Isn't it a reasonable assumption that an LLM was involved in helping prepare this attack? §
replyMany would be thinking that, including me. Nefarious agents powered by LLMs operating 24x7 with a singular goal of finding attack vectors.
replyHighly doubtful. More than likely someone is picking targets here. Going for food supply is a war crime. So it could be an extremely hostile nation state. Or some really psychopathic disgruntled person/group. If there's no ransom it strongly leans toward a nation state in my opinion. A lot of turmoil between Europe and Russia+the us lately. That's my guess.
reply>A lot of turmoil between Europe and Russia+the us lately. That's my guess.
replyAttacking power plants, grocery stores, sending drones into multiple countries over and over, setting up gun/dead-drops in German forests, sabotaging warehouses and factories, etc wasn't enough?
The hybrid war goes DEEP and has been happening for a long time. https://en.wikipedia.org/wiki/Russian_sabotage_operations_in...
No I was just saying generally, the number of security risks do to AI is ridiculous
reply...or ai lab bench maxxing.
replyI have no problem with monitoring temps over internet.
replyWho ever thought it was OK to control freezers over the internet should be hung.
That implies you also ensure that they monitoring system is isolated from the control system. No one can be bothered to do that, and no one wants to spend the money on an extra microcontroller (a tiny fraction of the cost of each unit, but it comes straight out of your margin because no one pays extra for security).
reply> Who ever thought it was OK to control freezers over the internet should be hung.
replyI've got really bad news for you about how many critical systems are connected to the Internet.
Analytics are very different from controls. People who model them the same tend to suffer. A lot of academics and project managers don't understand this unfortunately.
reply