10 ms·

Yes, I cancelled claude subscription a few weeks ago because sonnet 5 "ended a chat" over my calling something retarded. Unbelievably irritating for some pile

Yes, I cancelled claude subscription a few weeks ago because sonnet 5 "ended a chat" over my calling something retarded. Unbelievably irritating for some pile of bits to get uppity with me; will never pay for such.

throw1234567891 2mo ago
[flagged]
- embedding-shape 2mo ago
  You really managed to zoom in on the right issue here. Seems really weird to steer/configure/train a LLM/platform to literally close the session if you happen to use bad word too much, regardless if it's accurate or not. They don't get offended, they shouldn't pretend as such, and I should be able to tell it go fuck itself without it playing victim and closing the conversation.
  - hhh 2mo ago
    Why would you disrespect your own tools?
    - embedding-shape 2mo ago
      It's tools, who cares? I should be able to use a hammer, mistakenly hit my thumb with it, call it a bunch of names and maybe even throw it to the ground and stomp on it, and then be able to pick it up again, clean it and continue working with it. It's a thing, it shouldn't pretend to be a individual with feelings.
      - throw1234567891 2mo ago
        Yeah, but the hammer doesn’t generate hits based on an algorithm. You hold the hammer, if you hit yourself with it, why do you call it stupid? I bet you don’t call the hammer stupid when someone else hits you with it.
        embedding-shape 2mo ago
        Have you never used tools, been hurt by how you used them, know you are responsible yet take out your pain on the tool, verbally or otherwise? Feels like I'm having a conversation with a robot who hasn't experienced emotions or the very least never used physical tools.
        throw1234567891 2mo ago
        Must have been a long time ago.
        hhh 2mo ago
        no, why would I get mad at an abnormally shaped rock
        embedding-shape 2mo ago
        Because you're human, you're irrational and you do stupid shit.
  - throw1234567891 2mo ago
    Sounds like the new one was trained on senior engineer’s with ego’s, maybe even grey beard’s. Majority of the training data is low quality, so sounds like “yes, you’re right! It works! I commented out unit test’s and now all test’s pass!”, while the new one is a passive-aggressive “if you know better and you talk to me like that, fuck you do it yourself”, or simply turning around and walking away because they don’t need take the abuse anymore.
mdp2021 2mo ago
This is the second level of the implementation of unintelligence. The first was when they most obviously acritically repeated what they heard, "hearsay machines", "stochastic parrots". Intelligence requires assessment over every provisional output - a continuous cycle of criticisms over intuition. The second is proposing doctrinal biases, again without verification of the content - "hysterical reactive machines".
- semiquaver 2mo ago
  > Intelligence requires assessment over every provisional output - a continuous cycle of criticisms over intuition It’s not clear what you’re saying. Most humans don’t think this way, would you say they do not possess “intelligence”?
  - mdp2021 2mo ago
    > Most humans don’t think this way, would you say they do not possess “intelligence”? Substantially yes (but for the improper semantic in that 'possess'), we would say "they don't use it" and importantly "they don't train it". Intuition has a part, but it has to be confronted with reality (factual and logical).
n4r9 2mo ago
Interesting. Was it persistent, or just once? Their post about this functionality says: > This ability is intended for use in rare, extreme cases of persistently harmful or abusive user interactions. https://www.anthropic.com/research/end-subset-conversations https://www.anthropic.com/research/end-subset-conversations Regardless of the correctness of it, I'm curious to know why you thought such language was actually going to be helpful!

throw1234567891 2mo ago

[flagged]

embedding-shape 2mo ago
You really managed to zoom in on the right issue here. Seems really weird to steer/configure/train a LLM/platform to literally close the session if you happen to use bad word too much, regardless if it's accurate or not. They don't get offended, they shouldn't pretend as such, and I should be able to tell it go fuck itself without it playing victim and closing the conversation.
- hhh 2mo ago
  Why would you disrespect your own tools?
  - embedding-shape 2mo ago
    It's tools, who cares? I should be able to use a hammer, mistakenly hit my thumb with it, call it a bunch of names and maybe even throw it to the ground and stomp on it, and then be able to pick it up again, clean it and continue working with it. It's a thing, it shouldn't pretend to be a individual with feelings.
    - throw1234567891 2mo ago
      Yeah, but the hammer doesn’t generate hits based on an algorithm. You hold the hammer, if you hit yourself with it, why do you call it stupid? I bet you don’t call the hammer stupid when someone else hits you with it.
      - embedding-shape 2mo ago
        Have you never used tools, been hurt by how you used them, know you are responsible yet take out your pain on the tool, verbally or otherwise? Feels like I'm having a conversation with a robot who hasn't experienced emotions or the very least never used physical tools.
        throw1234567891 2mo ago
        Must have been a long time ago.
        hhh 2mo ago
        no, why would I get mad at an abnormally shaped rock
        embedding-shape 2mo ago
        Because you're human, you're irrational and you do stupid shit.
- throw1234567891 2mo ago
  Sounds like the new one was trained on senior engineer’s with ego’s, maybe even grey beard’s. Majority of the training data is low quality, so sounds like “yes, you’re right! It works! I commented out unit test’s and now all test’s pass!”, while the new one is a passive-aggressive “if you know better and you talk to me like that, fuck you do it yourself”, or simply turning around and walking away because they don’t need take the abuse anymore.

mdp2021 2mo ago

This is the second level of the implementation of unintelligence. The first was when they most obviously acritically repeated what they heard, "hearsay machines", "stochastic parrots". Intelligence requires assessment over every provisional output - a continuous cycle of criticisms over intuition. The second is proposing doctrinal biases, again without verification of the content - "hysterical reactive machines".

semiquaver 2mo ago
> Intelligence requires assessment over every provisional output - a continuous cycle of criticisms over intuition It’s not clear what you’re saying. Most humans don’t think this way, would you say they do not possess “intelligence”?
- mdp2021 2mo ago
  > Most humans don’t think this way, would you say they do not possess “intelligence”? Substantially yes (but for the improper semantic in that 'possess'), we would say "they don't use it" and importantly "they don't train it". Intuition has a part, but it has to be confronted with reality (factual and logical).

n4r9 2mo ago

Interesting. Was it persistent, or just once? Their post about this functionality says: > This ability is intended for use in rare, extreme cases of persistently harmful or abusive user interactions. https://www.anthropic.com/research/end-subset-conversations https://www.anthropic.com/research/end-subset-conversations Regardless of the correctness of it, I'm curious to know why you thought such language was actually going to be helpful!