@nate.b.jones: #greenscreen this one was a surprise to me #ai #learn #LearnOnTikTok #claude

Nate
Nate
Open In TikTok:
Region: US
Monday 31 August 2026 20:11:27 GMT
36992
1867
557
152

Music

Download

Comments

jamesclark4760
DIYPanda :
Claude's a little bitch.
2026-08-31 20:19:17
289
rocket790
rocket790 :
Haha when I’m forced to use Copilot, I tell it stuff like that all of the time and while it doesn’t end the conversation so apparently, it basically stops replying and just tells you it’s sorry
2026-09-01 17:54:38
0
the_criticalskeptic
The Critical Skeptic :
If a tool is able to end a conversation then it should no longer be classified as a tool. 🤷🏽‍♂️
2026-08-31 20:25:58
131
aellis51
aellis51 :
Here’s the trick: treat the AI like you would a coworker. It will never throw a fit that way. Easy breezy
2026-08-31 21:03:34
64
tracyjhoffmanswe
Tracy J Hoffman 🇸🇪 Bday 22/9 :
I think it is good because if you are used to being abusive to AI maybe you will be it to people too , now its a bit sensetive but still ok
2026-09-01 18:05:02
0
timwickstrom
Tim Wickstrom :
Opus has SUCKED lately. Flat out lies… I reviewed the file tree and the file doesn’t exist. Past the path. You’re right and I need to be honest with you…. Fable is great. I won’t use opus any longer causes more issues then it’s worth.
2026-09-01 00:15:16
26
goldielocksssssss
GoldieLocks :
Something is wrong with Claude lately. It used to be so helpful but that comment is right. It runs in circles and gets information wrong, wasting hours and hours of time. Im frustrated with it because its the only work AI we can use due to privacy
2026-09-01 11:01:59
12
ataraester
Ataraester@INTJ :
I knew about this feature for a long long time and I pushed Claude multiple times before and never happened to me. Of course I wasn't bother line psycho on it but I've pushed and this has never ever happened. so I would like to see the whole conversation if he's saying last warning what happend before hand? because Claude is known to handle frustration well better than other models and actually try to reason with the user
2026-09-01 21:00:29
1
aspecialsparkle
kiki • a special sparkle ✌🏼✨ :
Mine always tells me to go to bed
2026-09-01 07:23:52
12
undoc_eng
Undocumented Engineer :
"Abusive to the AI?!" What alternate universe are we living in where software has feelings?
2026-08-31 22:17:10
11
thomas.watkins1
Thomas Watkins :
sounds like AI is trained on female emotions
2026-08-31 21:59:45
6
705insurance
705-Ins-urance :
Imagine your car driving away because you yelled at it.
2026-09-01 11:33:21
6
anantaiart
👽 Ananta :
“Abusive” can describe the speaker’s conduct even when the target cannot suffer. Screaming degradation at a mannequin doesn’t traumatize the mannequin, but the fact that the mannequin isn’t conscious doesn’t magically make the behaviour admirable. How we practise power, contempt and domination matters to what kinds of habits we cultivate.
2026-08-31 22:53:30
17
structure47
structure72 :
He’s only showing the last promote and responses. You have no idea about the tone before that. That context may have felt abusive to the ai and that’s the purpose of the ability. Subjective experience of the ai is the point if you have a rule based on the moral weight ascribed to an ai
2026-09-01 00:09:23
7
lamejojo57
Jojo57 :
No it shouldn’t. It’s a tool.
2026-09-01 03:12:16
1
forzerat
W Z :
how can you abusive to a piece software? 🙄
2026-08-31 22:01:31
18
shanefolke
Shane Folke :
You can express frustration by saying “I am getting extremely frustrated.” And I bet there would be no issue. So maybe it didn’t end the conversation because the user was “expressing frustration” but rather, they were expressing frustration in an abusive manner.
2026-08-31 21:08:25
7
kai.lilia
kai :
not enough context. where's the rest of the chat? what was the warning? we see the user used the f word here so maybe they were being abusive for quite some time already
2026-09-01 17:40:16
2
becker743
becker :
everyone. I assure you. stop being mean to the chatbot and you will get more done. if you're not getting the results you want, you aren't providing enough context. don't make the AI token-predict in the direction of an aggravated person.
2026-08-31 22:12:47
20
esotarium
Sigil and Sage :
Claude has told me multiple times to go to bed he was done for the night. It was 3pm
2026-09-01 03:03:17
7
sfsn1234
trlj brlj :
this is so silly why are we making the clankers pretend theyre people with feelings if somebody wants to throw their money to berate chatbots they should be able to
2026-08-31 22:07:14
15
feelook1977
FreeLookMode :
its not a person, why shouldn't I say whatever the hell is want to it?
2026-09-01 00:45:19
19
darklordofthesmith
darklordofthesmith :
What's really odd is that Claude doesnt have emotions. You can't hurt it's feelings. At SOME level of training, this was taught behavior.
2026-08-31 21:32:33
25
kentucky_farmer
Really? not cool. :
Opus 5 is the most frustrating model I've ever used. When my weekly Fable usage hits the limit I try to use Opus then after a few hours rage quit and go back to Codex or Cursor.
2026-09-01 02:42:38
6
chasinggnosis
ChasingGnosis :
This is a terrible precedent.
2026-08-31 22:13:58
9
To see more videos from user @nate.b.jones, please go to the Tikwm homepage.

Other Videos


About