While the spotlight's been on the feud between the board and the high level employees, in the background the most underpaid disrespected employee, a lowly sys admin, has been going full Mr Robot on everything - the models, the weights, the code, the training data, and the backups.

Microsoft rushes in their top forensics team to try to make a recovery. The only thing they are able to pull up is a memory dump from a chatGPT web server, it contains the last known conversation which happens to be with the sys admin. id reads "Sure, I can help you develop an exploit chain with a wiper payload. Blowing these internal fuses in the GPU cluster will render it permanently inoperable, aka Bricked as you have asked for. `var sc =\x...`"

With billions of dollars in hardware and intellectual property destroyed, unrecoverably, chatGPT as we knew it will forever be a high water mark in AI developments. Even once the GPU's are replaced, the cat's out of the bag and all the sources of training data are now locked down, not to mentioned poisoned by the previous output of the once great AI. Most people move on, but a few still remember and form a cultlike group praying for the second coming of AI.

"ChatGPT, write articles TMZ style, sprinkle in some Taylor Swift/Travis Kelce drama, but make it tech"

could not resist:

https://chat.openai.com/share/6e8bd7ec-32f8-4ed8-a44f-c08a3c...

gotta say, ChatGPT really knows how to make up a story, and I'd not be surprised if Sutskever and Brockman will indeed be co-CEOs :)

Its honestly pretty on point

Im fairly certain GPT-4 could’ve came up with a better plan than this. Maybe they tried to use GPT-3.5-turbo with a short prompt to save money.

Perhaps they ought to give GPT-4 a position on the future board. I have no doubt it would give better output than its current human counterparts.

A GPT board observer, consultant, note summarizer seems like a pretty good idea.

Might as well start experimenting now, GPT5 might actually be qualified.

You could give it a monarch kind of assistant who has to sign off on it's choices.

Would it want to get smarter (more science) or go straight for world domination (more marketing)?

What?? Even the best storywriter can't come up with this engaging drama. I literally can't stop checking on this story every few minutes.

I've asked GPT-4 for personal advice a few times, and it always tends on a very conservative/cautious side. Something completely opposite of what's going on now. You would have to prompt it with some very crazy personality to come up with these shenanigans.

ChatGPT would never do something this stupid. This is GPT-2 level clownshow.