Path: ...!weretis.net!feeder8.news.weretis.net!reader5.news.weretis.net!news.solani.org!.POSTED!not-for-mail
From: Mild Shock <janburse@fastmail.fm>
Newsgroups: sci.math
Subject: Re: Good Bye Stack-Overflow
Date: Fri, 15 Nov 2024 03:37:15 +0100
Message-ID: <vh6c4r$6mlv$7@solani.org>
References: <vh6c2c$6mlv$5@solani.org> <vh6c3i$6mlv$6@solani.org>
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Injection-Date: Fri, 15 Nov 2024 02:37:15 -0000 (UTC)
Injection-Info: solani.org;
	logging-data="219839"; mail-complaints-to="abuse@news.solani.org"
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101
 Firefox/91.0 SeaMonkey/2.53.19
Cancel-Lock: sha1:TPzhd8fGIxBzwRPHYJoQyhEvmT4=
X-User-ID: eJwNirkBACAIxFaSH8c5EfcfQZImRUycvELdXO2NA70ztbFDbK1Gn35FUN6eLHW4UjFTBgrM5S2XJQgr7wdfqhXl
In-Reply-To: <vh6c3i$6mlv$6@solani.org>
Bytes: 3416
Lines: 82

My objection was then that Prolog didn't do
the proof only ChatGPT did it.

"But Prolog is supposed to search for proofs.
What is the proof?"

It then continued with attempts like:

what_if_befriend(Person, Friend, Goal) :-
     assert(friend(Person, Friend)),  % Assert the new friendship
     Goal,  % Check the goal with the new friendship
     retract(friend(Person, Friend)).  % Retract after checking the goal

Only the above will not work, since
assert/1 cannot return a variable binding.

So what is the magic behind lambda-Prolog
and hypothetical reasoning? lambda-Prolog
not being in the scope of ChatGPT?

Mild Shock schrieb:
> But ChatGPT has still to learn a lot, it
> cannot make the link from Prolog to
> lambda-Prolog to natural deduction.
> 
> Somehow after some questioning ChatGPT
> presented me this Prolog code:
> 
> likes(john, pizza).
> likes(mary, sushi).
> likes(sam, burgers).
> 
> friend(john, mary).
> friend(mary, sam).
> 
> likes_what_friends_like(Person, Thing) :-
>      friend(Person, Friend),
>      likes(Friend, Thing).
> 
> I then asked for a Prolog query that
> solves the following question:
> 
> "who should john befriend so that we
> can assume he likes burgers"
> 
> Interestingly ChatGPT strated using
> Prolog and natural deduction on the meta
> level, mentioning modus ponense, and answering
> 
> Proof via Natural Deduction:
> Assumption: Assume friend(john, X) (where X = sam).
> 
> Apply the Rule: Since friend(john, sam) is true, and likes(sam, burgers) 
> is also true, we apply the rule likes_what_friends_like(john, Thing) and 
> infer likes_what_friends_like(john, burgers).
> 
> Conclusion: Therefore, likes_what_friends_like(john, burgers) holds, 
> which implies that John would like burgers if he befriends Sam.
> 
> Mild Shock schrieb:
>> Hi,
>>
>> The cest pool of bit rot is finally gone:
>>
>> (Bit rot, because they did everything in a
>> facist way ignoring your copyright to prevent
>> you from deleting you own out dated stuff)
>>
>> CoincidenceIDontThinkSo
>> https://9gag.com/gag/abAG4oE
>>
>> The Fall of Stack Overflow
>> https://observablehq.com/@ayhanfuat/the-fall-of-stack-overflow
>>
>> The downfall is related to the appearance of
>> ChatGPT. I feel pitty for Ulrich Neumerkel false
>> and others who spammed stackoverflow for years
>>
>> with their infinite dif/2 nonsense.
>>
>> Bye
>