ChatGPT says that that it is a few years out of date often when you ask
a "political" type question.
So I have been experimenting with Google's AI - Bard.Google.com. And I
like it.
Recently, I have asked it about some programming. Like "please code
this:" It isn't always accurate, however if often gives a different way
of looking at a problem.
For example, based on the 40 year old Jol Panel instruction, I have
been developing a universal Panel or Form instruction that works in
MVS/Zos, Windows and Linux. I asked Bard about setting environment
variables at the CURRENT level, and it suggested stuffing the keyboard
with SET statements. And it sort of worked, except it wouldn't work in
an Windows BAT file - the keyboard and the next statement in the BAT
file got all tangled up! Ce la vie!
Anyway, it is all happening - although somewhat slowly. Where, for
example, do you store the results of a form so that a Clist can examine
the results? Or can a Rexx program use the Panel? And so on!
Clem
Dean Kent wrote:
Which brings up another 'interesting' anecdote. I used chatGPT to
'write' a set of bylaws for a new non-profit for a youth sports
club. I asked it over a dozen times with different wording, and it
came back with a wide variety of results - some that were long and
included many sections, and others that were short and included only
what might be considered 'necessary' sections. I have a friend who
is an attorney, and he sent me the template that LexisNexis
provides. It had a lot more content, some which the IRS now
indicates is 'preferred' or even required. So I ended up using the
template since I could just remove or ignore sections that weren't
pertinent. The template also had a variety of options (variables, if
you will) for wording depending on, for example, if the corporation
has a CEO or President and whether board officers can also be
corporate executives, etc.
That caused me to make the decision that I would not consider using
chatGPT for creating legal documents. Again, YMMV.
On 9/5/2023 12:27 PM, Steve Thompson wrote:
And so we can now understand that when a paralegal or newly minted
attorney uses it to find case law for points and authorities, it will
will make them up to match what was being searched for when it
prepares a motion it was asked for using the results of the search.
And some attorneys got a judge quite angry with them when they didn't
tell the court this, but the opposing council pointed out they could
not find any such case listed in the pleadings/motion. Then the
judge's people also could not find same.... This is the kind of thing
that concerns me about AI today. Once it has been taught enough to
learn on its own....
Steve Thompson
On 9/5/2023 12:46 PM, Dean Kent wrote:
I spent a bit of time playing with chatGPT to see what it could
do. So did my two sons - one an MS in biotech, the other a PhD in
theoretical physics. We all came to the same conclusion - chatGPT
is a very, very good Google search that can filter many different
possible 'answers' and come to one that is 'most likely' based on
various factors. It has little to no creativity or understanding of
what it is asked to do. Not surprising, but different than what
the popular press seems to say about it.
One of my questions was to write a simple sort routine in HLASM. It
came back with a template containing the entry/exit code, and then a
comment *insert sort routine here*. After doing that with many
different simple tasks, I came to the conclusion that the problem
chatGPT has with assembler (but not with C, Python, Java, etc.) is
that there are so few searchable examples of code in assembler.
So the quality of the results, for any question, depends upon what
exists out on the Internet. Again, not surprising.
As another example, I have an interest in what is called 'historical
analysis'. There are a number of books on the subject, so I asked
chatGPT to compare/contrast two of the books. Then two other
books, etc. In literally every case it came back with the same
introductory text and conclusion - but inserted a couple of
paragraphs that was similar to a book review for each book and
compared the 'differences'. Not very impressed.
My PhD son uses it to find obscure hypotheses and formulas that
would otherwise require a great many hours (or days) of searching.
My MS son uses it in a similar fashion to ferret out alternative
options for the various cell growing and protein extraction for his
job. A very useful tool, but not yet SkyNet...
YMMV.
On 9/5/2023 9:36 AM, Bill Johnson wrote:
We are all retired. The other 2 went before me. I went in
July 2022. You’re an idiot regardless. What are you afraid of? That
a computer can do what you do? That your “skills” aren’t all that
impressive and can be automated away?
Sent from Yahoo Mail for iPhone
On Tuesday, September 5, 2023, 12:25 PM, David Spiegel
<00000468385049d1-dmarc-requ...@listserv.ua.edu> wrote:
Hi Bill,
I have a better idea.
Why don't you and the 2 buddies who helped you modify the IEFUSI
fix it?
Probably because you don't have the wherewithal (even with 2 helpers).
Regards,
David
On 2023-09-05 12:04, Bill Johnson wrote:
Lol, how about going to chatgpt and asking the same question. So
that cut and paste isn’t a factor.
Sent from Yahoo Mail for iPhone
On Tuesday, September 5, 2023, 12:02 PM, David Spiegel
<00000468385049d1-dmarc-requ...@listserv.ua.edu> wrote:
Hi Steve,
It won't. The first executable statement is missing a comma between
operands.
Regards,
David
On 2023-09-05 11:43, Steve Thompson wrote:
I doubt it will assemble. And even if it does, the results are
unpredictable, other than it will probably ABEND for one reason or
another.
There are no DCB, OPEN, CLOSE macros while GET and PUT are being
used.
Me thinks this AI system is confusing a few different assembly
languages together. I wonder how close they came for DOS I/O.
Steve Thompson
On 9/5/2023 11:20 AM, Tom Marchant wrote:
You're right, Tom. That is not a program. Certainly not one that
will
do what it claims to do.
-- Tom Marchant On Mon, 4 Sep 2023 10:42:51 -0700, Tom Brennan
<t...@tombrennansoftware.com> wrote:
I can't be sure I formatted it properly, but after looking over
the
code, I have nothing to say but WTF? 😄
PRINT NOGEN
TITLE 'Simple Addition Program'
** Define storage for input numbers and result
*
NUM1 DS F First input number
NUM2 DS F Second input number
RESULT DS F Result of addition
** Main program
*
MAIN C 0 NUM1 Check if NUM1 is zero
BE ZERO Branch to ZERO if true
** Read the first number from input
*
GET NUM1,NUMIN Read NUM1 from input
LA 0,NUM1 Load NUM1 into register
** Read the second number from input
*
GET NUM2,NUMIN Read NUM2 from input
A NUM1,NUM2 Add NUM1 and NUM2
ST NUM1,RESULT Store the result in RESULT
** Print the result
*
PUT RESULT,NUMOUT Print the result
** Terminate the program
*
SR 15,15 Set return code to 0
BR 14 Return to caller
** Define input and output areas
*
NUMIN DC F'0' Input buffer for numbers
NUMOUT DC F'0' Output buffer for result
ZERO DC F'0' Constant zero
END MAIN End of program
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email tolists...@listserv.ua.edu with the message: INFO
IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO
IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO
IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO
IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to lists...@listserv.ua.edu with the message: INFO IBM-MAIN
