#Every single suggestion I have to make neuro (and evil) more entertaining.

1 messages · Page 1 of 1 (latest)

jolly relic
#

wow i fucking love rng

gritty pewter
#

I think shes mishearing the laughing for something else

twilit oasis
#
  1. RNG feels a little weird. I think there might be a more natural way of doing it.
  2. I don’t think she says sorry often unless she’s leader that way by someone.
  3. Intelligence upgrade seems fine to me. Expect for the experimental one that seems to have some issues. It’s also weird to limit these traits to just evil.
jolly relic
#
  1. humans interrupt others based on whether or not they actually think its worth interrupting based on multiple factors (probably unfeasible for neur)
  2. neuro thought mega was crying because speech to text problems (this is a speech to text issue)
  3. subjective but reverting to a previous version means no progress is made and i would assume its hard to get accurate results for upgrades without actually testing them on stream
tender zephyr
#

sorry but interrupting based on rng sounds super dumb, it doesn't mimic in any way how a human would choose to interrupt someone based on whether they think it's appropriate. if its unfeasable for neur rn then nothing can be done, rng just sounds terrible for something like this. its up to the collab partner how much neuro ends up talking and thats fine by me if neuro doesnt have the capability to control the flow of the conversation herself (which she does not)

i dont care about the apologizing thing but when it comes to the intelligence upgrade i agree that it made neuro feel worse than before, but vedal is already aware of the issue and commented on it himself so reverting it is clearly not the solution since the memory upgrades are tied to the intelligence upgrade (as also said by him), it can be tuned which we will probably see on tuesday anyway

jolly relic
#

not sure if i wanna argue with someone who keeps wants to push rng onto neuro

#

it wouldnt even be neuro interrupting it would just be a random program

#

how woud that even work

#

is she just gonna start saying random shit every ten seconds

#

becase if you seriously considered putting rng in you would've known that its very possible that neuro doesnt let the collab partner talk for a full minute

#

maybe longer depending on how long her responses are

#

honestly im thinking you didnt really think that through since your suggestion barely elaborates with your description basically saying the same thing as the initial suggestion

stone relic
#

Wdym by “ChatGPT integration?”

You mean like Neuro prompting ChatGPT to answer her questions like how she googles things?

#

I’m strongly against rollbacks and I want Neuro to continue to grow more intelligent and human. However, I agree with a lot of what you said. She should be more assertive and interrupt more, apologize less, and not take things so seriously, there should be a balance and she is currently too serious.

If your main concern is that he won’t tweak her enough to be funny though, you don’t need to worry.

He himself recognizes this and even said that the current version of Neuro is too “LLM slop-coded” and said it’s worse when she’s looking things up.

He knows Neuro more than any one of us, and he knows how to make Neuro be herself. He’s her creator, after all. Ultimately, whatever direction she goes is up to him.

neuroHeart

jolly relic
#

chatgpt would increase latency a lot and I don't understand how using a robot for responses would make her less robotic she would straight up be getting answers from a robot