>

Darius the Dumb has posted yet one more #veryStupidByLowIQaa article.

On 7/16/2023 8:16 PM, bitrex wrote:
> On 7/16/2023 12:13 PM, Clive Arthur wrote:
>> On 16/07/2023 16:25, bitrex wrote:
>>
>> <snip>
>>
>>> The bear, named Orion, nodded solemnly, his voice resonating deep
>>> within Tanya's soul.
>>
>> Well, if he's to be named after a constellation, it might have been
>> better to use an Ursa rather than the hunter.
>>
>> A few weeks back, I asked ChatGPT how to design a baseband OFDM
>> communication link, as that's what I've been doing for a while.  The
>> answer was of no practical use to me, nor would it have been at the
>> start of the project /however/, with just a little massaging, it would
>> have made a very good presentation to management, all the right
>> buzzwords etc and without any of that pesky detail.
>>
>
> I've tried to get it to draw ASCII schematics of e.g. "An op amp in
> inverting configuration with a gain of 1" and the results are...ah,
> creative..

I found them to be a disaster.

On Saturday, July 15, 2023 at 9:06:23 PM UTC-5, Don Y wrote:
> On 7/15/2023 5:22 PM, Dean Hoffman wrote:
> >> If an AI improved your medical care, would you campaign to ban
> >> them on the grounds that they displace doctors and other
> >> medical practitioners? Or, improved the fuel efficiency of
> >> a vehicle? Or...
> >>
> >> [I.e., does it all just boil down to "is *my* job threatened?"]
> >
> > Does it bother you that AI resurrected a Brazilian singer for a car commercial?
> > <https://www.theguardian.com/world/2023/jul/14/brazil-singer-elis-regina-artificial-intelligence-volkswagen>
> Why should it "bother" me?
>
> Would it bother you if an AI painted another Rembrandt? Even if
> you knew it wasn't done *by* the Master? Would it be, somehow, less
> artistic? Is the beauty/value in the work itself? Or, the /provenance/?

It doesn't seem right that the woman is supposedly endorsing something she might not have heard of. Let's take an extreme example. Suppose someone could get away with something like altering Martin Luther King's "I have a Dream" speech. Suppose someone could alter the play at home plate of a baseball game before it got to the viewers. Could someone get away with altering a speech by a presidential candidate?
I remember sportscaster Warner Wolf's line "Let's go to the videotape".
>
> Even if I had a personal collection of genuine articles, any
> perceived additional value they held (vs. the wannabes) would still to
> be held (among people who assign value to uniqueness). I.e., the
> owner of a "modern equivalent" could never pass his off as a
> "long lost original"...

I sometimes watch the show Pawn Stars. The host calls in experts to check things
to see if they're original. He deals in genuine articles. Someone brought in a Stradivarius violin.
Someone else brought in a poem supposedly written by Jimi Hendrix. Both were fake. It matters to people if
it's real, original and not a knockoff.
It would be good if there is some sort of marking to distinguish AI generated vs. original human made.

>
> I'm annoyed that D Adams wasn't a more prolific writer. I'd *welcome*
> anything that an AI could f* it mimicked his wit and intellect.
> (Or, would you be snobbish and avoid it out of "loyalty" to the original
> artist?)

On 7/17/2023 3:51 AM, Dean Hoffman wrote:

>> Would it bother you if an AI painted another Rembrandt? Even if you knew
>> it wasn't done *by* the Master? Would it be, somehow, less artistic? Is
>> the beauty/value in the work itself? Or, the /provenance/?
>
> It doesn't seem right that the woman is supposedly endorsing something she
> might not have heard of.

Do you *really* think all of the endorsers on TV 9social media) etc. REALLY
use the products that they are pushing? Does Joe Namath (or Tom Selleck
or Alex RIP Trebeck) have a reverse mortgage? Does Newt Gingrich use
Title Lock and actually believe it offers value?

<https://www.ksl.com/article/50334898/ads-claim-title-thieves-can-steal-your-home-but-can-you-really-lose-your-house>

Giuliani used to pitch some malware product (the guy who sees fraud everywhere
yet can't seem to prove any of it?)

> Let's take an extreme example. Suppose someone
> could get away with something like altering Martin Luther King's "I have a
> Dream" speech. Suppose someone could alter the play at home plate of a
> baseball game before it got to the viewers. Could someone get away with
> altering a speech by a presidential candidate?

That's using AI to commit fraud. I can cut and paste audio clips
(from magnetic tape) together and make it sound like you are
saying something else -- using 60 year old technology. Yet,
we haven't banned tape recorders.

[It's been possible to design a speech synthesizer that sounds
like a given person for many years, now. If I call you (at some
time of day when you are unlikely to pickup) sounding like your
wife and ask you to transfer $X from checking to some other
account, would you do so?]

> I remember sportscaster Warner Wolf's line "Let's go to the videotape".
>>
>> Even if I had a personal collection of genuine articles, any perceived
>> additional value they held (vs. the wannabes) would still to be held
>> (among people who assign value to uniqueness). I.e., the owner of a
>> "modern equivalent" could never pass his off as a "long lost original"...
>
> I sometimes watch the show Pawn Stars. The host calls in experts to check
> things to see if they're original. He deals in genuine articles. Someone
> brought in a Stradivarius violin. Someone else brought in a poem supposedly
> written by Jimi Hendrix. Both were fake. It matters to people if it's
> real, original and not a knockoff.

That's provenance. You wouldn't buy any article WHOSE VALUE LIES
ENTIRELY IN IT'S PROVENANCE without proof of same. (I've got the
original copy of the Declaration of Independence, here -- I'll let
you have it for $19.95...)

That's why such experts exist and why things like NFTs exist.

> It would be good if there is some sort of marking to distinguish AI
> generated vs. original human made.

The problem comes from The Unwashed Masses who will believe anything
peddled to them by a huckster SOUNDING genuine (covid started in a lab,
masks don't work, there are microchips in the vaccines, etc.).

There's little you can do to dissuade these people from believing
what they WANT to believe. *EVIDENCE* to the contrary!

We have all sorts of hucksters LEGALLY promoting bad ideas and
there's nothing you can really do to stop them -- if they
stay within the bounds of the law ("I didn't CLAIM this was
a medicine that would cure your terminal illness. I merely
let you vest your hope in that self-delusion -- and profitted
from YOUR *need* to believe")

Again, if an existing technology can do these things, why
shouldn't an AI be able to do them, too?

>> I'm annoyed that D Adams wasn't a more prolific writer. I'd *welcome*
>> anything that an AI could f* it mimicked his wit and intellect. (Or,
>> would you be snobbish and avoid it out of "loyalty" to the original
>> artist?)

On 7/17/2023 11:34 AM, Don Y wrote:
> That's using AI to commit fraud.  I can cut and paste audio clips
> (from magnetic tape) together and make it sound like you are
> saying something else -- using 60 year old technology.  Yet,
> we haven't banned tape recorders.
Which reminded me of an amusing anecdote...

A friend made a mix tape many years ago (college) -- for
parties. One of the tunes was Yellow Submarine (Beatles).
But, he had manually (audio tape!) edited the song so that
*one* of the choruses was:
We all live in a yellow submarine
Yellow submarine, yellow submarine, yellow submarine
This is amazingly easy to do -- just by manually positioning
the tape for dubbing.

And, few people (esp in the party atmosphere) would ever notice
the "fraud". (and, those who had an inkling that something was
wrong wouldn't be able to (easily) verify their suspicions

On 7/16/2023 7:20 PM, Don Y wrote:
> On 7/16/2023 1:12 PM, ehsjr wrote:
>> On 7/15/2023 3:56 PM, Don Y wrote:
>>> I'm trying to come to a rational/consistent opinion wrt AI
>>> and it's various, perceived "threats".
>>
>> I think the "guts" of their fear can be understood by
>> a mental experiment. (snipping to get there)
>>
>> <snip>
>>>
>>> I rely heavily on tools that are increasingly AI-driven
>>> to verify the integrity of my hardware and software designs;
>>> should they be banned/discouraged because they deprive
>>> someone (me!?) of additional billable labor hours?
>>
>> Imagine an AI that produces work indistinguishable from
>> what you produce. How would you react to the loss of
>> salary, the loss of recognition, the loss of reputation,
>> the loss of the sense of accomplishment, etc?  That's
>> what the actors face.
>
> I'd figure I have to find another line of work *or* a way
> to differentiate my "efforts" from those that the machine
> can generate.  The more "creative/original" the type of work,
> the easier it would be to make that differentiation.

With that type of (good in my opinion) thinking you can't
understand how the actors fear AI. I suppose that renders
your original question unanswerable, to you. If you can't
imagine AI producing work indistinguishable from your own,
the thought experiment fails.

Ed

>
> E.g., I would be hard-pressed to differentiate a *hole* dug
> by me from a hole dug by a machine.  OTOH, I suspect an
> original/unique timepiece that I designed would be far more
> appreciated by a "discerning buyer" than something likely
> "purely functional" (and possibly of incredible accuracy!)
> designed by machine.
>
> [Much of my career for the past several decades has been in
> creating products that haven't existed, previously.  How
> effective would an AI be at imagining solutions to incompletely
> specified needs?]
>
> It would be hypocritical to take advantage of "gains"
> (in technology, AI, etc.) when they benefit me and
> work to outlaw them when they don't, eh?
>

On 7/17/2023 11:57 AM, ehsjr wrote:
> On 7/16/2023 7:20 PM, Don Y wrote:
>> On 7/16/2023 1:12 PM, ehsjr wrote:
>>> On 7/15/2023 3:56 PM, Don Y wrote:
>>>> I'm trying to come to a rational/consistent opinion wrt AI
>>>> and it's various, perceived "threats".
>>>
>>> I think the "guts" of their fear can be understood by
>>> a mental experiment. (snipping to get there)
>>>
>>> <snip>
>>>>
>>>> I rely heavily on tools that are increasingly AI-driven
>>>> to verify the integrity of my hardware and software designs;
>>>> should they be banned/discouraged because they deprive
>>>> someone (me!?) of additional billable labor hours?
>>>
>>> Imagine an AI that produces work indistinguishable from
>>> what you produce. How would you react to the loss of
>>> salary, the loss of recognition, the loss of reputation,
>>> the loss of the sense of accomplishment, etc?  That's
>>> what the actors face.
>>
>> I'd figure I have to find another line of work *or* a way
>> to differentiate my "efforts" from those that the machine
>> can generate.  The more "creative/original" the type of work,
>> the easier it would be to make that differentiation.
>
> With that type of (good in my opinion) thinking you can't
> understand how the actors fear AI. I suppose that renders
> your original question unanswerable, to you.  If you can't
> imagine AI producing work indistinguishable from your own,
> the thought experiment fails.

In "this" industry, one is ALWAYS evolving their skillset.
You don't expect to start out doing X and finish off doing X.

I started out designing hardware in SSI/MSI, different logic
families were just different design constraints; computing
setup and hold times for worst case timing, verifying operating
limits with temperature and Vcc, ensuring hazard-free decoding,
etc.

Then, PLAs came along -- different design techniques, tools,
skills, etc. Bipolar PROMs to replace junk logic.

Then, standard cell. Then, full custom. Now, folks write VHDL
to do what I had to do "by hand". (test vector generation, anybody?)

Ditto in software. ASM and "scope loops" gave way to HLLs and
symbolic debugging. Foreground/background systems gave way to
multitasking. Then multiprocessing. Then distributed systems.

I.e., there is ALWAYS change and it's an opportunity (unless you
are a stick-in-the-mud).

I suspect the vocations that are at risk are those that don't have
such evolving skillsets. What does a TV presenter evolve into
(in the NORMAL course of their career)? Ditto accountants,
lawyers, actors, writers, artists, etc.

They can possibly adopt other media and tools. But, their basic
skillset is effectively immutable. I suspect that they know this,
subconsciously. An actor can try to move into directing/producing.
But, how many directors/producers does that industry need? An
accountant can become a CPA or CFO, etc. but, again, that's a
limited opportunity.

EVERY painter can learn to use a roller... spray gun... etc.

On Sun, 16 Jul 2023 22:41:29 -0500, John S <Sophi.2@invalid.org>
wrote:

>On 7/16/2023 8:16 PM, bitrex wrote:
>> On 7/16/2023 12:13 PM, Clive Arthur wrote:
>>> On 16/07/2023 16:25, bitrex wrote:
>>>
>>> <snip>
>>>
>>>> The bear, named Orion, nodded solemnly, his voice resonating deep
>>>> within Tanya's soul.
>>>
>>> Well, if he's to be named after a constellation, it might have been
>>> better to use an Ursa rather than the hunter.
>>>
>>> A few weeks back, I asked ChatGPT how to design a baseband OFDM
>>> communication link, as that's what I've been doing for a while.  The
>>> answer was of no practical use to me, nor would it have been at the
>>> start of the project /however/, with just a little massaging, it would
>>> have made a very good presentation to management, all the right
>>> buzzwords etc and without any of that pesky detail.
>>>
>>
>> I've tried to get it to draw ASCII schematics of e.g. "An op amp in
>> inverting configuration with a gain of 1" and the results are...ah,
>> creative..
>
>
>I found them to be a disaster.

Yes ! They look like something I might try to draw while on some
strong psychedelic drug.

boB

On Saturday, July 15, 2023 at 3:57:16 PM UTC-4, Don Y wrote:
> I'm trying to come to a rational/consistent opinion wrt AI
> and it's various, perceived "threats".
>
> I can understand how a person that can be *replaced* by an AI
> would fear for their livelihood. But, that (to me) isn't a
> blanket reason for banning/restricting AIs. (we didn't
> ban *calculators* out of fear they would "make redundant"
> folks who spent their days totaling columns of figures!
> or back hoes out of fear they would make ditch diggers
> redundant).
>
> The uproar in the "artistic" world implying that they are
> outright *stealing* their existing works seems a stretch,
> as well. If I wrote a story that sounded a hellofalot
> like one of your stories -- or painted a picture that
> resembled one of yours -- would that be "wrong"? (e.g.,
> imagine the number of variants of "A Sunday Afternoon..."
> you could come up with that would be *different* works
> yet strongly suggestive of that original -- should
> those "expressions" be banned because they weren't
> created by the original artist?
>
> How could a talking head justify his claim to "value" wrt
> an animated CGI figure making the same news presentation?

https://www.youtube.com/watch?v=cYdpOjletnc

> I rely heavily on tools that are increasingly AI-driven
> to verify the integrity of my hardware and software designs;
> should they be banned/discouraged because they deprive
> someone (me!?) of additional billable labor hours?
>
> If an AI improved your medical care, would you campaign to ban
> them on the grounds that they displace doctors and other
> medical practitioners? Or, improved the fuel efficiency of
> a vehicle? Or...
>
> [I.e., does it all just boil down to "is *my* job threatened?"]

This could happen, if 'actors' keep going out on strike:

https://www.imdb.com/title/tt0258153/

On 7/17/2023 7:06 PM, Michael Terrell wrote:
>> How could a talking head justify his claim to "value" wrt
>> an animated CGI figure making the same news presentation?
>
> https://www.youtube.com/watch?v=cYdpOjletnc

Yes, I made this reference to SWMBO and it just went "woosh",
over her head. <frown>

>> I rely heavily on tools that are increasingly AI-driven
>> to verify the integrity of my hardware and software designs;
>> should they be banned/discouraged because they deprive
>> someone (me!?) of additional billable labor hours?
>>
>> If an AI improved your medical care, would you campaign to ban
>> them on the grounds that they displace doctors and other
>> medical practitioners? Or, improved the fuel efficiency of
>> a vehicle? Or...
>>
>> [I.e., does it all just boil down to "is *my* job threatened?"]
>
> This could happen, if 'actors' keep going out on strike:
>
> https://www.imdb.com/title/tt0258153/

I don't think the "AI threat" just applies to actors, writers, etc.

A good deal of MANY jobs can be replaced by a "smart monkey"...
even moreso by a VERY smart monkey!

We already see "nurse practitioners" doing what doctors *used*
to do (though under the supervision of a doctor). Wait until
the *doctors* act under the supervision of an *AI* doctor!
(where does all that "prestige" go once YOU are delegated to
that subservient role?)

Tattoo artists? Physical therapists? Accountants?

Will we see self-driving garbage trucks? (navigate to next house;
park adjacent to trash container; activate grabber to lift
container into truck; step-and-repeat)

Mail delivery? (see above)

Think about the number of jobs that are largely static in
their skillsets and of limited promotion paths. I suspect
all will be targeted, eventually (talking heads should be
the first as they REALLY add no value! Imagine a Maxine
Headroom, Maxwell, Maximillion, etc.)

On Tuesday, July 18, 2023 at 1:34:41 AM UTC-4, Don Y wrote:
> On 7/17/2023 7:06 PM, Michael Terrell wrote:
> >> How could a talking head justify his claim to "value" wrt
> >> an animated CGI figure making the same news presentation?
> >
> > https://www.youtube.com/watch?v=cYdpOjletnc
> Yes, I made this reference to SWMBO and it just went "woosh",
> over her head. <frown>
> >> I rely heavily on tools that are increasingly AI-driven
> >> to verify the integrity of my hardware and software designs;
> >> should they be banned/discouraged because they deprive
> >> someone (me!?)
> >> If an AI improved your medical care, would you campaign to ban
> >> them on the grounds that they displace doctors and other
> >> medical practitioners? Or, improved the fuel efficiency of
> >> a vehicle? Or...

My doctor was on vacation a while back. I had to show her replacement the proper way to apply the triple layer wraps to my legs.

> >> [I.e., does it all just boil down to "is *my* job threatened?"]
> >
> > This could happen, if 'actors' keep going out on strike:
> >
> > https://www.imdb.com/title/tt0258153/
> I don't think the "AI threat" just applies to actors, writers, etc.
>
> A good deal of MANY jobs can be replaced by a "smart monkey"...
> even moreso by a VERY smart monkey!

Some could be replaced by a brain dead monkey.

> We already see "nurse practitioners" doing what doctors *used*
> to do (though under the supervision of a doctor). Wait until
> the *doctors* act under the supervision of an *AI* doctor!
> (where does all that "prestige" go once YOU are delegated to
> that subservient role?)

Like the holographic doctor on Star Trek Voyager? (BTW, it is streaming on Paramount Plus. It is free. if you use Walmart Plus.)

Nurses have traditionally done the grunt work. Sadly there are a few that weren't even good at that.I was pissing blood, so I went to the VA clinic, only to be turned away by' Nurse Rached' I told her that it was a bladder infection and asked to be tested . She threw a 'Sloman', ranting that it was obviously kidney stones, and that I couldn't possibly know what I was talking about. I made the hour long ride to a VA hospital, only to be told that I had a bladder infection, that should have been treated at my local clinic..

When I had a 'Third Nerve Palsey' in my right eye, she told be to go buy F...ing eye drops and stpp my f..ing whining. That required seven trips to the VA hospital, an MRI and over six months to heal. If I had listened to her, the damage would have been permanent. I couldn't open the eyelid, nor move the eye if I used my finger to lift it. Even two years later, they would occasionally unlock if I turned my head too fast, to look at something closer.

She refused to give me my first Glucose meter, then ranted that I wasn't keeping a log to bring to my appointments. I was 'informed' that it was impassible to use one properly without attending a class. It was sitting on her desk, so I took it, opened the box and coded it. I use the sample to verify it and then tested my blood while she was yelling at me.

A good AI would beat care like that, any day.

> Tattoo artists? Physical therapists? Accountants?
>
> Will we see self-driving garbage trucks? (navigate to next house;
> park adjacent to trash container; activate grabber to lift
> container into truck; step-and-repeat)
>
> Mail delivery? (see above)
>
> Think about the number of jobs that are largely static in
> their skillsets and of limited promotion paths. I suspect
> all will be targeted, eventually (talking heads should be
> the first as they REALLY add no value! Imagine a Maxine
> Headroom, Maxwell, Maximillion, etc.)

They are already using the 'Wind them up and watch them walk into the wall' model!

On 7/18/2023 10:31 AM, Michael Terrell wrote:
>
> My doctor was on vacation a while back. I had to show her replacement the
> proper way to apply the triple layer wraps to my legs.

You should be grateful that the replacement didn't get all "huffy"
about "being shown"! Some have egos that get in the way of good care...

[One thing I liked about my PCP (he retired recently) was that he would
always try to answer my questions -- even if it meant digging out
medical texts to pour over in the exam room (while the NEXT patient
waited! :< ) ]

>>>> [I.e., does it all just boil down to "is *my* job threatened?"]
>>>
>>> This could happen, if 'actors' keep going out on strike:
>>>
>>> https://www.imdb.com/title/tt0258153/
>> I don't think the "AI threat" just applies to actors, writers, etc.
>>
>> A good deal of MANY jobs can be replaced by a "smart monkey"... even
>> moreso by a VERY smart monkey!
>
> Some could be replaced by a brain dead monkey.

And this brings up an uncomfortable truth. What do we (as a society) "do"
with those folks who really can't contribute in the New World Order?
Surely we can't put them out to pasture. Do we *preserve* menial jobs
just to give them a place in society?

E.g., we try to accommodate Down's Syndrome kids, autistics, etc.
instead of institutionalizing them. What about folks who don't
fall in these categories but are just "too stupid" (said inelegantly)?

>> We already see "nurse practitioners" doing what doctors *used* to do
>> (though under the supervision of a doctor). Wait until the *doctors* act
>> under the supervision of an *AI* doctor! (where does all that "prestige"
>> go once YOU are delegated to that subservient role?)
>
> Like the holographic doctor on Star Trek Voyager? (BTW, it is streaming on
> Paramount Plus. It is free. if you use Walmart Plus.)

I don't think there will be much more effort than a cgi 2D image, at most.
I think people still would want to relate to a "human appearing" entity
even if it is just an embelishment over "text output".

OTOH, does a *doctor* need to see a cgi entity to accept a set of
recommendations issued by it?

> Nurses have traditionally done the grunt work.

Yeah, I had a lover many years ago who would remind me of that,
pretty regularly! :> She had a lot of disdain for many doctors
("*I* caught his MISTAKE and he ragged on me as if *I* had done
something wrong...)

> Sadly there are a few that
> weren't even good at that.I was pissing blood, so I went to the VA clinic,
> only to be turned away by' Nurse Rached' I told her that it was a bladder
> infection and asked to be tested . She threw a 'Sloman', ranting that it was
> obviously kidney stones, and that I couldn't possibly know what I was
> talking about. I made the hour long ride to a VA hospital, only to be told
> that I had a bladder infection, that should have been treated at my local
> clinic.

The consolation is that she will likely receive care from someone
equally inept, in her lifetime. *AND*, it will doubly annoy her
because she will THINK that *she* was always providing excellent care...

> When I had a 'Third Nerve Palsey' in my right eye, she told be to go buy
> F..ing eye drops and stpp my f..ing whining. That required seven trips to
> the VA hospital, an MRI and over six months to heal. If I had listened to
> her, the damage would have been permanent. I couldn't open the eyelid, nor
> move the eye if I used my finger to lift it. Even two years later, they
> would occasionally unlock if I turned my head too fast, to look at something
> closer.

Wow.

I often drive a friend to the local VA hospital as the walk from the
parking lot to the appropriate "sub-building" is quite a hike for him.
I can, instead, drive him to the door closest to his destination
and then go park the car (and walk back to where he is getting his care).

First time I did this, he gave me his handicapped tag so I could
park in one of the handicapped spaces (don't know why as *I* don't
need that).

I was stunned to find it wasn't a "set of spaces" but, rather, a
frigging parking *lot*! When I made that observation to him,
later, he said "Lots of guys here with problems..." and just
directed his eyes out the windows of the waiting room we were
in to watch the folks moving by...

[So, Thank You for Your Service}

> She refused to give me my first Glucose meter, then ranted that I wasn't
> keeping a log to bring to my appointments. I was 'informed' that it was
> impassible to use one properly without attending a class. It was sitting on
> her desk, so I took it, opened the box and coded it. I use the sample to
> verify it and then tested my blood while she was yelling at me.
>
> A good AI would beat care like that, any day.

Sorry that your experience has been so shitty. One thing I noticed
from my visits to the VA here (see above) is how polite and attentive
they have seemed to be. It was refreshing when contrasted to some of the
folks in "private practice" that don't hesitate to let you know
(and pay for!) that they are having a bad day...

>> Tattoo artists? Physical therapists? Accountants?
>>
>> Will we see self-driving garbage trucks? (navigate to next house; park
>> adjacent to trash container; activate grabber to lift container into
>> truck; step-and-repeat)
>>
>> Mail delivery? (see above)
>>
>> Think about the number of jobs that are largely static in their skillsets
>> and of limited promotion paths. I suspect all will be targeted, eventually
>> (talking heads should be the first as they REALLY add no value! Imagine a
>> Maxine Headroom, Maxwell, Maximillion, etc.)
>
> They are already using the 'Wind them up and watch them walk into the wall'
> model!

Which returns to the original question. If it's not YOUR job that's
being made redundant, then what reason NOT to exploit AIs?

On Tuesday, July 18, 2023 at 3:11:43 PM UTC-4, Don Y wrote:
> On 7/18/2023 10:31 AM, Michael Terrell wrote:
> >
> > My doctor was on vacation a while back. I had to show her replacement the
> > proper way to apply the triple layer wraps to my legs.
> You should be grateful that the replacement didn't get all "huffy"
> about "being shown"! Some have egos that get in the way of good care...

She was being trained for wound care, so she gladly accepted my help. It can look right and work OK, but there are tricks that allow you to unwrap them, instead of needing scissors to cut them off.

> [One thing I liked about my PCP (he retired recently) was that he would
> always try to answer my questions -- even if it meant digging out
> medical texts to pour over in the exam room (while the NEXT patient
> waited! :< ) ]

I had one VA doctor answer a very simple question wrong, after 30 seconds on the internet. Epsom Salt is labeled, ''Not for use by Diabetics'. He told me that it dries out your skin. That was wrong, in two ways. It was used as a laxative at one time, but it affected your electrolytes. The second error was that it softens dead skin, so it helps remove rough, dead skin. The resit is revealing healthy skin without risk of scracthes or tearing the skin because it rolls off after soaking. This was when a new, nuch larger VA clinic opened about the sae distance south of me. I was told that there was a two year waiting list, but I applied foor a transfer. Two weeks later, they transferred me to a new doctor. Sadly, she had the same ego. Both of them were from Inda, and they played their, "Patients are all low class' card to the hilt.

She didn't last long. After her, most of my doctors are Veterans who worked in Military hospitals. They show us respect, and ask if we need anything more than just a regular checkup.

> >> A good deal of MANY jobs can be replaced by a "smart monkey"... even
> >> moreso by a VERY smart monkey!
> >
> > Some could be replaced by a brain dead monkey.
> And this brings up an uncomfortable truth. What do we (as a society) "do"
> with those folks who really can't contribute in the New World Order?
> Surely we can't put them out to pasture. Do we *preserve* menial jobs
> just to give them a place in society?
>
> E.g., we try to accommodate Down's Syndrome kids, autistics, etc.
> instead of institutionalizing them. What about folks who don't
> fall in these categories but are just "too stupid" (said inelegantly)?
> >> We already see "nurse practitioners" doing what doctors *used* to do
> >> (though under the supervision of a doctor). Wait until the *doctors* act
> >> under the supervision of an *AI* doctor! (where does all that "prestige"
> >> go once YOU are delegated to that subservient role?)
> >
> > Like the holographic doctor on Star Trek Voyager? (BTW, it is streaming on
> > Paramount Plus. It is free. if you use Walmart Plus.)
> I don't think there will be much more effort than a cgi 2D image, at most..
> I think people still would want to relate to a "human appearing" entity
> even if it is just an embelishment over "text output".
>
> OTOH, does a *doctor* need to see a cgi entity to accept a set of
> recommendations issued by it?
>
> > Nurses have traditionally done the grunt work.
>
> Yeah, I had a lover many years ago who would remind me of that,
> pretty regularly! :> She had a lot of disdain for many doctors
> ("*I* caught his MISTAKE and he ragged on me as if *I* had done
> something wrong...)
>
> > Sadly there are a few that
> > weren't even good at that.I was pissing blood, so I went to the VA clinic,
> > only to be turned away by' Nurse Ratched' I told her that it was a bladder
> > infection and asked to be tested . She threw a 'Sloman', ranting that it was
> > obviously kidney stones, and that I couldn't possibly know what I was
> > talking about. I made the hour long ride to a VA hospital, only to be told
> > that I had a bladder infection, that should have been treated at my local
> > clinic.
>
> The consolation is that she will likely receive care from someone
> equally inept, in her lifetime. *AND*, it will doubly annoy her
> because she will THINK that *she* was always providing excellent care...
>
> > When I had a 'Third Nerve Palsey' in my right eye, she told be to go buy
> > F..ing eye drops and stop my f..ing whining. That required seven trips to
> > the VA hospital, an MRI and over six months to heal. If I had listened to
> > her, the damage would have been permanent. I couldn't open the eyelid, nor
> > move the eye if I used my finger to lift it. Even two years later, they
> > would occasionally unlock if I turned my head too fast, to look at something
> > closer.
> Wow.
>
> I often drive a friend to the local VA hospital as the walk from the
> parking lot to the appropriate "sub-building" is quite a hike for him.
> I can, instead, drive him to the door closest to his destination
> and then go park the car (and walk back to where he is getting his care).

They have a shuttle at the Gainesville VA hospital, since the parking lot is larger than the hospital. The Ocala CBOC just got a golf cart to take you from the from door, to your car.
> First time I did this, he gave me his handicapped tag so I could
> park in one of the handicapped spaces (don't know why as *I* don't
> need that).
>
> I was stunned to find it wasn't a "set of spaces" but, rather, a
> frigging parking *lot*! When I made that observation to him,
> later, he said "Lots of guys here with problems..." and just
> directed his eyes out the windows of the waiting room we were
> in to watch the folks moving by...
>
> [So, Thank You for Your Service}

You're welcome.

The VA system has the highest average age and percentage of disabled patients of any other provider in the United States. They also offer many services that other hospitals don''t. They also do research in fields like treating TBI. They created 'The Million Veterans Program' that requested a DNA sample, The databse will also take your medical history, to look for connections to diseases.

> > She refused to give me my first Glucose meter, then ranted that I wasn't
> > keeping a log to bring to my appointments. I was 'informed' that it was
> > impassible to use one properly without attending a class. It was sitting on
> > her desk, so I took it, opened the box and coded it. I use the sample to
> > verify it and then tested my blood while she was yelling at me.
> >
> > A good AI would beat care like that, any day.
>
> Sorry that your experience has been so shitty. One thing I noticed
> from my visits to the VA here (see above) is how polite and attentive
> they have seemed to be. It was refreshing when contrasted to some of the
> folks in "private practice" that don't hesitate to let you know
> (and pay for!) that they are having a bad day...

99% are good, but a few idiots and bullies slip though the hiring process.

> >> Tattoo artists? Physical therapists? Accountants?
> >>
> >> Will we see self-driving garbage trucks? (navigate to next house; park
> >> adjacent to trash container; activate grabber to lift container into
> >> truck; step-and-repeat)
> >>
> >> Mail delivery? (see above)
> >>
> >> Think about the number of jobs that are largely static in their skill sets
> >> and of limited promotion paths. I suspect all will be targeted, eventually
> >> (talking heads should be the first as they REALLY add no value! Imagine a
> >> Maxine Headroom, Maxwell, Maximillion, etc.)
> >
> > They are already using the 'Wind them up and watch them walk into the wall'
> > model!
> Which returns to the original question. If it's not YOUR job that's
> being made redundant, then what reason NOT to exploit AIs?

I created an Expert System' to help troubleshoot a very complex circuit board over 20 years ago. That was a low grade AI that got its input from the test fixture, then listed what needed to be checked on the monitor.

The original software simply stopped when there was a problem ,and not tel you what test had failed. On top of that, there were many errors in the test software. I politely asked the ET who built it to fix his mistakes. I was brushed off with, "I don't remember building it. Go away. Fix the damned thing,yourself." Then he bitched when I did just that. He was told, "You know better than to tell Michael that, if you don't want him to o it."

On 7/18/2023 6:43 PM, Michael Terrell wrote:

>> [One thing I liked about my PCP (he retired recently) was that he would
>> always try to answer my questions -- even if it meant digging out medical
>> texts to pour over in the exam room (while the NEXT patient waited! :< )
>> ]
>
> I had one VA doctor answer a very simple question wrong, after 30 seconds on
> the internet. Epsom Salt is labeled, ''Not for use by Diabetics'. He told me
> that it dries out your skin. That was wrong, in two ways. It was used as a
> laxative at one time, but it affected your electrolytes. The second error
> was that it softens dead skin, so it helps remove rough, dead skin. The
> resit is revealing healthy skin without risk of scracthes or tearing the
> skin because it rolls off after soaking.

Then why the warning/contraindication?

> This was when a new, nuch larger VA
> clinic opened about the sae distance south of me. I was told that there was
> a two year waiting list, but I applied foor a transfer. Two weeks later,
> they transferred me to a new doctor. Sadly, she had the same ego. Both of
> them were from Inda, and they played their, "Patients are all low class'
> card to the hilt.

Yes, when my PCP retired, an Indian couple came in to take his place.
"No thank you". My experience has been that they have an "attitude":
"I'm the doctor, you will do what I say!"

By contrast, my PCP would give *advice*. Then, we'd figure out what
I was *willing* to do ("No, let's defer the medication route and see
what I can do with dietary changes...")

He was smart enough to trust my own self-assessment (and, I left him
a back-door where he could bring up his solution at a later date if
I failed to achieve my goal)

> She didn't last long. After her, most of my doctors are Veterans who worked
> in Military hospitals. They show us respect, and ask if we need anything
> more than just a regular checkup.

My friend's sole complaint is that they have made some "recommendations"
and he's not keen on accepting that course of treatment (open heart surgery
with an estimated poor survivability). He claims that his refusal of
treatment can jeopardize his continued care (???). So, his solution is
to keep rescheduling appointments related to *that* care...

>>> When I had a 'Third Nerve Palsey' in my right eye, she told be to go
>>> buy F..ing eye drops and stop my f..ing whining. That required seven
>>> trips to the VA hospital, an MRI and over six months to heal. If I had
>>> listened to her, the damage would have been permanent. I couldn't open
>>> the eyelid, nor move the eye if I used my finger to lift it. Even two
>>> years later, they would occasionally unlock if I turned my head too
>>> fast, to look at something closer.
>> Wow.
>>
>> I often drive a friend to the local VA hospital as the walk from the
>> parking lot to the appropriate "sub-building" is quite a hike for him. I
>> can, instead, drive him to the door closest to his destination and then go
>> park the car (and walk back to where he is getting his care).
>
> They have a shuttle at the Gainesville VA hospital, since the parking lot is
> larger than the hospital. The Ocala CBOC just got a golf cart to take you
> from the from door, to your car.

This is the local VA:
<https://www.google.com/maps/@32.1811412,-110.9642396,675m/data=!3m1!1e3?entry=ttu>
My buddy gets care near the rotary at the center of the image.
Note there are only 8-12 handicap spots, there. So, you'd
have to park "out front" (the black strips that look like barracks are
PV covered parking areas) and hike to your destination, through
the building (there may be an electric shuttle, indoors -- but I've
never encountered it!)

If you are "able-bodied", it's just an annoyingly long walk
(most hospitals, here, are similar examples of sprawl).

But, if you have any health or mobility issues, it can be
brutal. When we walk BACK to the rotary, after his appointment
(I let him sit while I continue the trek to fetch the car),
he needs to stop a few times to catch his breath. Sometimes,
a passing unoccupied wheelchair may come along...

>> First time I did this, he gave me his handicapped tag so I could park in
>> one of the handicapped spaces (don't know why as *I* don't need that).
>>
>> I was stunned to find it wasn't a "set of spaces" but, rather, a frigging
>> parking *lot*! When I made that observation to him, later, he said "Lots
>> of guys here with problems..." and just directed his eyes out the windows
>> of the waiting room we were in to watch the folks moving by...
>>
>> [So, Thank You for Your Service}
>
> You're welcome.
>
> The VA system has the highest average age and percentage of disabled
> patients of any other provider in the United States.

No doubt -- and for "good" (dubious choice of words) reason!

> They also offer many
> services that other hospitals don''t. They also do research in fields like
> treating TBI. They created 'The Million Veterans Program' that requested a
> DNA sample, The databse will also take your medical history, to look for
> connections to diseases.

IIRC, Martin has made references to a similar program run by the NHS (?)

Of course, with our disjointed private medical services, there's no
central agency to coordinate such efforts -- outside of the VA.

[I installed a Reading Machine at this VA in ~1978. It was likely the
only one in all of AZ (we had only built 50, at the time)]

>>>> Think about the number of jobs that are largely static in their skill
>>>> sets and of limited promotion paths. I suspect all will be targeted,
>>>> eventually (talking heads should be the first as they REALLY add no
>>>> value! Imagine a Maxine Headroom, Maxwell, Maximillion, etc.)
>>>
>>> They are already using the 'Wind them up and watch them walk into the
>>> wall' model!
>> Which returns to the original question. If it's not YOUR job that's being
>> made redundant, then what reason NOT to exploit AIs?
>
> I created an Expert System' to help troubleshoot a very complex circuit
> board over 20 years ago. That was a low grade AI that got its input from the
> test fixture, then listed what needed to be checked on the monitor.

Expert Systems are probably the easiest AIs to "relate to". They're
intuitive -- even if complex. But, they typically don't "learn".
A knowledge engineer is responsible for crafting them, relying on
his own specific knowledge of the application domain/problem space.

I.e., it's no smarter than the person who creates it.

But, it can be very thorough -- even moreso than its creator
because humans tend to forget details whereas the AI won't.

I've tried to build most of my AIs as expert systems ("Production
Systems") that are dynamically modified by a neural net. So,
the rules can change ("learn") but, more importantly, a human
can inspect the rules, as they exist at any given point in time,
and understand WHY a particular decision was made/action taken.

> The original software simply stopped when there was a problem ,and not tel
> you what test had failed. On top of that, there were many errors in the test
> software. I politely asked the ET who built it to fix his mistakes. I was
> brushed off with, "I don't remember building it. Go away. Fix the damned
> thing,yourself." Then he bitched when I did just that. He was told, "You
> know better than to tell Michael that, if you don't want him to o it."

Letting "engineers" write code is almost always a mistake. Someone with
knowledge of the PROBLEM space needs to figure out what the code should do
and HOW it should do it as well as how it should interact with the
ACTUAL user(s).

[I can't begin to count the number of times I've "intentionally" crashed
programs to point out "unfortunate" assumptions that their writers had
made in their designs -- without even having a formal notion of the
problem that they were trying to solve! :< ]

Otherwise, you get code that walks itself into a corner or forces a user
to do X when he really wants to do Y.

[I visited a website, recently, to create an account. I was offered a
choice of authentication strategies, at one point. I opted for the 2FA
option. Then, realized I wasn't happy with any of the "second factors"
they had implemented. Ah, but there's no way to "go back" to the point
before you made that choice! And, if you log out and log back in, it
cheerfully returns you to this same point in your account configuration
process. <frown> So, abandon the account and start over again! I
wonder if they have a GC process that periodically scans the accounts
to close out those that haven't been finalized in N days...?]

On 7/18/2023 8:52 PM, Don Y wrote:
>> The original software simply stopped when there was a problem ,and not tel
>> you what test had failed. On top of that, there were many errors in the test
>> software. I politely asked the ET who built it to fix his mistakes. I was
>> brushed off with, "I don't remember building it. Go away. Fix the damned
>> thing,yourself." Then he bitched  when I did just that. He was told, "You
>> know better than to tell Michael that, if you don't want him to o it."
>
> Letting "engineers" write code is almost always a mistake.  Someone with

I should have said "design" instead of "write" though "design" is often
a misstatement of the process! <frown>

> knowledge of the PROBLEM space needs to figure out what the code should do
> and HOW it should do it as well as how it should interact with the
> ACTUAL user(s).
>
> [I can't begin to count the number of times I've "intentionally" crashed
> programs to point out "unfortunate" assumptions that their writers had
> made in their designs -- without even having a formal notion of the
> problem that they were trying to solve!  :< ]
>
> Otherwise, you get code that walks itself into a corner or forces a user
> to do X when he really wants to do Y.
>
> [I visited a website, recently, to create an account.  I was offered a
> choice of authentication strategies, at one point.  I opted for the 2FA
> option.  Then, realized I wasn't happy with any of the "second factors"
> they had implemented.  Ah, but there's no way to "go back" to the point
> before you made that choice!  And, if you log out and log back in, it
> cheerfully returns you to this same point in your account configuration
> process.  <frown>  So, abandon the account and start over again!  I
> wonder if they have a GC process that periodically scans the accounts
> to close out those that haven't been finalized in N days...?]
>
> Of course, this is true of many things -- failing to consult the
> stakeholders about their needs and IMPOSING your own notion of a
> solution.
>

On Tuesday, July 18, 2023 at 11:52:15 PM UTC-4, Don Y wrote:
> On 7/18/2023 6:43 PM, Michael Terrell wrote:
>
> >> [One thing I liked about my PCP (he retired recently) was that he would
> >> always try to answer my questions -- even if it meant digging out medical
> >> texts to pour over in the exam room (while the NEXT patient waited! :< )
> >> ]
> >
> > I had one VA doctor answer a very simple question wrong, after 30 seconds on
> > the internet. Epsom Salt is labeled, ''Not for use by Diabetics'. He told me
> > that it dries out your skin. That was wrong, in two ways. It was used as a
> > laxative at one time, but it affected your electrolytes. The second error
> > was that it softens dead skin, so it helps remove rough, dead skin. The
> > resit is revealing healthy skin without risk of scratches or tearing the
> > skin because it rolls off after soaking.
> Then why the warning/contraindication?

The warning on the package doesn't say why you shouldn't use it. I did some research to find what it does to a Diabetic's Electrolytes. With so may disabled Veterans having Diabetes, he should have know about something sold in most stores. Add a little color an perfume, then sell it to women ats bath salts. They would use it, if it dried out your skin.

> > This was when a new, nuch larger VA
> > clinic opened about the same distance south of me. I was told that there was
> > a two year waiting list, but I applied foor a transfer. Two weeks later,
> > they transferred me to a new doctor. Sadly, she had the same ego. Both of
> > them were from Inda, and they played their, "Patients are all low class'
> > card to the hilt.
> Yes, when my PCP retired, an Indian couple came in to take his place.
> "No thank you". My experience has been that they have an "attitude":
> "I'm the doctor, you will do what I say!"
>
> By contrast, my PCP would give *advice*. Then, we'd figure out what
> I was *willing* to do ("No, let's defer the medication route and see
> what I can do with dietary changes...")
>
> He was smart enough to trust my own self-assessment (and, I left him
> a back-door where he could bring up his solution at a later date if
> I failed to achieve my goal)
> > She didn't last long. After her, most of my doctors are Veterans who worked
> > in Military hospitals. They show us respect, and ask if we need anything
> > more than just a regular checkup.
> My friend's sole complaint is that they have made some "recommendations"
> and he's not keen on accepting that course of treatment (open heart surgery
> with an estimated poor survivability). He claims that his refusal of
> treatment can jeopardize his continued care (???). So, his solution is
> to keep rescheduling appointments related to *that* care...
> >>
> >> I often drive a friend to the local VA hospital as the walk from the
> >> parking lot to the appropriate "sub-building" is quite a hike for him. I
> >> can, instead, drive him to the door closest to his destination and then go
> >> park the car (and walk back to where he is getting his care).
> >
> > They have a shuttle at the Gainesville VA hospital, since the parking lot is
> > larger than the hospital. The Ocala CBOC just got a golf cart to take you
> > from the from door, to your car.
> This is the local VA:
>
> <https://www.google.com/maps/@32.1811412,-110.9642396,675m/data=!3m1!1e3?entry=ttu>

Here is one of the clinics I use. This has Wound care, the other doesnt
https://www.va.gov/north-florida-health-care/locations/ocala-va-clinic/

My PCP is at this clinic: https://www.va.gov/north-florida-health-care/locations/the-villages-va-clinic/

> My buddy gets care near the rotary at the center of the image.
> Note there are only 8-12 handicap spots, there. So, you'd
> have to park "out front" (the black strips that look like barracks are
> PV covered parking areas) and hike to your destination, through
> the building (there may be an electric shuttle, indoors -- but I've
> never encountered it!)
>
> If you are "able-bodied", it's just an annoyingly long walk
> (most hospitals, here, are similar examples of sprawl).
>
> But, if you have any health or mobility issues, it can be
> brutal. When we walk BACK to the rotary, after his appointment
> (I let him sit while I continue the trek to fetch the car),
> he needs to stop a few times to catch his breath. Sometimes,
> a passing unoccupied wheelchair may come along...
> >> First time I did this, he gave me his handicapped tag so I could park in
> >> one of the handicapped spaces (don't know why as *I* don't need that).
> >>
> >> I was stunned to find it wasn't a "set of spaces" but, rather, a frigging
> >> parking *lot*! When I made that observation to him, later, he said "Lots
> >> of guys here with problems..." and just directed his eyes out the windows
> >> of the waiting room we were in to watch the folks moving by...
> >>
> >> [So, Thank You for Your Service}
> >
> > You're welcome.
> >
> > The VA system has the highest average age and percentage of disabled
> > patients of any other provider in the United States.
> No doubt -- and for "good" (dubious choice of words) reason!
> > They also offer many
> > services that other hospitals don''t. They also do research in fields like
> > treating TBI. They created 'The Million Veterans Program' that requested a
> > DNA sample, The databse will also take your medical history, to look for
> > connections to diseases.
> IIRC, Martin has made references to a similar program run by the NHS (?)
>
> Of course, with our disjointed private medical services, there's no
> central agency to coordinate such efforts -- outside of the VA.
>
> [I installed a Reading Machine at this VA in ~1978. It was likely the
> only one in all of AZ (we had only built 50, at the time)]
> >>>> Think about the number of jobs that are largely static in their skill
> >>>> sets and of limited promotion paths. I suspect all will be targeted,
> >>>> eventually (talking heads should be the first as they REALLY add no
> >>>> value! Imagine a Maxine Headroom, Maxwell, Maximillion, etc.)
> >>>
> >>> They are already using the 'Wind them up and watch them walk into the
> >>> wall' model!
> >> Which returns to the original question. If it's not YOUR job that's being
> >> made redundant, then what reason NOT to exploit AIs?
> >
> > I created an Expert System' to help troubleshoot a very complex circuit
> > board over 20 years ago. That was a low grade AI that got its input from the
> > test fixture, then listed what needed to be checked on the monitor.
> Expert Systems are probably the easiest AIs to "relate to". They're
> intuitive -- even if complex. But, they typically don't "learn".
> A knowledge engineer is responsible for crafting them, relying on
> his own specific knowledge of the application domain/problem space.
>
> I.e., it's no smarter than the person who creates it.

I did this on a 486 computer, over 25 years ago. The original test software would read voltages but he faiiled to have the proper sign on many of them.. If it was a negative voltage, his failure to look for a negative number failed on every board. I corrected his errors, then started adding more tests, a message on the screen for the current test, then the defects as I found them. One day, the head of thee test line stormed up to my bench, yelling that I hadn't tested the board in his hand. I ran the test in front of him, and it passed. I ran it a second time, and it failed the first test. I looked at the board. A 10K resistor and a .01uF capacitor had been switch during the build. It read the right voltage, before the capacitor charged. To eliminate that, I simply copied the first test, to the end pf the software. I had the program updated before he shut up. I switched the two SMD components and sent him to the cleaning room, That board interfaced our in house embedded controller with the rest of a $20,000 radio. The controller used the MC68340. It also used a pair of Dallas 2K*8 Battery backed NVRAM to store the settings These were often a problem, because they left the factory with random data

These systems are also no smarter than the tech who runs them.

> But, it can be very thorough -- even moreso than its creator
> because humans tend to forget details whereas the AI won't.
>
> I've tried to build most of my AIs as expert systems ("Production
> Systems") that are dynamically modified by a neural net. So,
> the rules can change ("learn") but, more importantly, a human
> can inspect the rules, as they exist at any given point in time,
> and understand WHY a particular decision was made/action taken.
>
> > The original software simply stopped when there was a problem ,and not tel
> > you what test had failed. On top of that, there were many errors in the test
> > software. I politely asked the ET who built it to fix his mistakes. I was
> > brushed off with, "I don't remember building it. Go away. Fix the damned
> > thing,yourself." Then he bitched when I did just that. He was told, "You
> > know better than to tell Michael that, if you don't want him to do it."
>
> Letting "engineers" write code is almost always a mistake. Someone with
> knowledge of the PROBLEM space needs to figure out what the code should do
> and HOW it should do it as well as how it should interact with the
> ACTUAL user(s).

On 7/20/2023 12:33 AM, Michael Terrell wrote:
> Here is one of the clinics I use. This has Wound care, the other doesnt
> https://www.va.gov/north-florida-health-care/locations/ocala-va-clinic/

Wow, this is a lot smaller! Is it more like a "doctor's office".
I assume they have less staff and equipment on hand? And, that
there is probably something similar, here? Yet, my buddy goes
to the VA *hospital* likely because they have the more sophisticated
imaging equipment?

[I will have to ask him if there is another, smaller place that he also
visits or if the hospital services all veterans' needs in the area]

> My PCP is at this clinic:
> https://www.va.gov/north-florida-health-care/locations/the-villages-va-clinic/

I
>
don't think he has a PCP but, rather, different doctors for different
ailments. (I know he's complained of needing to find a PCP)

>>> I created an Expert System' to help troubleshoot a very complex circuit
>>> board over 20 years ago. That was a low grade AI that got its input from
>>> the test fixture, then listed what needed to be checked on the monitor.
>> Expert Systems are probably the easiest AIs to "relate to". They're
>> intuitive -- even if complex. But, they typically don't "learn". A
>> knowledge engineer is responsible for crafting them, relying on his own
>> specific knowledge of the application domain/problem space.
>>
>> I.e., it's no smarter than the person who creates it.
>
> I did this on a 486 computer, over 25 years ago.

A 486 was a decent sized machine (as was the 386). And, software
was "sized" more appropriately, back then. The bloat that came with
the linsux craze hadn't struck -- I ran NetBSD on a 386 with 13MB (that's
an M not a G) of RAM and a 40MB disk quite comfortably.

This explains why so many apps "hang" or "go wonky" (because they don't
catch() all errors and just let them propagate until they die)... folks
no longer understand what they are coding and its ramifications.

(my latest NetBSD INSTALL kernel is 18MB. Granted, it needs to support
as many hardware configurations as possible -- I'll build a new kernel
eliding all the unnecessary support -- but that still wouldn't fit in
physical RAM in that first 386 system mentioned above)

> The original test software
> would read voltages but he faiiled to have the proper sign on many of them.

Ooops!

> If it was a negative voltage, his failure to look for a negative number
> failed on every board. I corrected his errors, then started adding more
> tests, a message on the screen for the current test, then the defects as I
> found them. One day, the head of thee test line stormed up to my bench,
> yelling that I hadn't tested the board in his hand. I ran the test in front
> of him, and it passed. I ran it a second time, and it failed the first test.
> I looked at the board. A 10K resistor and a .01uF capacitor had been switch
> during the build. It read the right voltage, before the capacitor charged.
> To eliminate that, I simply copied the first test, to the end pf the

Yeah, one of my early POSTs had to be reengineered to accommodate a
failure in the (hardware) refresh circuitry. You obviously wanted to
run the test as quickly as possible (because the system can't boot to
the point of being responsive until then) so you could be back *checking*
written values faster than the refresh interval and everything looks good!

[I have a similar potential problem with "live" testing of released memory
pages in my VMM system... you want to check them quickly so you aren't
wasting MIPS *and* so you can get them returned to the scrubbed page pool
for reuse but need to make sure you've given the page "time" to noticeably
fail]

> software. I had the program updated before he shut up. I switched the two
> SMD components and sent him to the cleaning room, That board interfaced our
> in house embedded controller with the rest of a $20,000 radio. The
> controller used the MC68340. It also used a pair of Dallas 2K*8 Battery
> backed NVRAM to store the settings These were often a problem, because they
> left the factory with random data

NVRAM makes for some delightful thinking regarding POST/BIST/RESET, etc.
You don't want to scrub it until you KNOW that it has nothing in it.
So, you need to assure yourself (ideally without asking a human) that
it is *intended* to CURRENTLY have valid contents -- which you must then
check, correct and test. And, barring that, be able to set it's contents
to a "sane" state so the software doesn't act on bogus values.

> These systems are also no smarter than the tech who runs them.

They can be made so -- if that is the intent. But, you have to
think about how you interface to the user/tech and what you expect
of him. E.g., the "check engine" syndrome is often a lazy approach
to error/fault handling ("Well, what am I supposed to be CHECKING?!")

>>> The original software simply stopped when there was a problem ,and not
>>> tel you what test had failed. On top of that, there were many errors in
>>> the test software. I politely asked the ET who built it to fix his
>>> mistakes. I was brushed off with, "I don't remember building it. Go
>>> away. Fix the damned thing,yourself." Then he bitched when I did just
>>> that. He was told, "You know better than to tell Michael that, if you
>>> don't want him to do it."
>>
>> Letting "engineers" write code is almost always a mistake. Someone with
>> knowledge of the PROBLEM space needs to figure out what the code should
>> do and HOW it should do it as well as how it should interact with the
>> ACTUAL user(s).
>
> Most engineers should never be allowed to solder, as well.

Engineers often think they are more skilled/knowledgeable than they actually
are. "It's just soldering".

I let the folks on the line do all of that work for me. First, they are
more efficient at it. But, they also will follow-up with a detailed
inspection (whereas I may just give that sort of work a cursory looking over).

> I have had similar results trying to get a new cell phone activated. I got a
> message that setup was completed. It then deactivated the old phone, but
> didn't activate the new phone. Now, I'm supposed to call support, without a
> phone to fix their mess.

The last step of the process should have "closed the loop": "Please dial
XYZ on your new phone to complete the process...". As such, if any of
the preceding steps had gone awry and NOT BEEN DETECTED AS HAVING DONE SO,
there's be no way to know, for sure.

> I tried to use the chat , but the jerk kept telling
> me that he wasn't receiving the information that I posted over 20 time, but
> he replied to every other message. I'm about ready to take it out to the
> driveway and crush it. It was sent to replace a non 5G phone, after Verizon
> bought the company, so I didn't pay have for it. If I do that, I'll lose my
> long time phone number.

Aren't cell phones wonderful? So many more times you NEED to interact
with the Phone Company than you had to previously! <grin>

We give out the home phone number when we need to. And everyone wants to
send you an SMS (for authentication, to let you know when your dinner
reservation is ready, to...).

"Oh, we don't get texts..."
"Huh?"
"It's a land line. But, you can CALL us now that you've insisted on
having a contact number..."

[We have disabled SMS on the cell phones as well. And, voice mail.
They exist for *our* convenience, not yours! So, if *we* want to make
(or receive) a call, we will take steps to do so. But, if YOU think
you can exploit them to gain access to us -- inconvenience us -- at
YOUR convenience, you're going to be disappointed! <grin> Sheesh,
all these people jumping to react to each SMS, phone, etc. "alert".
I'm sure Pavlov would be proud of how well you have CONDITIONED
YOURSELF! :> ]

>> Of course, this is true of many things -- failing to consult the
>> stakeholders about their needs and IMPOSING your own notion of a
>> solution.
>
> Like when I asked for a computer to store service information, back in the
> '80s. Instead of spending a few hundred for a C64 and drive, they dragged in
> a beat up Atari 850(?) It had some crap ware database, but it took several
> minutes every time you added notes. The drive soulded like a cement mixer,
> and it frequently tried to defrag the floppy witch took at least 15
> minutes.

I worked with a group that relied on a database that was limited to ~16K
records. Exceeding that limit caused the DBMS to lock up which complicated
efforts to elide contents to get it down below that limit. So, they had
to routinely check to see how many records (rows) they were using and
"edit out" records that they didn't deem essential to avoid that problem.

On Thursday, July 20, 2023 at 11:13:09 AM UTC-4, Don Y wrote:
> On 7/20/2023 12:33 AM, Michael Terrell wrote:
> > Here is one of the clinics I use. This has Wound care, the other doesnt
> > https://www.va.gov/north-florida-health-care/locations/ocala-va-clinic/
> Wow, this is a lot smaller! Is it more like a "doctor's office".
> I assume they have less staff and equipment on hand? And, that
> there is probably something similar, here? Yet, my buddy goes
> to the VA *hospital* likely because they have the more sophisticated
> imaging equipment?
>
> [I will have to ask him if there is another, smaller place that he also
> visits or if the hospital services all veterans' needs in the area]
> > My PCP is at this clinic:
> > https://www.va.gov/north-florida-health-care/locations/the-villages-va-clinic/

This area has a very high percentage of retired Veterans year round, along with Snowbirds.
> I don't think he has a PCP but, rather, different doctors for different
> ailments. (I know he's complained of needing to find a PCP)

Some areas have a long waiting list because of a shortage of doctors, or in some cases they have a handful of patients who need one, but not enough to be allowed to hire another doctor. When I went into the system, you pretty much had to wait for other Veterans to die, as your name moved up the list with each.

This was caused by Bill Clinton, when he opened the system to more Veterans, without adding staff or additional clinics.
They finally updated the VA Hospital in Gainesville. They doubled the number of patient rooms and added some services, but other services are still located at nearby sites. When that got their MRI system, it was installed away from the hospital itself. There is a tunnel under the highway, to the Shands teaching hospital. The MRI equipment was built to open into the tunnel, and it has a removable roof to install or replace equipment.

Gainesville has many tiny CBOC around the area It is a very busy college town and with this being Florida there are a lot of homeless Veterans who move here from other states. This way, they can provide basic care to those with no transportation, or transport them to the hospital, if needed.

> >>> I created an Expert System' to help troubleshoot a very complex circuit
> >>> board over 20 years ago. That was a low grade AI that got its input from
> >>> the test fixture, then listed what needed to be checked on the monitor.
> >> Expert Systems are probably the easiest AIs to "relate to". They're
> >> intuitive -- even if complex. But, they typically don't "learn". A
> >> knowledge engineer is responsible for crafting them, relying on his own
> >> specific knowledge of the application domain/problem space.
> >>
> >> I.e., it's no smarter than the person who creates it.
> >
> > I did this on a 486 computer, over 25 years ago.
> A 486 was a decent sized machine (as was the 386). And, software
> was "sized" more appropriately, back then. The bloat that came with
> the linsux craze hadn't struck -- I ran NetBSD on a 386 with 13MB (that's
> an M not a G) of RAM and a 40MB disk quite comfortably.

This had 4 MB, and was picked up used at a thrift store for $30. The code was in MS's Basic. which limited what could be done .I had no access o a compiler or other language, so I had to make due.

The problem with some techs is that they can't parse even simple instructions on the screen. You have to explain thins to them, and answer questions for each instruction. This isn't common, but it happens. For instance, I rebuilt a test fixture and added a 10 pole, 12 position rotary switch to simplify testing. Start at one, and do a test, and continue. There were three terminals for a DVM. One was ground, one was positive and the other negative.. I had one idiot who could follow 'Connect meter to + and Ground' or any other configuration when there were only three. No one else had a problem, but he had an attitude like Sloman in thart everyone else was always wrong. It tested the wiring for the front panel of a Telemetry receiver.

> This explains why so many apps "hang" or "go wonky" (because they don't
> catch() all errors and just let them propagate until they die)... folks
> no longer understand what they are coding and its ramifications.

Many have the attitude that their code should be the only thing running orher than an OS.

> (my latest NetBSD INSTALL kernel is 18MB. Granted, it needs to support
> as many hardware configurations as possible -- I'll build a new kernel
> eliding all the unnecessary support -- but that still wouldn't fit in
> physical RAM in that first 386 system mentioned above)
> > The original test software
> > would read voltages but he failed to have the proper sign on many of them.
>
> Ooops!
>
> > If it was a negative voltage, his failure to look for a negative number
> > failed on every board. I corrected his errors, then started adding more
> > tests, a message on the screen for the current test, then the defects as I
> > found them. One day, the head of thee test line stormed up to my bench,
> > yelling that I hadn't tested the board in his hand. I ran the test in front
> > of him, and it passed. I ran it a second time, and it failed the first test.
> > I looked at the board. A 10K resistor and a .01uF capacitor had been switch
> > during the build. It read the right voltage, before the capacitor charged.
> > To eliminate that, I simply copied the first test, to the end pf the
>
> Yeah, one of my early POSTs had to be reengineered to accommodate a
> failure in the (hardware) refresh circuitry. You obviously wanted to
> run the test as quickly as possible (because the system can't boot to
> the point of being responsive until then) so you could be back *checking*
> written values faster than the refresh interval and everything looks good!
>
> [I have a similar potential problem with "live" testing of released memory
> pages in my VMM system... you want to check them quickly so you aren't
> wasting MIPS *and* so you can get them returned to the scrubbed page pool
> for reuse but need to make sure you've given the page "time" to noticeably
> fail]

It's only going to get worse. I've had numerous programmers either tell me, It's no my problem, I finished that job" or "I don't even remember writing that program".

> > software. I had the program updated before he shut up. I switched the two
> > SMD components and sent him to the cleaning room, That board interfaced our
> > in house embedded controller with the rest of a $20,000 radio. The
> > controller used the MC68340. It also used a pair of Dallas 2K*8 Battery
> > backed NVRAM to store the settings These were often a problem, because they
> > left the factory with random data.
> NVRAM makes for some delightful thinking regarding POST/BIST/RESET, etc.
> You don't want to scrub it until you KNOW that it has nothing in it.
> So, you need to assure yourself (ideally without asking a human) that
> it is *intended* to CURRENTLY have valid contents -- which you must then
> check, correct and test. And, barring that, be able to set it's contents
> to a "sane" state so the software doesn't act on bogus values.
> > These systems are also no smarter than the tech who runs them.
> They can be made so -- if that is the intent. But, you have to
> think about how you interface to the user/tech and what you expect
> of him. E.g., the "check engine" syndrome is often a lazy approach
> to error/fault handling ("Well, what am I supposed to be CHECKING?!")

The problem we had with it was sometimes the garbage data from the factory would keeep a board from booting, because it was out of range. The quick fix was to unplug both, and switch positions. In a few cases, you had to replace one. I was able to erase them on one of our oddball EPROM programmers that did 5V only 2716 chips.

> >>
> >> Letting "engineers" write code is almost always a mistake. Someone with
> >> knowledge of the PROBLEM space needs to figure out what the code should
> >> do and HOW it should do it as well as how it should interact with the
> >> ACTUAL user(s).
> >
> > Most engineers should never be allowed to solder, as well.
>
> Engineers often think they are more skilled/knowledgeable than they actually
> are. "It's just soldering".

I did my own rework. I had several Ungar Loner irons. Two with small chisel tips, and the thrd had a special 0.015" tip to solder the MC68340 chips, and others with that density.

When I worked at a Defense Plant in the late '70s, I was doing QA on the PRC77 Manpack Radio. Soldering was a union job, but they ran into a job they couldn't do. Their supervisor knew that I had a business of my own, so he asked to borrow me. The Karen of a Union Rep was glaring at me. The idea of someone from mud level management teaching 'her girls' to solder pissed her off! The problem was obvious. That company used the old Weller WTPC soldering stations, and they simply didn't produce enough heat to melt the solder on a bad audio output transformer. I used two irons and quickly removed the solder, then again to solder in the new transformer. Just to make her eve madder, I reached for the second broad, "Ladies? I'm only going to sow you this one more time!" She was about to explode. She had 'informed me' on my first day that no man knew how to solder. She had to eat her words, twice..

On 7/20/2023 8:18 PM, Michael Terrell wrote:
>> [I will have to ask him if there is another, smaller place that he also
>> visits or if the hospital services all veterans' needs in the area]
>>> My PCP is at this clinic:
>>> https://www.va.gov/north-florida-health-care/locations/the-villages-va-clinic/
>
>>>
> This area has a very high percentage of retired Veterans year round, along
> with Snowbirds.

Ditto here. Plus three active bases. (total state population is only ~6M)

>> [I have a similar potential problem with "live" testing of released
>> memory pages in my VMM system... you want to check them quickly so you
>> aren't wasting MIPS *and* so you can get them returned to the scrubbed
>> page pool for reuse but need to make sure you've given the page "time" to
>> noticeably fail]
>
> It's only going to get worse. I've had numerous programmers either tell me,
> It's no my problem, I finished that job" or "I don't even remember writing
> that program".

When you sign a contract to develop a product, "support" is something
that is legally enforceable. If you *don't* remember what you did,
it will be *you* who pays for it when/if a problem arises.

So, it behooves you to do things in ways that "make sense" and document
the reasons behind your design choices so they "make sense" when you
revisit it months/years later.

>>> These systems are also no smarter than the tech who runs them.
>> They can be made so -- if that is the intent. But, you have to think about
>> how you interface to the user/tech and what you expect of him. E.g., the
>> "check engine" syndrome is often a lazy approach to error/fault handling
>> ("Well, what am I supposed to be CHECKING?!")
>
> The problem we had with it was sometimes the garbage data from the factory
> would keeep a board from booting, because it was out of range.

But a sanity check (and integrity check) should be the first thing
that happens before the data is used.

I break my nonvolatile parameters into many groups and implement
checks on each group. So, I can hope that the "most important"
settings are preserved even if some of the less important ones
aren't.

And, that a glitch in one of the less important settings doesn't
invalidate the more important settings (requiring them to be
reset).

It's possible (though unlikely with well designed algorithms) that
the NVRAM can randomly power up in a state that *seems* intact.
And, there's nothing I can do about that.

But, even after a group of parameters indicates they *appear* to
be intact, I can examine individual parameters to see if they
make sense in the context of the other parameters. Just by
verifying any rules that apply to the parameter *set*.

E.g., a netmask of 0x000fff0f is probably bogus!

> The quick fix
> was to unplug both, and switch positions. In a few cases, you had to replace
> one. I was able to erase them on one of our oddball EPROM programmers that
> did 5V only 2716 chips.

>>>> Letting "engineers" write code is almost always a mistake. Someone
>>>> with knowledge of the PROBLEM space needs to figure out what the code
>>>> should do and HOW it should do it as well as how it should interact
>>>> with the ACTUAL user(s).
>>>
>>> Most engineers should never be allowed to solder, as well.
>>
>> Engineers often think they are more skilled/knowledgeable than they
>> actually are. "It's just soldering".
>
> I did my own rework. I had several Ungar Loner irons. Two with small chisel
> tips, and the thrd had a special 0.015" tip to solder the MC68340 chips, and
> others with that density.

For thru-hole designs, I would knock up prototypes as it was quicker
than putting together a kit of parts to send to someone else.

But, for SMT designs (esp components top and bottom), I'd prefer
someone with better eyes and steadier hands... and, the "real"
inspection followup instead of just doing a quick visual.

>> The last step of the process should have "closed the loop": "Please dial
>> XYZ on your new phone to complete the process...". As such, if any of the
>> preceding steps had gone awry and NOT BEEN DETECTED AS HAVING DONE SO,
>> there's be no way to know, for sure.
>
> Instead, you are instructed to turn the phone off, then reboot it and
> attempt to make a call. If it desn't work, wait and try it again.

Yeah, what is this nonsense about rebooting everything? Blame MS
for getting people to consider that "normal". I never wrote "if
things get hosed, unplug power for 60 seconds..." in a manual.

Why *should* they "get hosed"?

> Another phone setup got a server error, but the phone had a number assigned,
> so I can't try again, even though it was never activated.

> Motorola really pissed me ff. You can't use an Email address with a period
> in your name on some models.

You can't use a colon in a filename with MS. Or a backslash. Or...

Interesting to see how folks place arbitrary restrictions on things.

"Um, why?"

>>> I tried to use the chat , but the jerk kept telling me that he wasn't
>>> receiving the information that I posted over 20 time, but he replied to
>>> every other message. I'm about ready to take it out to the driveway and
>>> crush it. It was sent to replace a non 5G phone, after Verizon bought
>>> the company, so I didn't pay have for it. If I do that, I'll lose my
>>> long time phone number.
>> Aren't cell phones wonderful? So many more times you NEED to interact with
>> the Phone Company than you had to previously! <grin>
>
> This is what happens when Itinerant burger flippers move into tech support.

In their defense, it has got to be a nightmare dealing with Joe Randoms
complaining about things that don't work and not being able to explain
EXACTLY what they are doing to *create* a problem.

Likewise, keeping machines configured properly, current on updates, etc.
I spend at least a day a week on this sort of crap. (and, it's never DONE;
change X and Y needs to be tweaked, etc.)

>> [We have disabled SMS on the cell phones as well. And, voice mail. They
>> exist for *our* convenience, not yours! So, if *we* want to make (or
>> receive) a call, we will take steps to do so. But, if YOU think you can
>> exploit them to gain access to us -- inconvenience us -- at YOUR
>> convenience, you're going to be disappointed! <grin> Sheesh, all these
>> people jumping to react to each SMS, phone, etc. "alert". I'm sure Pavlov
>> would be proud of how well you have CONDITIONED YOURSELF! :> ]
>
> I'm still a night owl, after working night shifts for decades. I've had
> calls at 0500, begging for a donation to some liberal politician, or
> another. I have a long list of DC area phone numbers that have tried to call
> that Magic Jack for money. It can receive voice mail or text, but it cant
> uplink voice without chopping it into a string of broken bits and gaps of
> silence.

I designed a box to screen calls so we only tend to hear the phone ring
when its someone we want/need to talk to (and then, only when we are
likely WANTING (or willing) to talk to them! I.e., the phone only rings
for a very select few people after 9PM -- even fewer after midnight!)

[I can't do that for SMS so its easier to just disable that service]

On 2023-07-18, Michael Terrell <terrell.michael.a@gmail.com> wrote:
> On Tuesday, July 18, 2023 at 1:34:41 AM UTC-4, Don Y wrote:
>> On 7/17/2023 7:06 PM, Michael Terrell wrote:
>> >> How could a talking head justify his claim to "value" wrt
>> >> an animated CGI figure making the same news presentation?
>> >
>> > https://www.youtube.com/watch?v=cYdpOjletnc
>> Yes, I made this reference to SWMBO and it just went "woosh",
>> over her head. <frown>
>> >> I rely heavily on tools that are increasingly AI-driven
>> >> to verify the integrity of my hardware and software designs;
>> >> should they be banned/discouraged because they deprive
>> >> someone (me!?)
>> >> If an AI improved your medical care, would you campaign to ban
>> >> them on the grounds that they displace doctors and other
>> >> medical practitioners? Or, improved the fuel efficiency of
>> >> a vehicle? Or...
>
> My doctor was on vacation a while back. I had to show her replacement the proper way to apply the triple layer wraps to my legs.
>
>> >> [I.e., does it all just boil down to "is *my* job threatened?"]
>> >
>> > This could happen, if 'actors' keep going out on strike:
>> >
>> > https://www.imdb.com/title/tt0258153/
>> I don't think the "AI threat" just applies to actors, writers, etc.
>>
>> A good deal of MANY jobs can be replaced by a "smart monkey"...
>> even moreso by a VERY smart monkey!
>
>
> Some could be replaced by a brain dead monkey.
>
>
>> We already see "nurse practitioners" doing what doctors *used*
>> to do (though under the supervision of a doctor). Wait until
>> the *doctors* act under the supervision of an *AI* doctor!
>> (where does all that "prestige" go once YOU are delegated to
>> that subservient role?)
>
>
> Like the holographic doctor on Star Trek Voyager? (BTW, it is streaming on Paramount Plus. It is free. if you use Walmart Plus.)
>
> Nurses have traditionally done the grunt work. Sadly there are a few that weren't even good at that.I was pissing blood, so I went to the VA clinic, only to be turned away by' Nurse Rached' I told her that it was a bladder infection and asked to be tested . She threw a 'Sloman', ranting that it was obviously kidney stones, and that I couldn't possibly know what I was talking about. I made the hour long ride to a VA hospital, only to be told that I had a bladder infection, that should have been treated at my local clinic.
>
> When I had a 'Third Nerve Palsey' in my right eye, she told be to go buy F..ing eye drops and stpp my f..ing whining. That required seven trips to the VA hospital, an MRI and over six months to heal. If I had listened to her, the damage would have been permanent. I couldn't open the eyelid, nor move the eye if I used my finger to lift it. Even two years later, they would occasionally unlock if I turned my head too fast, to look at something closer.
>
> She refused to give me my first Glucose meter, then ranted that I wasn't keeping a log to bring to my appointments. I was 'informed' that it was impassible to use one properly without attending a class. It was sitting on her desk, so I took it, opened the box and coded it. I use the sample to verify it and then tested my blood while she was yelling at me.
>
> A good AI would beat care like that, any day.

Who holds the gold makes the rules. the AI you are likely to get will
be designed to reduce costs, not to improve service.

--
Jasen.
🇺🇦 Слава Україні