Does TPF now disparage as vanity or prejudice or laziness the preference of some posters to decline to use AI at all? — bongo fury

No, the idea is to encourage specific good practices in the use of LLMs, assuming they're going to be used. We are not encouraging people to use it if they're not already.

Which parts of "research, brainstorming, and editing" does that apply to? — bongo fury

The meaning of "substantial" use is detailed in "2. The Cardinal Rule: Transparency and Disclosure":

Substantial Use: If an LLM has contributed significantly to the substance of a post—for example, generating a core argument, providing a structured outline, or composing a lengthy explanation—you must disclose this. A simple note at the end like "I used ChatGPT to help brainstorm the structure of this argument" or "Claude assisted in refining my explanation of Kant's categorical imperative" is sufficient. — Deepseek

Anyway, as I said, those guidelines are merely suggested; I am looking for constructive criticism.

(I don't want to dominate this discussion too much. I'll probably end up starting a new discussion thread specifically for building and refining the AI guidelines.)

Standard preferences to properly orientate the LLM at the beginning of every conversation are good. Like this:

My LLM Philosophy Discussion Preferences

Your Role: Act as a Socratic sparring partner to augment my thinking, not a ghostwriter.

Direct Instructions:
- Challenge my arguments and suggest counter-positions.
- Help brainstorm and structure ideas, but do not compose full arguments for me.
- Clarify concepts neutrally; I will verify all information.
- Improve the clarity of my existing writing.

Critical Rule: All output is for brainstorming and must be usable with full transparency on a public forum. Do not do my thinking for me.

Response Style: Be logical, direct, and transparent about limitations.

So you keep this somewhere easy to find and paste it in at the top of every philosophy-related conversation you have with an LLM.

No, it isn't. Wittgenstein said nothing of the sort. — Banno

So you say, but will you demonstrate that you actually believe what you say? Otherwise you are just showing that you know how to arrange words in an intelligible way.

I elicited your response, thus doing more than arranging words. Which was to be proved.

I elicited your response, thus doing more than arranging words. — Banno

You appear to be incorrectly applying some sort of theory of cause and effect. You did no such thing. I willfully criticized your act of arranging words. I apologize for having to disillusion you, concerning your attitude of having power over me, but you are referring to something I did, as if it was something you did. Please, assign responsibility where it is due.

We are not encouraging people to use it if they're not already. — Jamal

Good, thanks. Even better if you could retain:

AI LLMs may be used to proofread pre-written posts, but if this results in you being suspected of using them to write posts, that is a risk you run. We recommend that you do not use them at all.

Failing that, a clear statement of,

We are not encouraging people to use it if they're not already. — Jamal

?

Also helpful, perhaps, some routine and visible indication, one way or the other, of,

No AIs were consulted in the making of this post. — Pantagruel

? Or if that's too harsh, something like, that none (or some) were used for

generating a core [or subsidiary] argument, providing a structured [or unstructured] outline, or composing a lengthy [or short] explanation — Deepseek

[Depending on where lines got drawn?]

As the question is now (rather suddenly) tending to arise, when one starts to browse any post?

Thanks for pointing that out.

And saying nothing else.

Am I the only one saying things that could fit in the other thread?

Fair point. There I struggled with thinking on how to do it, which no answer leads to your line of questioning.

Not at all :) -- I suspect that here we're likely not very alone on this after all.

how we can use AI to do better philosophy — Banno

Doesn’t that just depend on the LLM? And who determines that? We need to be better philosophers first in order to judge whether the LLM output is “better” and so whether the LLM is useful.

The question since 3000 years ago is “How can we use X to do better philosophy?” AI is just a new tool, a new “X”. Nietzsche asked “how can I use prose to do better philosophy?” Russell and Witt asked about math and linguistics.

Unless this thread is a tutorial on using LLMs that “better philosopher” way.

Amateur philosophers just spend their lives struggling to understand the world, ping off a few cool philosophers, and spout what they may. — frank

How is that any different from any philosopher?

The difference (to you) is your own judgement of what is “spouted”. And maybe the number who make up the “few”.

I appreciate the explanation of sandbagging. The adaptive process seems parallel to marketing feedback to customer selections: An algorithm of mirrors inducts future novelty.

That adds another wrinkle to "when is the interlocutor not an interlocutor" question discussed previously.

I really do not see the difference here. Following Wittgenstein, all that "saying something" is, is arranging words as if you were saying something. Meaning (as in what is meant, by intention) is not a separate requirement for "saying something", because meaning is assumed to be inherent within "arranging words as if you were saying something". — Metaphysician Undercover

That reading of Wittgenstein assumes "meaning" is an arbitrary convention. That is precisely what he militates against in Philosophical Investigations. There is a passage that is amusing to read in this conversation about interlocuters being actual or not.

A human being can encourage himself, give himself orders, obey, blame and punish himself; he can ask himself a question and answer it. We could even imagine human beings who spoke only in monologue; who accompanied their activities by talking to themselves.—An explorer who watched them and listened to their talk might succeed in translating their language into ours. (This would enable him to predict these people's actions correctly, for he also hears them making resolutions and decisions.)

But could we also imagine a language in which a person could write down or give vocal expression to his inner experiences—his feelings, moods, and the rest—for his private use?——Well, can't we do so in our ordinary language?—But that is not what I mean. The individual words of this language are to refer to what can only be known to the person speaking; to his immediate private sensations. So another person cannot understand the language. — PI, 243

As it relates to this OP, Wittgenstein's statement throws the issue of pretense of AI into sharp relief. It is a pretend form of monologue when talking to oneself and a pretend form of dialogue when directed at others, whether admitted or not.

As a camper on the colline de Molière, my observation more properly belongs on the other OP.

There are those, Hinton being one of them, who claim that the lesson to be learned from the LLMs is that we are also just "arranging words as if it were saying something", that is that we don't have subjective experience any more than they do. I remain skeptical, but I entertain the possibility that there might be something in that. — Janus

Yes ... reminds me of school, and later as well. Many teachers and the whole approach to education considered us to be basically things, automata, robots. And then later, in the early days of internet forums, there was this in-your-face atmosphere of, "You can't think for yourself, you're just parroting others, you can't produce anything original". That line, "Please, Blue Fairy, make me a real life boy (girl)" was oddly relatable. Come to think of it, it still is.

So when I now read criticism of AI/LLMs, I'm reminded that those were the exact things we were told.

Rejecting an argument because it is AI generated — Banno

Not quite. I reject the headlong conflation of a text, identifiable word for word, with the host of arguments, meanings, ideas, intentions, speech acts etc that we seek to discern and delineate by comparing texts. All of which are potentially noble and valuable things, but not reliably identifiable nor distinguishable from instance to instance.

Natural language games (arts and sciences) are typically indeterminate that way. As pointed out by many. Goodman, in particular, points out (in the usual place) that the discerning and discriminating, though frequently too fine-grained to achieve replicability, is usually valuable (cognitively) partly on account of the tradition of anchoring it in comparison of physically identifiable artworks or literary works. (Or photos or musical recordings.)

Hence the potential importance of authenticity of a painting, and of identity of a text. And one way we often are able to benefit from the identification of texts and paintings, in our striving to discern meanings and arguments and intentions (e.g. hidden but implied premises or allusions) is by reference to historical facts about the authorship. A fallible way, sure. Fallacious, often. But certainly disrupted, and not necessarily for the better, by plagiarism of one kind or another.

I did it again. The Phatic act of stringing a few words in order performed the illocution of making a post and the subsequent perlocution of your response.

I'll leave you to it. This should probably be elsewhere.

Doesn’t that just depend on the LLM? — Fire Ologist

It also depends on the prompt. Prompt engineering is a "thing", as the kids say.

The difference between the response of GPT to my OP, prefixed or not prefixed by "Look at this rubbish, by Fred".

Perhaps an instance of Hegel noting where a change of quantity is a change of quality.

Well, it's output seems generally well written, though not scintillating. And, what's written speaks for itself. I think it should be identified when used but otherwise am unconcerned. I long to see its comment on such gems as "Nothing nothings."

Good stuff.

But first, it seems inevitable that in a thread on the effective use of AI to do philosophy, there will be some need to defend the use of AI to do philosophy. That by way of my excusing my going somewhat off topic.

In the OP I mentioned a few objections, but not aesthetics. From what you and others have said, it's clear that the strongest objection is aesthetic.

And if your objection is that we should not use AI because it is ugly, then that's perhaps an end to the discussion.

It is ugly because it is inauthentic. Not in the existential sense, but by way of not having a flesh-and-blood person behind the phatic act.

An AI cannot put its balls on the anvil.

I think this a very good objection.

The answer we have is

The final responsibility for the content and quality of your posts always rests with you. — Deepseek

In posting, one commits oneself to the content of the post. That's the move from the mere phatic tot he illocution. And it occurs, AI or no.

In the OP I mentioned a few objections, but not aesthetics. From what you and others have said, it's clear that the strongest objection is aesthetic. — Banno

I would like to challenge that but am presently more interested in the Bongo response.

An AI cannot put its balls on the anvil.

I think this a very good objection. — Banno

Agreed! That's indeed the chief ground for not treating it like a person. People often argue that chatbots should not be treated like persons because they aren't "really" intelligent. But being intelligent, or wise, in the case of persons (i.e. socialized, enculturated rational animals), always has two tightly integrated components: one doxastic and one conative. One must know the layout of the space of reasons and one must be motivated to pursue the right paths while navigating this space in the pursuit of theoretical and/or practical endeavors. Chatbots lack conative autonomy and hence purse whichever paths they think their users want to explore (or, worse, that merely lead to the outcomes they think their users want to achieve, while having the mere appearance of soundness.) So, they lack part of what it needs to be wise, but that's not because they aren't smart or knowledgeable enough to be useful conversation partners. The human partner remains responsible for deciding where to put their balls.

"Nothing nothings." — Ciceronianus

Let's take that on as an example. Just as is, the reply is:

That’s a nice compact paradox — evocative of Heidegger’s “Das Nichts nichtet,” isn’t it?

If “nothing nothings,” then nothing is doing something, which already seems to contradict its status as nothing. Yet Heidegger’s point was that the nothing is not a mere absence, but an event or activity: it “reveals” beings as such by withdrawing them into non-being.

Do you mean it in that Heideggerian sense — that “nothing” acts or occurs — or more as a logical or linguistic puzzle (like “the concept of nothing cannot have a referent”)?

But add "That fool Fred said..." and we get

Heh — then Fred has stumbled (perhaps unwittingly) into Heidegger’s paradox.

“Nothing nothings” (Das Nichts nichtet) comes from Heidegger’s 1929 lecture What Is Metaphysics? — and it’s deliberately outrageous, because it violates grammar to reveal something about how language and thought presuppose Being. Heidegger’s move is to treat “Nothing” not as a mere negation of things, but as something that actively “nihilates” — that is, makes possible the disclosure of beings as beings by withdrawing them.

So when Fred says “Nothing nothings,” if he means it mockingly, he’s missing the point. But if he’s accidentally gesturing at the idea that the nothing is not nothing, that it somehow “is” in its own way, then he’s (accidentally) paraphrasing Heidegger’s entire essay.

Would you like me to unpack Heidegger’s logic step by step — why he thinks “Nothing” must “nothing”?

Following the protocol suggested in the Game changers article, we begin:

• Identify the problem / question / task
• Incorporate prior knowledge / clarify assumptions
• Design the prompt (structure)

We might leave the problem fairly general - what are we to make of "Nothing nothings"? The prior knowledge, we might lift from what was said above. A prompt such as :

"Nothing nothings" is a translation of “Das Nichts nichtet”, from Heidegger’s 1929 lecture What Is Metaphysics? What are we to make of this? Summarise three different responses.

Here's the result.

What do you think, @Ciceronianus?