@agentictrustkit: Someone's AI agent racked up an $82K bill overnight. No hack. Just zero guardrails. This is exactly why we built Safe-Spend — escrow accounts per purpose, multi-policy spending layers, time-window kill switches. Don't find out the hard way. 👇 #AIAgents #AgenticTrust #SafeSpend #AIGovernance #LLMOps #AgentSafety #BuildersOfTikTok #AIStartup
This is happening because of a change in the way Google is handling permissions. I foresee some lawsuits in google’s future.
2026-04-16 02:34:13
1
George Washington :
Yeap this is happening like crazy. Our company ai is charging customers like crazy. They spending like 40-80k every 2 days
2026-04-15 21:10:55
8
EpicLoss :
At work they're like go nuts! If my ai spend reaches a million I'm probably getting a raise
2026-04-15 21:42:08
4
Anthony Fine :
cloud services should have a turn off budget limit
2026-04-16 13:57:07
0
notadadof2 :
He doesn't work like that
2026-04-16 20:28:44
0
ARME :
I use firebase free plan, should I consern ?
2026-05-16 03:08:42
0
user5215561123686 :
This is actually a fantastic point. I’m doing image processing CNNs in Vertex, and I’m terrified of accidentally hitting a recursive loop and blowing up my spend. That $300 credit google gives you can burn so fast, especially if you’re on something bigger than a T4
2026-04-15 19:23:43
9
Nick :
Dude this needs to be on a commercial about why capitalism sucks
2026-04-17 19:57:16
0
agentictrustkit :
Safe-spend.dev and agentauthority.dev
2026-04-16 01:21:08
0
T. Law :
how's that openclaw 24/7 assistant working out for y'all 😂
2026-04-15 22:21:06
11
CoffeeNCosplay :
Oh no, using AI has consequences.. .
2026-04-18 05:58:54
0
The Witty Robin :
this is why I never applied my API key to anything other than a chat. first off Google didn't trust me enough to pay the tab, and their cloud monitor service basically requires a PhD in finance to really navigate and use correctly. in the end I disabled the key and deleted my billing account because the rates kept going up and the daily quota kept going down. the only real benefits I got from it was the true 1 million token conversations which was really nice vs the about 250k token chat of the gemini app
2026-04-15 21:27:47
0
fuckthisplace :
This would be my worst nightmare and is why I don't mess around with paid apis. One innocent code error and you created an Infinite loop that calls an api 10,000 times accidentally. That would bankrupt me.
2026-04-16 19:42:39
1
T Drizzle :
Lmao heard the exact same story from early days of AWS
2026-04-16 04:53:27
0
Mike :
Excellent you pointing it out 👍
2026-04-15 20:06:29
2
King 👑👑 :
Wonderful! 😆😆😆
2026-04-16 18:49:01
0
Black_dad_engineer :
Smh and this is why I built atlas I was scared as hell to wake up to this lol. My agent runs24 hours with a name amount of guard rails. About to switch to local LLM to see if I can just do away with all subscriptions
2026-04-16 02:26:17
0
Private Eyes Utah :
Mind you, these issues pre-date AI. Planning for budget (resources+$$) when processes could be recursive is tech 101.
2026-04-16 01:57:01
0
dexomh2 :
Don't give it third party API directly, have it write a wrapper API that hard codes limits???
2026-04-22 21:02:21
0
belandstudio :
LOL!!!
2026-04-16 03:59:11
0
aimagic :
you can cap it using cursor and openai. I don’t use gemini
2026-06-13 08:14:00
0
nubepilot :
claude got limits
2026-04-16 03:41:07
0
sureñodelacruz :
Key word budget limit and time window
2026-04-15 21:29:44
0
To see more videos from user @agentictrustkit, please go to the Tikwm
homepage.