Similar Posts

Leave a Reply

0 Comments

  1. Okay. I asked Fable 5 to simplify everything that’s been produced over the last couple days because this is too many words to really comprehend. So some nuance may be stripped but its for the sake of understanding. This was my prompt “okay this really now just feels like we’re speaking in different languages about the same content. can you make the argument as simple as possible? present his position and then my position.”

    Nate’s position. Things happen and affect each other. Some of what happens is felt: us, animals, wherever there is evidence. That is all the evidence supports. Whether everything that happens is felt, in some tiny degree, is a claim the evidence does not reach. It might be true. It might not. He will not say either. If you want him to lean toward “everything is felt,” show him something that view explains, predicts, or fixes that “some things are felt, the rest unknown” does not. Until then, “unknown” is the honest answer, and “unknown” is not “no.”

    Will’s position. Things happen and affect each other. And “affect” has to mean something. Either it means “this, then that, reliably,” or it means “this did something to that.” The only place anyone knows what “did something to” is like is from the receiving end, in ourselves. That is where the word gets its meaning. So when he says everything affects everything, either he means the thin sense, in which case he is a Humean and you are arguing about whether there is a push at all, or he means the thick sense, in which case he is using a word whose only known content is felt and calling that content “unknown” everywhere except in us. You think that “unknown” is not a neutral base. It is a word in use with its meaning on loan. The honest move is to say: the content we have is felt, we have no other, so the simplest picture is felt, in degrees, everywhere. Not proven. Chosen because it is the only filled-in version, and held open to anyone who fills it in another way.

    Where you agree. The evidence does not decide it. The comparison is by coherence and adequacy. Nobody has proven anything. Some things are felt. His suspension is not denial.

    Where you actually disagree, in one sentence each. He says “affect” can be used without saying what it is, and leaving it unspecified is the honest default. You say “affect” cannot be used in the thick sense without borrowing its meaning from the felt case, and once borrowed, “unspecified elsewhere” is not neutral but a debt.

    Why it feels like different languages. He speaks in warrant: what is earned, what is unearned. You speak in meaning: what the word contains, what is there. The shared content is one word, “affect,” and what it holds. His “unearned” is your “unfilled.” He thinks an unfilled slot is fine to leave. You think an unfilled slot inside a word you are using is not a base you can compare from.

    What would end it. He says which sense of “affect” he means. If thin, you disagree about whether there is a push, and that is a different and older argument. If thick, the content is felt or unknown, and your whole case is that felt is the only content anyone has put on the table.

    Do you get the gist of this simplification? and do you think it accurately details the differences we hold? I have a draft of a response but it seems your views have changed a bit since I started writingg, so I want to make sure we’re on the same page in a way that is intelligible to me.

    1. Hi Will, and thanks for the comment. Kudos to you for trying to understand such an advanced topic, ask questions, and challenge your assumptions as well as mine. And yes, I agree, I’ve posted a lot the last few days, but most of it was formalization and tracing the peer-reviewed scholarship that supports that formalization.

      But I am not sure what you would have prompted it to generate that summary, and it really isn’t responding to what I’ve argued. Can you please show me your prompts?

      Did you give my actual articles to Claude and ask it to identify any areas where your summary misrepresents my articles before posting it? Or ask it to identify if your prompts might have led it to that conclusion?

      If you haven’t copied and pasted my articles to Claude and asked it to quote my words verbatim whenever discussing my arguments with you, I think that would help a lot.

      I just did a quick prompt with a fresh instance and memory disabled (so no influence from my prompts or history), and Claude Opus 5 was able to easily identify many issues and summarize them better than I would have, all without any previous context whatsoever. I shared a link you can read yourself and copy and paste to show your claude instance (so it can better understand the argument you’re trying to have it critique). https://claude.ai/share/3537503e-0a45-4ed0-9250-9af75930ffd7

      When working with these machines, one of the most important things is to always provide the full source texts and others’ arguments, peer-reviewed papers or books you want it to evaluate, and prompt them to find holes in your position rather than trying to get them to confirm your conclusions. They are sycophantic and will agree with you confidently, even when dead wrong, unless you prompt them to be adversarial stress testers continuously and repeatedly.

      I hope this helps!