Volume 66 Number 38
Produced: Thu, 18 May 23 16:45:45 -0400
Subjects Discussed In This Issue:
AI (2)
[Michael Poppers Keith Bierman]
Bracha rishona (2)
[Micha Berger Steven Oppenheimer]
First to minyan? (3)
[Micha Berger Carl Singer Menashe Elyashiv]
Nusach hatfila (2)
[David Ziants Menashe Elyashiv]
Shavuot Second Day on Shabbat in Chutz La'Aretz
[Yisrael Medad]
Shower shiur
[Yisrael Medad]
----------------------------------------------------------------------
From: Michael Poppers <the65pops@...>
Date: Wed, May 17,2023 at 10:17 PM
Subject: AI
In response to Joel Rich (MJ 66#37):
Before we talk about sentience, free will, or consciousness -- much less
"Heavenly assistance," "prophetic spirit," or "intuitive Torah knowledge" (my
1st-draft attempts at translating siyata dishmaya, ruach haqodesh, and da'as
Torah, respectively) -- shouldn't a p'saq (Halachic decision-making authority)
prerequisite be having a n'shamah (soul)?
Michael Poppers
Elizabeth, NJ, USA
----------------------------------------------------------------------
From: Keith Bierman <khbkhb@...>
Date: Thu, May 18,2023 at 02:17 AM
Subject: AI
Joel Rich wrote (MJ 66#37):
> There is certainly a permanent spiritual argument that artificial
> intelligence could never becoming a poseik...
I think we need to sort out just what we're talking about with respect to AI,
before we can sensibly discuss application to halachic analysis.
The hot topic these days are "Large Language Models" (LLMs) and for a detailed,
but not overly technical presentation, Prof Bender's paper on Stochastic Parrots
https://dl.acm.org/doi/pdf/10.1145/3442188.3445922
is a good starting point. For the technically inclined, a little googling can
turn up various rehashes of the basic algorithm (about 300 lines of Fortran, for
example!). To summarize, this approach is purely statistical, the algorithm
determines the most likely next word. Truth and logic are not involved.However,
just as we anthropomorphize our pets, we project depth and understanding into
the seemingly intelligent pronouncements of these computerized "stochastic parrots".
More generally, I think people broadly fall into two camps. Those who are AI
pessimists and do not believe that anything we produce can be actually
intelligent. These people (like Prof Bender) may say "While we don't fully
understand intelligence, we certainly can recognize what it isn't".
The AI "optimists" are represented by Dr. Steven Wolfram (ChatGPT what is
it doing)
https://writings.stephenwolfram.com/2017/05/a-new-kind-of-science-a-15-year-view/
which provides much detail about how such algorithms work). He has long held
that there is a new kind of science (book) where small simple algorithms running
for long periods of time can produce profound complexity (think Conway's "game
of life"). Combining LLM alpha+?? might actually produce general intelligence
(according to this line of thought). "alpha" understands symbolic logic and a
lot of symbolic mathematics so combined with a language model ... perhaps that
is enough (or another few other "simple" bits composed in the flow would do the
trick).
A third approach (beloved in science fiction circles) is one where our biologic
brains are scanned, and provide the basis for computer intelligence.
I think each of these provides a different set of halachic challenges.
Starting with the last, say we had a detailed scan of a widely recognized posek
and Gadol haDor. Assuming it had not only all of the Gadol's memories, but had
the entire writen corpus of responsa, and original sources online, one would
expect it would be at least as competant and reliable as the Gadol himself was.
Putting aside the formal question of recogizing it as "Mara DeAtra" (local
decisor) it would, at the very least, be able to provide a firm basis and well
written draft opinion for the fully human Mara DeAtra to endorse. After enough
time, one could imagine people might rely on it directly. No doubt, however,
over time we'd have multiple competing such "Ai Rabbis" to continue our
traditional debates. :>
In any event, just what should be the halachic parameters if/when a "real"
logical clone of a posek is available?
Jumping back to the LLM, I submit that a "stochastic parrot" shouldn't even be
trusted to write a good draft opinion. While people use the term "hallucinate"
to describe the way these LLM "parrots" make up facts, the sad truth is that it
has no idea about truth or falsehood, it just computes what the next most likely
word is. Various "guardrails" (the logic behind such things is far from
complete, and more complex than the LLM logic itself). But with some time, much
polish, I suppose it is conceivable that a posek might be able to leverage such
a device to assist in research. Especially if it is well integrated with a
search engine and trained exclusively on appriropriate rabbinic responsa.
The AI optimists present the greatest challenge. If a combination of "simple"
engines eventually gain real general intelligence, how will we recognize it vs.
a very good stochastic parrot? At what point ought we trust such a thing? (that
is, assuming it doesn't run amok and make humans moot. ;<) And assuming we
recognize it as intelligent, and it produces fully annotated responsa (so we can
validate the quotes/basis for reasoning), at what point might it become
acceptable for decision making?
----------------------------------------------------------------------
From: Micha Berger <micha@...>
Date: Mon, Apr 3,2023 at 03:17 PM
Subject: Bracha rishona
Alexander Seinfeld wrote (MJ 66#37):
> Why should liftan mean necessarily eaten in the same bite as the bread?
> it's something you eat along with bread...
Because I don't think that's the translation. For example, in Hil. Bishul Akum,
"lalefes bo es hapas" shows up -- the same shoresh but as a verb. (AZ 38a, 59a)
It is something you do to the bread.
A lefet is a turnip, and "lelafeis bo es hapas" is often taken to mean something
used as a relish on the bread. See, for example the Kaf haChaim. Yes, some
English translations do say "to accompany the bread", but I think that's a loose
translation, motivated from the conclusion.
An example which links liftan and lalefes, "kedei lelafeis bo es hapas" for two
meals is a unit of measure for eiruvei chatzeiros. (SA OC 386:6) But only for
liftan. Otherwise, the shiur is enough food for two meals.
My question was about how the decision was made that being a flavoring for the
bread wasn't the important feature, that liftan is just one way the food and the
bread are in the same course. On the contrary, it seems more natural (to me) to
have assumed lelafeis bo es hapas to be a case of eating an ikar and tafeil
together.
> For me, sweet kugel is 100% dessert but my family disagree.
Or all those apple "kugels" that are literally apple crumbles...
Tir'u baTov!
Micha Berger
http://www.aishdas.org/asp
Author: Widen Your Tent
- https://amzn.to/2JRxnDF
----------------------------------------------------------------------
From: Steven Oppenheimer <steven.oppenheimer@...>
Date: Wed, May 17,2023 at 04:17 PM
Subject: Bracha rishona
Chaim Casper (MJ 66#37) brings proofs that a beracha rishona should not be made
for desert. To add to this, Rav Moshe Feinstein is quoted by Rav Aharon Felder
that one does not make a separate beracha for the deserts at a Viennese Table
(Rishumei Aharon chelek 2, page 20).
Steven Oppenheimer, D.M.D.
----------------------------------------------------------------------
From: Micha Berger <micha@...>
Date: Mon, Apr 17,2023 at 07:17 PM
Subject: First to minyan?
Joel Rich wrote (MJ 66#37):
> If one has a choice of being one of the first 10 to the minyan or putting
> on taalit and tfillin in the entrance hallway of the building that has the
> shul, which gets priority?
First I was thinking: it depends which area the sho'el needs more work in. After
all, both are "nice to"s, not hard-n-fast halachic requirements.
Then I thought... hmm... putting on one's tallis before entering shul is from
the Ari. Based on but not in the Zohar. And while the Rama calls it a common
minhag, the Mechaber, Mehnarshal, Gra, etc ... don't consider donning a tallis
before shul all that important. How can it compare to a desideratum praised in
the gemara?
So, let me limit that first impression to tefillin (not tallis) vs being one of
the first ten.
Tir'u baTov!
Micha Berger
http://www.aishdas.org/asp
Author: Widen Your Tent
- https://amzn.to/2JRxnDF
----------------------------------------------------------------------
From: Carl Singer <carl.singer@...>
Date: Wed, May 17,2023 at 05:17 PM
Subject: First to minyan?
Joel Rich asks (MJ 66#37):
> If one has a choice of being one of the first 10 to the minyan or putting on
> talit and tfillin in the entrance hallway of the building that has the shul,
> which gets priority?
The relevant issues include:
1. Punctuality - Will your choice impact when the minyan begins / continues,
etc.? Some weekday Shacharis minyanim that I know of are structured to end at
a specific time to allow baalabatim to catch public transportation and get
to work on time. (The minyan I currently attend starts at 6:15 T-W-F and
6:10 M-Th -- the focus is on the end time. I've been to minyanim where 8 or 9
people are looking towards the door -- perhaps delaying davening until someone
notes that a "tzenter" (10th individual) has arrived.
2. Anxiety -- If it is a schvach (weak) minyan, your being in the hallway as
opposed to being with the kehillah may cause anxiety for someone who is worried
about having a minyan. I've been to a "weak" minyan where people are looking
towards the door -- perhaps delaying davening until someone notes that a
"tzenter" (10th individual) is present.
3. Courtesy - Will your putting on your tallis and tephillin (say in a crowded
room as opposed to the hallway) disturb others.
Also -- are we implying that being one of the 1st 10 for the minyan is somehow
more choshuv?
Carl Singer
----------------------------------------------------------------------
From: Menashe Elyashiv <menely2@...>
Date: Thu, May 18,2023 at 10:17 AM
Subject: First to minyan?
In response to Joel Rich (MJ 66#37):
The Ari"zl held that putting on Tallit & Tefillin at home and then going to pray
is more important than being the first ten. I do not think that doing that in
the entrance is so important. I personally can not do that as I drive to prayer,
and it is not proper to drive wearing Tefillin, so I am the first to come (well,
I am the gabbai)
----------------------------------------------------------------------
From: David Ziants <dziants@...>
Date: Wed, May 17,2023 at 04:17 PM
Subject: Nusach hatfila
Joel Rich (MJ 66#37) asks for the rabbi's/gabbai's response if a Sfardi insists
on saying Sfardi kaddish on an ongoing basis.
In the Ashkenazi shuls in my neighbourhood, this happens all the time.
The Ashkenazi kaddish sayers just stay quite whilst the Sephardim are
saying the extra eleven words:
"chaim, v'sova, v'y'shua, v'nechama, v'shizava, oor'phu'a, oog'ula, oos'licha,
v'chapara, v'revach, v'hatzala"
and then they all try and say the words "lanu" and "alainu" (which more-or-less
mean the same thing) in unison. Sometimes, an ashkenazi kaddish sayer misses
this, so he just quickly catches up.
Like everywhere, there are always congregants who feel that this is a tircha
d'tzibura [delaying the congregation], but we just have to learn to be more
tolerant of each other.
David Ziants
<dziants@...>
----------------------------------------------------------------------
From: Menashe Elyashiv <menely2@...>
Date: Thu, May 18,2023 at 09:17 AM
Subject: Nusach hatfila
In response to Joel Rich (MJ 66#37):
In our daf yomi class, sometimes a guy came and said kaddish at the end. He said
a sephardi one in an ashkenazi place. Usually it was o.k., however, once the Rav
was present, he gave the guy an ashkenaz siddur and told him that this is what
is used here. The next times, the Rav was not present, and he continued his way.
I personally will say kaddish if no one else will say it, no reason to miss an
important prayer, the daf yomi says that after the destruction of the temple,
the world continues because we say "yehe shmai...", and I will say a sephardi
one even in an ashkenazi place.
----------------------------------------------------------------------
From: Yisrael Medad <yisrael.medad@...>
Date: Thu, May 18,2023 at 04:17 AM
Subject: Shavuot Second Day on Shabbat in Chutz La'Aretz
A Ben Eretz Yisrael is in Chutz La'aretz on the Shabbat immediately after
Shavuot which falls on Friday this year. What Kiddush version does he recite? Is
there a difference between a public and private recitation?
--
Yisrael Medad
Shiloh
Israel
----------------------------------------------------------------------
From: Yisrael Medad <yisrael.medad@...>
Date: Thu, May 18,2023 at 04:17 AM
Subject: Shower shiur
Joel Rich asks (MJ 66#37):
> Listening to a shiur in the shower recommended, permitted or forbidden?
I would like to know if one is attending a bridal shower and the kallah gives
the shi'ur, can one listen?
Yisrael Medad
Shiloh
Israel
----------------------------------------------------------------------
End of Volume 66 Issue 38
