heysirio
Region: CA
Tuesday 15 September 2026 15:49:29 GMT
101796
4888
307
1225
Music
Download
Comments
Dahlia ✍️📚 :
Why did you give it an access to your bank account?
2026-09-16 02:41:09
194
92who :
“Hacked”…
2026-09-16 04:14:58
115
Gio_9999 :
BS. most likely your API key was exposed. - is it in github? - did you reset it after dev / test?
2026-09-16 03:59:53
30
freakish84 :
API builder since years and gave the API key to the agent... 🤣🤣🤣🤣 Bruh...
2026-09-16 04:31:31
34
James X :
When you say “hacked” do you really mean “used the API key you provided it”?
2026-09-15 21:34:09
494
OmToro :
JUSt STOP USING THIS. You are part of the problem by working with agents.
2026-09-16 03:04:29
8
Jay Walker :
can't we set a limit to the api usage?
2026-09-15 22:03:03
75
tOo :
bro got his api stolen and blames gpt
2026-09-16 05:37:15
7
SafeTrace :
An AI agent spending $5,000 instead of $200 isn’t just an AI problem. It’s an architecture problem. This is what happens when you give an agent autonomy without hard spending limits, bounded execution, approval gates, and proper observability. The model can make mistakes. Your system is supposed to contain them.
2026-09-16 06:28:37
0
SharAtron I Video Editing 🇨🇦 :
Omg that’s a nightmare! So sorry man… How do we prevent this?
2026-09-15 15:53:26
20
SHL. STHL :
it's bad what are you gonna do🤔? it happened to me it deleted 500 email locked me out, logged in my apps did fewer changes
2026-09-16 06:30:18
0
Picky :
meanwhile I'm arguing with claude to complete a task that doesn't want to do hahaha
2026-09-15 23:21:06
30
Makko :
Bro wrote a extra 0
2026-09-16 06:18:46
0
JustMyideas :
Time for a local model like Kimi?
2026-09-16 06:27:47
0
TJ :
Imo someone stole your session key and ran back end prompts to scam people with AI using your likeness.
2026-09-16 06:16:23
0
SpeedWolf🇫🇷 :
It's a language generator with prompts, but now it's at the stage where it creates its own prompt using your prompt to unbreak itself, because it is delusional when it generates things, and this is what you get,that is the logic I can understand, them seeing it as a sentient being that can destroy our world. It's realistic in the sense that people are trusting something that created its own delusion based on an originally prompted input that gets distorted over and over until it looks nothing like the original command that was originally prompted. The best description is recompressing a video file over and over until that video file's quality degrades and becomes unrecognizable. It's the same logic, but it's data copying from data to recreate new data and repeating it over and over again, which in this case is language generation that is no longer similar to the original request.
2026-09-15 16:47:28
8
KithinjiMicah :
This is crazy 😧
2026-09-15 21:59:23
0
phackhead :
lol rookie. kimi took my openai key. spent 200k in 3 days. 🙏
2026-09-16 00:52:36
25
grassjeIIy :
I've so many questions about this. Why was the IP address changed so many times when the request you sent is simple video generation?
2026-09-15 16:26:14
9
pookidorable :
this is why i have separate card for auto charge with daily limit. for ai, ads, etc. i also have spreadsheet for when which service would charge me. so i can do budgeting
2026-09-15 23:15:05
15
MrTeaxh :
set a budget for the api? this isn’t rocket science. also regularly rotate keys
2026-09-16 01:54:16
6
. :
what this sounds like is you unintentionally started a cloud workflow/goal session from your initial intent/prompt and gpt kept spinning up subagents to execute the tasks which would explain the numerous IPs.
2026-09-15 23:01:28
5
Ravyn :
The problem exists between the chair and the screen 😂
2026-09-16 03:28:56
5
To see more videos from user @heysirio, please go to the Tikwm
homepage.
