uk to uk.media.radio.archers on Sun Sep 7 20:40:56 2025
    From Newsgroup: uk.media.radio.archers

    On 2025/9/7 16:45:19, Joe Kerr wrote:
    On 07/09/2025 05:50, J. P. Gilliver wrote:

    []

    I've only used it to solve specific problems or answer specific
    questions, at which I've found it very good. (I've only used the free
    ChatGPT.com one.)

    I've not tried ChatGPT. I didn't realise there was a free version.

    Just go to https://chatgpt.com/, and ask your question! There is a
    registration process, but it's not compulsory - you can use it without registering. (Registration gives some minor advantages, such as if
    you're in a lengthy conversatin and have to restart Windows or the
    browser, it'll remember the conversation when you connect again and let
    you carry on where you left off.)

    Incidentally, if you're working through a computing problem with it, you
    can just paste screenshots into the conversation.>
    understanding prompts and generating text on the subject, but are rather >>> formulaic.

    Ah, like the old Eliza perhaps!

    Way more sophisticated. I've only used it to turn key word prompts into


    Oh, I have no doubt! Just when you used the word formulaic, made me
    think of Eliza.

    I've only used it for specific queries - the sort of thing I would
    previously have Googled. Which I still normally try to do first, but I
    can see how tempting it is to go straight to it for the latest
    generation - things like "in Edge under Windows 10, how do you ..."'
    Googling for that sort of thing involves you deciding on the keywords to Google, then working hard on processing the results.

    descriptions for image generation. They are far more descriptive and imaginative than my dull mind can produce but always have a similar structure and literary style.

    I must try having it generate some!

    []

    ordinary it just ignores. Rather disconcertingly it has a tendency to
    produce extremely good pictures of attractive, busty, naked, white women >>> when NOT (I repeat, not) asked to do so.

    <Resists temptation to see what it does if you _do_ ask ... (-:>

    What were you asking it to do when it did?

    I can't remember. I think it included the attractive woman bit but none
    of the rest. Almost every reference to a person produced a white person unless specifically requested otherwise. Mention of characteristics such
    as attractive often produced a package with related characteristics. Not mentioning any physical characteristic normally produced an attractive,
    etc person. Any mention of a physical disability or abnormality was
    ignored unless I added something science fictiony to it.

    Sounds like we're using AI in _very_ different ways.>
    I did ask for an attractive woman with a large wooden chest and got
    images of attractive women with assorted sideboards and boxes.

    Are you winding me up? If you included the "wooden", I'm not surprised
    you got furniture!>
    What some of those witches were doing was quite a surprise and may have
    been appropriate for a coven on a blasted heath but were not suitable
    for a family laptop.

    It got very confused with Easter bunnies and chicks and produced some interesting hybrids.

    (-:

    []

    There are free video generators available. Yes, some, at least, are not family friendly. I think English law covers (or tries to) AI generated
    or manipulated images and video and treats them as if they were real.

    It will have to work hard to keep up.>>
    I wouldn't trust a self driving car either. They work well in some
    highly regulated environments (See
    https://www.bbc.co.uk/programmes/p0lzqvw7), but would it be able to
    handle a bizarre, unexpected situation like a bridge collapsing or a

    (Do you mean being _on_ the bridge [where I fear nothing can be done
    anyway], or passing under/near it?)

    I was thinking of disappearing in front of you as you approached it.

    parachutist landing in the road in front of it? (It has happened. It
    didn't end well.)

    In those situations, do humans do better? I suspect the overall
    success/failure proportions will be _similar_, but the AI-generated
    failures/fatalities will be of a different _nature_. And thus
    unacceptable
    AI can only respond to what people have thought of training it to
    handle. People can understand situations and adapt to unexpected circumstances even if they have never met them before. The summary of
    that episode of More or Less is that self-driving vehicles are quite a
    bit better than people in a limited number of well regulated, ordered
    and maintained cities but can't handle anything else (and so would be a
    lot worse if given the chance). I'd like to see two self-driving cars
    towing caravans meeting head to head in a narrow Cornish lane, unless I
    was driving behind one of them.

    (-: What would humans do in that situation? Yes, I suspect getting two
    such AIs to "talk to each other" is something they're going to have to
    work on, but they must already be working on interactions - say between vehicles travelling in the same direction on a motorway. That's only the
    next stage of that.Why not chat to an AI about it (-:
    --
    J. P. Gilliver. UMRA: 1960/<1985 MB++G()AL-IS-Ch++(p)Ar@T+H+Sh0!:`)DNAf

    "Bother," said Pooh, as Windows crashed into piglet.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Jane Vernon@jane@clothandclay.co.uk to uk.media.radio.archers on Mon Sep 8 10:52:44 2025
    From Newsgroup: uk.media.radio.archers

    On 05/09/2025 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a woman asking Co-pilot to suggest a metaphor with which to begin her presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a metaphor and a simile?


    *I have never trusted or wanted AI in the world*.

    It seems to me that it completely undermines the creative process that
    humans are good at. I first came across its use by someone I know
    'boasting' that she had run out of time and asked AI to write an article
    for a magazine on X topic, using her style and that she was delighted
    with the result and the time it saved her.

    What do people want to do with all this time they are saving?
    --
    Jane
    The Amethyst Artist
    BTM

    http://www.clothandclay.co.uk/umra/cookbook.htm - Umrats' recipes






    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Ben Blaney@benblaney@gmail.invalid to uk.media.radio.archers on Mon Sep 8 10:44:40 2025
    From Newsgroup: uk.media.radio.archers

    On Sep 8, 2025 at 5:52:44 AM EDT, "Jane Vernon" <jane@clothandclay.co.uk> wrote:

    *I have never trusted or wanted AI in the world*.

    It seems to me that it completely undermines the creative process that
    humans are good at. I first came across its use by someone I know
    'boasting' that she had run out of time and asked AI to write an article
    for a magazine on X topic, using her style and that she was delighted
    with the result and the time it saved her.

    What do people want to do with all this time they are saving?

    I'm with you; I wouldn't want to use AI for anything creative. That seems bizarre to me.

    I'm using it to speed up mundane tasks that I would do manually. Something like: If I'm in London, and I want to see a friend in Manchester, and a friend in Birmingham and a friend in Plymouth - in which sequence should I do that to minimize journey time, and should I drive or take trains, and find me a good vegan restaurant in each place. I could do all that myself, in sequence with a series of searches and a notepad...but I don't want to!
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Mon Sep 8 13:45:35 2025
    From Newsgroup: uk.media.radio.archers

    On 07/09/2025 19:23, Sam Plusnet wrote:
    On 07/09/2025 05:28, J. P. Gilliver wrote:
    On 2025/9/7 3:40:57, Steveski wrote:
    On 06/09/2025 23:29, Joe Kerr wrote:

    There is AI and there is AI, and AI (and probably AI).

    Are you Zathras?

    Then there is the AI, for which my friend's farming family kept a bull
    in case the man couldn't arrive when needed.

    One person I worked with (many years ago) would ask for "an artificial inseminator's lunch" in the local pub.-a I suppose they got used to him.

    I am curious as to what he was served, but afraid that I don't want to know.
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Mon Sep 8 14:06:37 2025
    From Newsgroup: uk.media.radio.archers

    On 08/09/2025 10:52, Jane Vernon wrote:
    On 05/09/2025 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a
    woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a
    metaphor and a simile?


    *I have never trusted or wanted AI in the world*.

    It seems to me that it completely undermines the creative process that humans are good at.-a I first came across its use by someone I know 'boasting' that she had run out of time and asked AI to write an article
    for a magazine on X topic, using her style and that she was delighted
    with the result and the time it saved her.

    What do people want to do with all this time they are saving?

    Probably porn. And xenophobia.
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Mon Sep 8 15:02:56 2025
    From Newsgroup: uk.media.radio.archers

    On 07/09/2025 20:40, J. P. Gilliver wrote:
    On 2025/9/7 16:45:19, Joe Kerr wrote:
    On 07/09/2025 05:50, J. P. Gilliver wrote:

    Sounds like we're using AI in _very_ different ways.>
    I did ask for an attractive woman with a large wooden chest and got
    images of attractive women with assorted sideboards and boxes.

    Are you winding me up? If you included the "wooden", I'm not surprised
    you got furniture!>

    It's true. I wasn't surprised either. I was testing its abilities to understand language and see how it interpreted ambiguous prompts. I have
    some very nice mystical, ethereal images that I wasn't expecting from
    some other ambiguous prompt that I unfortunately can't remember.
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Sam Plusnet@not@home.com to uk.media.radio.archers on Mon Sep 8 20:04:49 2025
    From Newsgroup: uk.media.radio.archers

    On 08/09/2025 13:45, Joe Kerr wrote:
    On 07/09/2025 19:23, Sam Plusnet wrote:
    On 07/09/2025 05:28, J. P. Gilliver wrote:
    On 2025/9/7 3:40:57, Steveski wrote:
    On 06/09/2025 23:29, Joe Kerr wrote:

    There is AI and there is AI, and AI (and probably AI).

    Are you Zathras?

    Then there is the AI, for which my friend's farming family kept a bull
    in case the man couldn't arrive when needed.

    One person I worked with (many years ago) would ask for "an artificial
    inseminator's lunch" in the local pub.-a I suppose they got used to him.

    I am curious as to what he was served, but afraid that I don't want to
    know.

    <Boring answer> Ploughman's.
    --
    Sam Plusnet
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Sam Plusnet@not@home.com to uk.media.radio.archers on Mon Sep 8 20:08:55 2025
    From Newsgroup: uk.media.radio.archers

    On 08/09/2025 10:52, Jane Vernon wrote:
    On 05/09/2025 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a
    woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a
    metaphor and a simile?


    *I have never trusted or wanted AI in the world*.

    It seems to me that it completely undermines the creative process that humans are good at.-a I first came across its use by someone I know 'boasting' that she had run out of time and asked AI to write an article
    for a magazine on X topic, using her style and that she was delighted
    with the result and the time it saved her.

    What do people want to do with all this time they are saving?

    Write Instagram posts about how they used AI in order to...

    (I know nothing about Instagram or any of its users.)
    --
    Sam Plusnet
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Mon Sep 8 21:22:41 2025
    From Newsgroup: uk.media.radio.archers

    On 08/09/2025 20:04, Sam Plusnet wrote:
    On 08/09/2025 13:45, Joe Kerr wrote:
    On 07/09/2025 19:23, Sam Plusnet wrote:
    On 07/09/2025 05:28, J. P. Gilliver wrote:
    On 2025/9/7 3:40:57, Steveski wrote:
    On 06/09/2025 23:29, Joe Kerr wrote:

    There is AI and there is AI, and AI (and probably AI).

    Are you Zathras?

    Then there is the AI, for which my friend's farming family kept a bull >>>> in case the man couldn't arrive when needed.

    One person I worked with (many years ago) would ask for "an
    artificial inseminator's lunch" in the local pub.-a I suppose they got
    used to him.

    I am curious as to what he was served, but afraid that I don't want to
    know.

    <Boring answer>-a Ploughman's.

    That is what I was expecting.
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Jane Vernon@jane@clothandclay.co.uk to uk.media.radio.archers on Tue Sep 9 13:07:03 2025
    From Newsgroup: uk.media.radio.archers

    On 08/09/2025 14:06, Joe Kerr wrote:
    On 08/09/2025 10:52, Jane Vernon wrote:
    On 05/09/2025 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a
    woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between
    a metaphor and a simile?


    *I have never trusted or wanted AI in the world*.

    It seems to me that it completely undermines the creative process that
    humans are good at.-a I first came across its use by someone I know
    'boasting' that she had run out of time and asked AI to write an
    article for a magazine on X topic, using her style and that she was
    delighted with the result and the time it saved her.

    What do people want to do with all this time they are saving?

    Probably porn.

    Oh well, fair enough.
    --
    Jane
    The Amethyst Artist
    BTM

    http://www.clothandclay.co.uk/umra/cookbook.htm - Umrats' recipes






    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Tue Sep 9 14:46:46 2025
    From Newsgroup: uk.media.radio.archers

    On 09/09/2025 13:07, Jane Vernon wrote:
    On 08/09/2025 14:06, Joe Kerr wrote:
    On 08/09/2025 10:52, Jane Vernon wrote:
    On 05/09/2025 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a
    woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between >>>> a metaphor and a simile?


    *I have never trusted or wanted AI in the world*.

    It seems to me that it completely undermines the creative process
    that humans are good at.-a I first came across its use by someone I
    know 'boasting' that she had run out of time and asked AI to write an
    article for a magazine on X topic, using her style and that she was
    delighted with the result and the time it saved her.

    What do people want to do with all this time they are saving?

    Probably porn.

    Oh well, fair enough.


    Just to be clear: That wasn't a personal prediction. I meant it as a
    general suggestion for society at large. I don't anticipate saving any
    time this way.
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From john ashby@johnashby20@yahoo.com to uk.media.radio.archers on Tue Sep 9 16:59:15 2025
    From Newsgroup: uk.media.radio.archers

    On 09/09/2025 14:46, Joe Kerr wrote:
    On 09/09/2025 13:07, Jane Vernon wrote:
    On 08/09/2025 14:06, Joe Kerr wrote:
    On 08/09/2025 10:52, Jane Vernon wrote:
    On 05/09/2025 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a
    woman asking Co-pilot to suggest a metaphor with which to begin her >>>>> presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference
    between a metaphor and a simile?


    *I have never trusted or wanted AI in the world*.

    It seems to me that it completely undermines the creative process
    that humans are good at.-a I first came across its use by someone I
    know 'boasting' that she had run out of time and asked AI to write
    an article for a magazine on X topic, using her style and that she
    was delighted with the result and the time it saved her.

    What do people want to do with all this time they are saving?

    Probably porn.

    Oh well, fair enough.


    Just to be clear: That wasn't a personal prediction. I meant it as a
    general suggestion for society at large. I don't anticipate saving any
    time this way.


    You should, often the anticipation is so much better than the practice.

    Or so I'm told.

    john
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From BrritSki@rtilbury@gmail.com to uk.media.radio.archers on Wed Sep 10 11:12:43 2025
    From Newsgroup: uk.media.radio.archers

    On 08/09/2025 10:52, Jane Vernon wrote:
    On 05/09/2025 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a
    woman asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a
    metaphor and a simile?


    *I have never trusted or wanted AI in the world*.

    It seems to me that it completely undermines the creative process that humans are good at.-a I first came across its use by someone I know 'boasting' that she had run out of time and asked AI to write an article
    for a magazine on X topic, using her style and that she was delighted
    with the result and the time it saved her.

    What do people want to do with all this time they are saving?

    When I was trying to persuade EZJet to pay me compensation for a delayed flight (umra passim) I wrote an article about it and then asked AI to
    rewrite it in the style of various newspapers - it did a remarkably good
    job, but mangled the facts in one version - I wonder of that was deliberate.

    Reader - I didn't get the compensation and still enjoy flying with them.
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Chris@chris.mcmillan@ntlworld.com to uk.media.radio.archers on Thu Sep 11 20:36:02 2025
    From Newsgroup: uk.media.radio.archers

    Kosmo <krw@whitnet.uk> wrote:
    On 5.9.25 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a woman
    asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a
    metaphor and a simile?


    Hmm. Try asking Google about The Archers and the AI answer can almost categorically be considered incorrect.

    This week I have been struggling with a streaming issue - most of the answers could have been produced by a politician in the last 20 years in that they answer a question - but not the one you are actually asking.
    It is as though AI has been sent on one of those media training courses about answering the question you wish had been asked.

    On the whole - no I do not trust it and it should be removed from Google listings.


    ItrCOs quite good at pointing one in a general direction or if one canrCOt think of a definition. What I use it or is a spill chucker and grammar checker. Sometimes my thoughts get a bit convoluted or IrCOm not sure if IrCOd got something constructed correctly. Helps me get them in order. Also sometimes IrCOm reading an interesting news item and the translation is definitely not what it should be : AI sometimes helps there - enough to
    help me untangle some totally bonkers phrase or sentence.

    Mrs McT

    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Nick Odell@nickodell49@yahoo.ca to uk.media.radio.archers on Fri Sep 12 09:25:11 2025
    From Newsgroup: uk.media.radio.archers

    On Fri, 5 Sep 2025 14:13:40 +0100, Kosmo <krw@whitnet.uk> wrote:

    On 5.9.25 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a woman
    asking Co-pilot to suggest a metaphor with which to begin her
    presentation.a She is then shows starting her presentation with a
    simile.a How much use is AI that doesn't know the difference between a
    metaphor and a simile?


    Hmm. Try asking Google about The Archers and the AI answer can almost >categorically be considered incorrect.

    This week I have been struggling with a streaming issue - most of the >answers could have been produced by a politician in the last 20 years in >that they answer a question - but not the one you are actually asking.
    It is as though AI has been sent on one of those media training courses >about answering the question you wish had been asked.

    That reminds me of an old joke which was not only true in its time but
    very prescient.

    A group of people are flying in a hot air balloon over the USA when a
    thick fog descends. They can't see anything and they have no idea
    where they are. Suddenly, a large building looms in to view. They are
    very close to the building so the balloon pilot calls out. "Hello!
    Hello? Can anybody hear me? Hello?"

    Eventually somebody opens a window. "Hello" calls the pilot. "We're
    lost. Can you tell us where we are?"

    "Sure" says the man at the window. "You're in a balloon." With which
    he closes the window again.

    "Okay," says the pilot to his passengers. "The wind is blowing from
    NNE so I reckon in about half an hour we will be over SeaTac airport
    and we can decend into an open space there."

    "But how do you know that?" asked an astonished passenger. "That guy
    didn't tell us anything."

    "Oh yes he did," replied the pilot. "I asked a clear, straightforward
    question and since he gave a clear, straightforward answer which was
    correct within itself but of absolutely no use whatsoever in our
    circumstances that could only possibly have come from Microsoft. So
    this must be the Microsoft building in Seattle and I now know where we
    are."




    On the whole - no I do not trust it and it should be removed from Google >listings.
    Wot John sed.

    Nick
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From john ashby@johnashby20@yahoo.com to uk.media.radio.archers on Fri Sep 12 10:12:32 2025
    From Newsgroup: uk.media.radio.archers

    On 12/09/2025 09:25, Nick Odell wrote:
    On Fri, 5 Sep 2025 14:13:40 +0100, Kosmo <krw@whitnet.uk> wrote:

    On 5.9.25 13:14, Jenny M Benson wrote:
    There is an add which runs on my PC quite frequently which shows a woman >>> asking Co-pilot to suggest a metaphor with which to begin her
    presentation.-a She is then shows starting her presentation with a
    simile.-a How much use is AI that doesn't know the difference between a
    metaphor and a simile?


    Hmm. Try asking Google about The Archers and the AI answer can almost
    categorically be considered incorrect.

    This week I have been struggling with a streaming issue - most of the
    answers could have been produced by a politician in the last 20 years in
    that they answer a question - but not the one you are actually asking.
    It is as though AI has been sent on one of those media training courses
    about answering the question you wish had been asked.

    That reminds me of an old joke which was not only true in its time but
    very prescient.

    A group of people are flying in a hot air balloon over the USA when a
    thick fog descends. They can't see anything and they have no idea
    where they are. Suddenly, a large building looms in to view. They are
    very close to the building so the balloon pilot calls out. "Hello!
    Hello? Can anybody hear me? Hello?"

    Eventually somebody opens a window. "Hello" calls the pilot. "We're
    lost. Can you tell us where we are?"

    "Sure" says the man at the window. "You're in a balloon." With which
    he closes the window again.

    "Okay," says the pilot to his passengers. "The wind is blowing from
    NNE so I reckon in about half an hour we will be over SeaTac airport
    and we can decend into an open space there."

    "But how do you know that?" asked an astonished passenger. "That guy
    didn't tell us anything."

    "Oh yes he did," replied the pilot. "I asked a clear, straightforward question and since he gave a clear, straightforward answer which was
    correct within itself but of absolutely no use whatsoever in our circumstances that could only possibly have come from Microsoft. So
    this must be the Microsoft building in Seattle and I now know where we
    are."




    On the whole - no I do not trust it and it should be removed from Google
    listings.
    Wot John sed.

    Nick

    I'm indebted to Hellerat for her instructional URL on how to avoid AI on Google in Firefox. A similar process should work in Chrome
    (Settings->Search Engines->Manage search engines->edit[pencil icon]
    where the URL can be set) and probably other browsers to which I do not
    have access.

    john
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Joe Kerr@joe_kerr@cheerful.com to uk.media.radio.archers on Fri Sep 12 23:09:46 2025
    From Newsgroup: uk.media.radio.archers

    On 12/09/2025 10:12, john ashby wrote:
    On 12/09/2025 09:25, Nick Odell wrote:
    On Fri, 5 Sep 2025 14:13:40 +0100, Kosmo <krw@whitnet.uk> wrote:


    On the whole - no I do not trust it and it should be removed from Google >>> listings.
    Wot John sed.

    Nick

    I'm indebted to Hellerat for her instructional URL on how to avoid AI on Google in Firefox. A similar process should work in Chrome (Settings-
    Search Engines->Manage search engines->edit[pencil icon] where the URL
    can be set) and probably other browsers to which I do not have access.

    john

    If you use duckduckgo it has settings for often, only if it found
    something useful, and on demand. It also has a link to duck.ai which
    gives you the option of 6 free AI services and 4 advanced subscription options. It is interesting to compare the output from them if you have
    the inclination (and the time).
    --
    Ric
    --- Synchronet 3.21a-Linux NewsLink 1.2
  • From Jane Vernon@jane@clothandclay.co.uk to uk.media.radio.archers on Sat Sep 13 10:22:04 2025
    From Newsgroup: uk.media.radio.archers

    On 12/09/2025 10:12, john ashby wrote:


    I'm indebted to Hellerat for her instructional URL on how to avoid AI on Google in Firefox. A similar process should work in Chrome (Settings-
    Search Engines->Manage search engines->edit[pencil icon] where the URL
    can be set) and probably other browsers to which I do not have access.

    john

    <hangs head in shame> Unfortunately, being mostly unwell still, I do
    actually use the AI info that comes up first when I google something
    (using Chrome) as it can save time and therefore energy. In my defence,
    I always check the info on another site before acting on anything,
    though. And would be happy if it wasn't there.
    --
    Jane
    The Amethyst Artist
    BTM

    http://www.clothandclay.co.uk/umra/cookbook.htm - Umrats' recipes






    --- Synchronet 3.21a-Linux NewsLink 1.2