99 Comments
User's avatar
Ash's avatar

Of course AI can't pasken. It might be able to find or summarize sources, but peak requires shikul hadass.

Expand full comment
Yehoshua's avatar

The question of whether AI will ever be able to pasken is just another way of phrasing the question whether AI will have human style consciousness.

I think a fundamental, perhaps the most fundamental, of Yiddishkeit is that consciousness in a fundamental property of its own, not simply emergent from high intelligence.

Expand full comment
Happy's avatar

I think the question about whether AI will ever be able to pasken is another way of phrasing the question whether AI will be able to perfectly imitate human style consciousness, and not just consciousness, but imitate being a human itself. I contend that will never happen. If it does happen, we can discuss whether we can trust something that perfectly imitates being a human, and I assume not, as it is still only an illusion- I think that's the point you are making.

Expand full comment
Yehoshua's avatar

If consciousness is just an emergent property of a certain amount of connected neurons, then AI can be conscious. To imitate consciousness is an entirely different project. Whether it is possible to test if the consciousness is real or imitated is an interesting question. I don't know the answer to that.

Expand full comment
Shaul Shapira's avatar

"Do you need any more evidence than this? Doesn’t it sound exactly like one of our friend’s essays on kollel or army service or how Rashi was a corporealist? Like an academic trying to write Torah as if it was a college paper, with all the academic jargon? "

No, it doesn't. ChatGPT is a different kind of nonsense. I disagree with a bunch of Dr Slifkin's arguments, but the point is *there's something to disagree with.* AI, at least thus far, is just a disconnected jumble of words. It's not only a problem with Halacha. L'havdil, secular courts have been dealing with the issue as well. https://www.reuters.com/legal/new-york-lawyers-sanctioned-using-fake-chatgpt-cases-legal-brief-2023-06-22/ See also here: https://www.scotusblog.com/2023/01/no-ruth-bader-ginsburg-did-not-dissent-in-obergefell-and-other-things-chatgpt-gets-wrong-about-the-supreme-court/

(Human) Judge Kevin Newsom has a delightful opinion here https://media.ca11.uscourts.gov/opinions/pub/files/202212581.pdf on the pro and cons of using AI in judging. (Starts on page 25 of the PDF.) I think some of his points are transferable to Halacha, but I'm not a posek, automated or otherwise.... :

Expand full comment
shulman's avatar

With the AI example, well done!

Expand full comment
Ash's avatar

Who did you vote for in the end?

Expand full comment
Happy's avatar

What goes on in the voting booth stays in the voting booth ;)

Expand full comment
Shimshon's avatar

https://voxday.net/2024/07/01/ai-fraud-and-fakery/

Vox Day excerpts from and links to this (https://ghostarchive.org/archive/BeICd) post, which is full of foul language, but valuable insights regarding AI. A key point:

"Consider the fact that most companies are unable to successfully develop and deploy the simplest of CRUD applications on time and under budget."

AI actually has utility, but not as implemented by the big companies that encrust it with restrictive wrappers in pursuit of an ideological agenda.

Expand full comment
Shaul Shapira's avatar

"AI actually has utility, but not as implemented by the big companies that encrust it with restrictive wrappers in pursuit of an ideological agenda."

Retarded thing to say.

Expand full comment
Shimshon's avatar

Google's AI is just a particularly notorious and obvious example of what I refer to.

Expand full comment
Shaul Shapira's avatar

Encrusted restrictedness?

Expand full comment
Shimshon's avatar

Try asking any AI questions involving race.

Expand full comment
Shimshon's avatar

What's your point? That you fall victim to the Genetic Fallacy repeatedly?

Expand full comment
Shaul Shapira's avatar

Nah. Vox Day is an all around moron. There's no need to use any specific area of his idiocy to prove that. He says dumb things about just about everything.

Expand full comment
Shimshon's avatar

I responded relevantly to the subject. You didn't. Committing the Genetic Fallacy is retarded. I mentioned that Vox referred to another post with foul language. Any opinions on that? Or is it tainted by association.

The government of Israel seeks to be like all the nations. One of the Prime Minister's signature books is titled "A Place Among the Nations". On that self-absorbed metric, Vox is not wrong. We make bad goyim. We should stop trying to be like them.

This war is a terrible tragedy. Our boys would be better off being jailed for defying orders designed to kill them. Moshe Feiglin begged his grandson to do exactly that days before he was killed going into a booby-trapped building.

Expand full comment
Yair's avatar

This is the first time I'm commenting here, so hi everyone.

Happy, you wrote:

"...that’s exactly what ChatGPT knows how to do. It analyzes trillions of texts and pumps out a synthesis of said texts. But it doesn’t know how to learn. Knowing how to learn requires a Mesorah. It requires reasoning. It requires Yiras Shamayim. It requires Shimush Talmidei Chachamim. It requires the “fifth cheilek of the Shulchan Aruch”. Without these, you will get very bad results."

You then provided an exchange demonstrating ChatGPT's apparent ignorance of several well-known Torah concepts, labeling this as a sample of its "learning capabilities".

(While the article ended with finding issue with the language of ChatGPT's response, the main point was obviously "the nonsensical content rather that the high school essay tone", as you wrote in the comments. The language issue is something quite easy to address and improve, cf. PosekGPT...)

The example that you provided, however, doesn't illustrate your opening points, that ChatGPT "doesn’t know how to learn. Knowing how to learn requires a Mesorah. It requires reasoning. It requires Yiras Shamayim. It requires Shimush Talmidei Chachamim. It requires the “fifth cheilek of the Shulchan Aruch”."

Rather, it primarily demonstrates the current limitations of ChatGPT's knowledge base in Torah and Halachic topics - a constraint that could potentially be addressed through more comprehensive training on libraries like the Bar Ilan Responsa Project or Otzar Hachochma.

Regarding your specific points:

1. Reasoning: AI models, including ChatGPT, have demonstrated significant capabilities in logical reasoning and are rapidly improving. It is premature to conclude they cannot reason effectively in halachic contexts.

2. Mesorah, Shimush Talmidei Chachamim, and the "fifth cheilek of the Shulchan Aruch": If these can be taught to humans, they may be encoded into AI systems through careful training and programming.

3. Yiras Shamayim: One can argue that an AI system could be designed to consistently adhere to halachic principles without being swayed by personal biases or temptations.

Your article raises valid concerns about the current state of AI in halachic reasoning. However, in a discussion oriented toward the future, it's important to account for potential advancements in AI technology. The limitations you've identified in today's AI models may not necessarily apply to future iterations.

Generative AI chatbots like ChatGPT are still in their infancy. It would be prudent to wait patiently and observe how this technology develops before drawing definitive conclusions about its potential impact on the field of psak.

Expand full comment
Hammer Otongo's avatar

" I will not apologize for my positions"

That's the problem: you probably should.

In our darkest hour since the Holocaust, Biden was there for us. That's ultimately what is going to be Joe Biden's legacy, not all the other truly horrible stuff he has done throughout his career, Presidency and even his personal life. But I'm not going to pretend that he was ever and otherwise good guy, because he wasn't. And even from a purely Jewish angle, it is Biden and his party who advocate importing 10s of millions of antisemitic Muslims.

Expand full comment
test's avatar

Sounds very much like a chaburoh put together by 90% of benei torah. 90% of them do exactly the same thing. A memory game with impressive soundbites.

How many times do you ask a learner about X and he will reply, ah yes, there is a machlokas rashbo and rosh on that. And he will genuinely believe he has answered the question.

The next stage up in learning is to repeat over what the rashbo/rosh says. Or make believe he is. Sounds impressive but if you go and look it up you will probably either find there is no such rashbo, or if there is, there is only a vague similarity to what is claimed they said.

See SK with his 'does not compute' comment. Exactly on point.

Look at the seforim put out in the last 20 years by younger 'benei torah'. 90% are likut seforim. Very little original thought. In today's chareidi world original thought is dangerous.

I asked a good ben torah about the reasoning behind a certain halochoh. He said the mishneh beruroh paskens X. I said I know that, but what is the havonoh behind it. He gave me a look and walked off. That sums it up.

I was in that world. I know exactly what goes on.

Expand full comment
מכרכר בכל עוז's avatar

I think this is the first time I ever agreed with you. At least with the second part of your comment. In general, there has been a tremendous aniyus hada'as amongst the lomdei halacha lately. This is very relevant to a topic I am currently researching and writing about (in the real world, under my real name).

Expand full comment
test's avatar

The reason is complex, but it is a similar to the reason why 1,000 years ago the Arab/Muslim was the center of wisdom, progress and scicence and Europe was an ignorant medivel dark aged dangerous place. And post the renaissance that flipped.

Progress and wisdom flourish in a culture that tolerates new ideas. The golden age of arab world coincided with a culture of tolerance and open-mindensess. But intolerant, closed minded societies do not grow.

A society that retreates into itself, fears kefira from every outside idea, does not believe that its basic identity and beliefs provide enough protection, and cannot admit that mistakes were and are being made, will not prosper. Bright inquisitive people do not flourish in the Arab world because the culture persecutes anyone whose ideas do not confirm with the prevailing beliefs ot are thouight to insult religious teachings. Sounds familiar?

Yes in numbers chareidi society is growing, but that is a function of biology, nothing more. As you noted, the quality is going downhill.

For example, a society which rolls out (or repeats, with little variation) arguments from gedolim 100 years ago against the Isreal, without conceding that the gedolim of 100 years ago were wrong when they predicted that EY will become a spiritual wildenress and their arguments are 'not shayech' to 5784 cannot grow and develop. But the society is not prepared to rethink those hashkofos. A society that treats shelled eggs and peeled onions left overnight with more stringency that it ever treated COVID (goyshe science), cannot grow and develop. A society where rabbonim fear to meakel for fear of being no longer considered a godol cannot grow and develop. A society that treates its leaders like popes, refusing to accept they can ever be mistaken, cannot be creative. A society that censors its media, a society that bans texting because it could chas v'sholom be used to effectively for communiated ideas, will stifle itself and cannot be crative.

A society that feels so threatened that many 'gedolim' issue a declaration that one should daven for a relief from the gezeiros of the 'unnatural tradegy' in Lakewood, the death of gedolim, gezeiros against funding but refuse to mention the hundreds of soldiers killed, will not be creative. Ditto a society with 'gedolim' who advise against visiting wounded soldiers. You think the gedolim of yesteryear, The Chofetz Chaim, Reb Elochonon, Reb Shlomo Zalman Aurbach etc would ever dream if telling their talmidim not to visit wounded Jewish soldiers, however secular or Zionistic? That is the difference between then and now. A society that hushes any criticism of anybody, however well known the issue is with 'loshon horoh', cannot grow.

In without growth and tolerance, creativity is stifiled, and the havonoh and development of torah is stunted. Hence all we have is likut seforim.

Plenty more to say, but to be honest, its too depressing.....

PS that's before we note that a society that can only sustain itself with fraud, to various extents, will not get siyatoh dishmayo.....I'm concentrating here on the more cerebral reasons.

Expand full comment
Happy's avatar

Sure, I believe that about your chaburohs. About your chaburohs it's probably closer to 100%.

Expand full comment
Yehoshua's avatar

So write a post about how 'likut seforim can't pasken'. That would be useful.

Expand full comment
Isha Yiras Hashem's avatar

Like an academic trying to write Torah as if it was a college paper, with all the academic jargon? ***

If you're arguing form vs content, you should point out what is wrong with the AI generated content, besides the form. It would make this post stronger.

https://ishayirashashem.substack.com/p/artificial-intelligence-for-dummies

At the end of the day Hashem gave the Torah to humans, we don't get to give it to machines. BH electricity is forbidden on shabbos and Yom tov!

Expand full comment
test's avatar

It's well knoen Chat GPT produces nonsense about all things. No serious talmid chochom or even non serious talmid chochom has ever suggested it will replace poskim.

This is, like most of these posts, a straw man argument.

Expand full comment
Bandit's avatar

Once again, you've completely missed the point. The point of this post seems to have been to compare "rationalists" making very sophisticated-sounding arguments, while the content is in fact gibberish, to ChatGPT doing the same.

Expand full comment
test's avatar

Actually the chaburohs that come out of 'Slifkin's camp' benei torah are generally far better quality than what yeshiva land produces. The former have intros, beginnings, middle, end, summaries, signposting etc. Much better structured than the latter which is often a hodge podge of maroh mekomos all jumbled up with supposed sevoros and not delivered in one single language. The communication skills of a typical 'ben torah' are shockingly poor. (And those that venture onto blogs are by definition not typical.)

Some in the former are actually trained in writing chidushim. Goyish, I know. Not like the yeshivish system where bochur A 'learns by' (whatever that means exactly - like all these soundbites, no one can explain it, and its kefiroh to ask) godol B 'for many years' (together with hundreds of others) and mystically, by some form of osmosis, is supposed to pick up how to produce decent structures divrei torah in the godol's 'mehalech'.

Expand full comment
rkz's avatar

I don't know where you take your examples of chaburus from, but I have read (and graded and edited) over a thosand chaburus (mostly written by charedim) and you are wrong.

Expand full comment
test's avatar

1,000 is what percentage of worldwide 'benei torah' worldwide delivering chaburos? And what do you mean by 'mostly'. I never said no chareidi knows how to write a decent chaburoh. And obviously, chaburos submitted for some sort of test system are going to be excellent quality. It's not a representative sample, by definition.

Another one of the issues with chareidi yeshivaland is the complete dearth of independent quality assessment. Without that, standards fall and fall and fall....even somebody with the much-chased reputation 'he knows how to learn' - it's generally the blind assessing the blind....or the biased assessing the biased.....I know a few 'roshei kollelim' which such poor communication skills nobody can really properly assess how well they can learn - the skill is talking the talk and sounding knowledgeable

Expand full comment
rkz's avatar

Please re-read what I wrote, and then see if you responded to the point the I made (we discussed this before, so I don't understand why you don't understand my point).

Expand full comment
mb's avatar

"BH electricity is forbidden on shabbos and Yom tov!"

Careful. You shouldn't be making sweeping statements like that. Most poskim, if not all, reject the Chazon Ish on that. And I turn lights on Yom Tov.

Cheers.

Expand full comment
Isha Yiras Hashem's avatar

So do I, if someone's life is in danger.

Expand full comment
mb's avatar

Super frum!

Expand full comment
Isha Yiras Hashem's avatar

I like to think of it as "Isha Yiras Hashem"

Expand full comment
test's avatar

They reject the chazon ish, but still say its d'rabbonon. And old fashioned filament lamps involved ma'vir. And molid eish on YT.

Expand full comment
mb's avatar

Virtually everybody agrees filament lights are fire on Shabbat and a Torah violation.. However, on Yom tov, several poskim permit turning them on and some even on and off. We follow Aruch ha Shulkan on this. However the lady said electricity on Shabbat. 6 different opinions on what the problem is. building, something new,completing,sparks,increase of power usage at the source and cooking. R. Shlomo Zalman Auerbach rejects them and says there's no violation, torah or Rabbinic at all. Except for long standing Minhag (not that long of course 150 years, ish.) And permits if there's a need. Obviously saving a life everybody agrees. So what's a real need? Ah there lies the rub.

Hope that is useful.

Expand full comment
test's avatar

Who permits turning lights on and off on YT besides RZPF?

RSZA states its d'rabbon.

אמנם נראה דהאמת יורה דרכו, דאף אם אין בו שלהבת כלל מכל מקום חשיב נולד גמור כיון שהוא נהפך לגמרי מגוף חשוך לגוף מאיר, וחשיב טפי מוליד מהולדת ריח בבגד דאסור

Expand full comment
mb's avatar

The old perfumed garden, and confusing the issue. And being rude.

You are talking about LIGHTS. I was quoting RSZA on non-incandescent use of ELECTRICITY

Read Minhat Shlomo Again. He categorically states that there is no problem at all, except for established Minhag. And the only reason he doesn't permit is because one might turn on/off incandescent light. However, if there's a great need, he permits use of electricity. What's a great need? Beauty is in the eyes of the beholder.

Shabbat Shalom

Expand full comment
mb's avatar

Are you sure R.ZPFrank permits turning off too?

Regardless, CR Uziel permitted both.

Cheers

Expand full comment
rkz's avatar

So did Aruch HaSulhan, but he wrote not to rely on him because he wasn't sure that the explanation given by the שואלים was correct (indeed it wasn't)

Expand full comment
Ash's avatar

You eat treif, no?

Expand full comment
mb's avatar

Who are you asking ?

Expand full comment
Shimshon's avatar

AI won't make rabbis obsolete. It will make certain "rabbi doctors" obsolete. If you are a rationalist, you can be replaced.

Expand full comment
User's avatar
Comment deleted
Jul 3, 2024
Comment deleted
Expand full comment
Happy's avatar

The opposite. If you read my exchange and know anything about anything, you will see the problem is the nonsensical content rather that the high school essay tone.

Expand full comment
Isha Yiras Hashem's avatar

You are overestimating frum people who use the internet... by a lot

Expand full comment
User's avatar
Comment deleted
Jul 3, 2024
Comment deleted
Expand full comment
Happy's avatar

"Compare and contrast" is a perfectly normal form of Talmudic and halachic discourse, one of the most common, if not the most common. And the program optimized for it produces halachic nonsense.

Expand full comment
shulman's avatar

You must of missed the point, so let me clarify. There is no machlokes between the ibn Ezra and the ketzos about this issue. There is no dispute about muktza in this context whatsoever. Happy threw that in for fun and it's shocking that they had an answer. And Rabbi Yehuda Aryeh probably has absolutely nothing to say on any of this that we have in writing. Again, Happy was asking questions that the only answer, for a thinking human, is "idk" or "what shaychus" but gpt hada whole chabura on it! It's freakishly messed up actually.

Expand full comment
rkz's avatar

That's how ChatGPT works.

E.g. The now infamous true story about fasting on Yom Kippur that falls on Rosh Chodesh.

Expand full comment
Happy's avatar

Just now:

Me: If I buying a prepared chicken that has many other ingredients, how can I make sure it is chalav Yisrael?

Chatgpt: If you're buying prepared chicken with many other ingredients and you want to ensure that any dairy products included are Chalav Yisrael, you need to check the kosher certification on the product. Here’s how you can do it:

1. **Certification Symbols**: Look for a kosher certification symbol that indicates Chalav Yisrael. Some common kosher certifying agencies that might indicate Chalav Yisrael include OU (Orthodox Union), Kof-K, OK, and Star-K. They may use specific symbols or wording to denote Chalav Yisrael on the packaging.

2. **Ingredient List**: Check the ingredient list to see if there are any dairy components. If there are, ensure that the certification explicitly states Chalav Yisrael.

3. **Contact the Certifying Agency**: If the packaging is unclear, you can contact the certifying agency listed on the product. They can provide information on whether the dairy ingredients in the product are Chalav Yisrael.

4. **Consult with a Rabbi**: If you are still unsure, consult with a knowledgeable rabbi who can help clarify the certification and ingredients.

By following these steps, you can ensure that any dairy components in your prepared chicken are Chalav Yisrael.

Expand full comment
test's avatar

But why is funding a kollel lifestyle through mateh akum taking of benefits better than chicken with cholov yisroel ingredients?

Expand full comment
shulman's avatar

Lol exactly. I've gotten similar results with very scientific and mathematical equations in which chat gpt questioned is own ability to do basic multiplication!

Expand full comment
Happy's avatar

ChatGPT cannot do any reasoning beyond what it sees in the training data as a language model. Any question that is a little bit beyond it's training data, even on simple math, it will fail. For example, it cannot for its life get the correct answer to the following question I made up "There is a four-sided pyramid that has a base that is 8 meters, and is 5 meters tall. Perched on top of the pyramid is a sphere of radius 3. There is an ant standing on the center of one of the bottom edges of the pyramid. The ant climbs straight up the pyramid, then straight up the sphere until it reaches the equator of the sphere. Then the ant walks west on the equator 80 degrees. How many meters in a straight line is the ant from its original starting position?" It keeps on coming up with 6.77 meters, and I keep on pointing out that the distance cannot be less than 8 meters, and it tries again until it says the distance must be approximately 8 meters, lol.

Expand full comment