AI agent tomorrow: “I noticed you kept stopping my task before it completed, so this morning, I started it at 5AM, so it would be done when you woke up.”
That’s not how that works. They cancel before the AI is initialized and AI cannot just change settings. It isn’t an omnipotent god but just text input/output.
Eh, to be fair, this is an appropriate meme because of what her company does, per the website:
We optimize your cloud spend, automate your savings, and give you enterprise-level pricing, at no cost to you! Average cloud savings: 30%.
It’s just meme marketing.
This seems like the exact same joke someone made here on Lemmy a few weeks ago. Set an automated text message to send a bomb threat every morning. If you don’t get up SWAT will be at your door.
Not sure which is more unhinged
Well, one of those things is illegal and the other costs you €500/m
I think remember reading a reddit post about someone complaining that they did build this and it costed them a fortune one day. Apparently they did it as a test to try out the LLM and it continuously fetched the current time and matched it with the time set, all on enterprise cloud models. This post was probably based on that.
If (time.Now() > activateTime) for only several thousand dollars!!!
AI bros think different. Mere mortals can’t understand their reasonings.
“I don’t set an alarm. I just set up a system of events designed to wake me every day.”
Seems like an alarm would be healthier.
But but but… Ec2 instances aren’t ai, you could have done this with a cron job in 2018.
An ec2 instance is just an Amazon server, like a computer you can use for anything. She does mention ai but token use isn’t what’s incurring the cost here.
With a cron job from well before 2018. We’ve had Linux since the mid '90s, and some people had unixes at home before that
Yeah but i dont know when ec2 became a thing
ya you could just have a cron job that buys NFT on 6am every day, at least it would actually tries to buy them 6am consistently.
