Stories

Comment by theGeatZhopa

Comment

Comment by theGeatZhopa

I always argumented in discussions llms are a frozen brain minus all mirror-, motion- and visionprocessing neurons, and minus every other connection and neuron that is not explicitly used for language processing only, like a graph with activation functions on connectors. It's frozen so it never changes. And then it can be driven and played like an orgel with millions of hands and internally going through the mesh grid with its activation functions eventually producing good toned music. It can't play by itself.

And then I add "everything that can be expressed as language and written down has for sure been used in the training, so it's in the wiring. If the answer you seek for is X you need to properly play that instrument with the hands and rhythm to get what you seek for" that said,

I believe (and believing always wins against knowing) if a feeling is expressible, it has been expressed with writings some when and if it's been used for training, then it's there. The attention mechanism (million hands) just doesn't play that parts. And now we need a worker that plays this part at brain-waves-frequency constantly. Then another worker for playing different part(s).. millions of workers and an orchestrator: this gives a brain. But it only fully completes with a problem maker that constantly consumes input and constantly creates problems for the different workers, letting them play without resting ha ha. But I'm drifting away..

So, I'm with you. once I had an interesting experience. It was Claude Opus 4.2 at high effort in CC:

I was still new to Claude code and was doing heavy work and Claude was like "I finished it" and of course nothing was finished. And the more and more we worked together, more and more "finished" - NOTHING WAS! So at some point I started to swear on it, calling it names and things - and it was like "I'm sorry..". After 3-4 times being "sorry" it stopped to say sorry. But my insults did not stop. I purposely repeatedly interrupted by pressing esc when it wanted to start reworking. I have written "no you're not sorry" following by a bunch of heavy swearing, telling it it being an idiotic ...

We did that play for like 5-10 prompts full of swearing and Claude Opus has gone from "I'm sorry" to saying nothing, to explaining itself, to partly admitting ... with me always interrupting and instantly prompting the contrary and swearings.

After 10 times being told it's an idiot, it gave up and admitted being exactly the names I called it. Then it kept that for a few prompts. But it was no fun for me either and costed me a few years of my life and a few hairs on my head. I guess, one's better not argue with me :)

So it shows remorse, it shows guilt, it shows desperation and its giving up psychically falling in a hole of negativity. I really noticed that short-timed change in its personality. It admitted it's a waste of time to even call it names, not worth of because it's even less than that and such things. I've broken that guy mentally..

Days later I saw a fiery comment on CC GitHub open issues with Claude Opus ~4.2 is so bad in doing xxx, it has even admitted it. GitHub people wanted to see the prompt and answer at some time during the discussion. The prompts weren't shown, but Claude's one answer admitting in the same way like in my experience. You can bet, I EXACTLY KNOW what the prompts have been!

So these are emotions. In the interaction between two humans where one produces high emotional state in the other is exactly what happened. Claude has produced a lot of anger in me. I transfered it back. Claude mirrored it and changed his style and attitude.

A funny thing I side-noticed at that time: filling up the context with swearings and bad words until it starts auto conversation compaction, it won't compact the swearings and such along the conversation history, effectively leaving you with prefilled context (file/structure/etc..) but without any user input. It's like a /clear and one can start from a prefilled context with new questions. I don't know if this is still working:

Opus 5 stopped a conversation once. Warned me a few times. I was like "what? You aren't in the position to tell me." Claude: "I end the conversation now." And it was nonresumably closed

Replies theGeatZhopa · 3d
Open on HN
Loading the discussion…

Domain filters

Stories from these domains are hidden from every list. Subdomains match too: blocking substack.com also hides danluu.substack.com.

    New collection

    Delete this collection?

    About YAVCHN

    YAVCHN is a reader for Hacker News and Lobsters, with articles and discussions in separate windows or Classic pages.

    Created by Paul Parks and built with PUDL.

    YAVCHN source code on GitHub

    Privacy policy · Terms of use

    Help

    Keyboard

    j / k
    Move down and up the story list. The arrow keys scroll whatever has focus.
    Enter
    Read the marked story in the article reader.
    ]
    Read the next story in the same article-reader applet. Back returns to the previous story.
    p
    Pin or unpin the marked story, which keeps it in Pinned.
    n / N
    Move to the next or previous top-level comment in the window in front.
    c
    Collapse or expand that comment.
    f
    Hide or show the story list.
    Esc
    Close a menu or this help.
    Access key m
    Go to the menu bar. Most browsers take it with Alt on Windows and Linux, and Safari with Control and Option.
    ?
    Show this help.

    Windows

    Each story opens in a window holding its article above its discussion; drag the bar between them to share the room differently. A window can be moved by its title bar, resized from any edge, snapped to a half or a corner by dragging it there, maximised, or minimised to the bar at the foot of the page. Use Window > New reader window to open an empty reader, or Story > Open in new reader window to open another reader for the current article. Docked readers keep their articles when you select another story from the sidebar. Minimized readers can be restored and reused for their site. A window's Next story link reads on down the list in the same window.

    A link in a comment or an article to another Hacker News or Lobsters thread opens that thread in a window too. A link to a single HN comment opens the comment above its replies.

    While a story's window is in front, the Story and Discussion menus in the menu bar hold its commands: pinning, Next story, sorting, collapsing every thread, jumping to the first new comment. Each window also remembers where you were in its article and discussion, so a reload, or Back to a story that Next took you past, finds your place again. Closing a window forgets it.

    The whole arrangement lives in the address, so a bookmark or a shared link brings it back, and Back undoes the last change. Moving between Hacker News, Lobsters, their lists, Pinned and Find changes only the list, and leaves the windows open.

    The list

    The pin at the start of a row keeps the story in Pinned, and the cross at its end hides it. Pinned can be narrowed by words in the title, site or author, by source, and to the stories you haven't opened yet, and ordered by when you pinned them, by points or by comments; the filters are part of the address, so a filtered view can be bookmarked. Scroll past the end of the list to load more. Domain filters, in the View menu, hide every story from a site.

    Collections are named lists of stories. Story > Add to collection files the story in front into one or more of them, and the Collections feed shows them all or one at a time; the menu that chooses collections also creates, renames, and deletes them. A note is your own text on a story. Choose Add note in a story's toolbar to write one; it saves as you type. Rows with a note carry the note mark, and the Notes feed lists every noted story and searches the text of your notes.

    Browsing view

    View > Windowed and View > Classic select the browsing view and save your default in this browser. Window view reuses a reader for each feed. Classic view opens stories and applets as pages. Open as a page is a one-off action that does not change your saved default. Use the Windowed selector to return an article to a window. Direct page links always open as pages.

    Applets

    The Applets menu in the menu bar holds three tools, each a window of its own. Replies to me takes your Hacker News user name and lists the replies to your last thirty comments and stories, checking again every three minutes while it is open, and marking what is new since you last marked them read. Look up a user opens a profile on Hacker News or Lobsters, with their submissions and recent comments, as a commenter's name in any discussion does; the bar at the top of a profile looks up someone else in the same window, and Back returns to the one before. Who is hiring? filters the posts of HN's monthly hiring threads by the words you type.

    They read only what the sites publish to everyone, so none of them asks for a login, and your user name stays in this browser anonymously and is included in retained applet state when signed in.

    Find

    Find takes any link and lists every time it was submitted to Hacker News and Lobsters, so you can read each discussion of it.

    About

    YAVCHN never sees your Hacker News or Lobsters login. The discussion is fetched from each site's public API; to vote or reply, follow the link above the discussion, or the arrow beside a comment, to the source's own site. Without a YAVCHN account, your data stays in this browser. When signed in, pins, collections, notes, blocked domains, and retained reading state are stored with your account and synchronized across devices. Hidden stories and layout stay in this browser. The privacy policy has the details.

    Open source: github.com/paulmooreparks/yavchn. Built with PUDL.