This is even more funny. Oh no, 15 TV boxes in the same network can produce 0.39 tokens per second while being completely computationally overwhelmed? Imagine all of your house appliances not reacting and you not catching on that there might be an issue that requires factory resets.
Not to mention the network lag and how often small models need to reiterate stuff because it got stuff wrong.
Exactly. People often forget that one of the reasons some worms were actually able to propagate that well a few decades ago was their size and efficiency.
IIRC a lot of them were under 1KB in total, a few dozen or hundreds of bytes. And this allowed them to slip through network and local file transfers unnoticed.
A replicating genAI model is like a semi-truck trying to slip through that tiny back alley next to your house at night while you can see all the fumes coming out from its exhaust and occasionally hear their obnoxious reverse alarm going off.
Yeh lol, what is a worm virus? This really is the .com bubble all over again
This is even more funny. Oh no, 15 TV boxes in the same network can produce 0.39 tokens per second while being completely computationally overwhelmed? Imagine all of your house appliances not reacting and you not catching on that there might be an issue that requires factory resets.
Not to mention the network lag and how often small models need to reiterate stuff because it got stuff wrong.
Exactly. People often forget that one of the reasons some worms were actually able to propagate that well a few decades ago was their size and efficiency.
IIRC a lot of them were under 1KB in total, a few dozen or hundreds of bytes. And this allowed them to slip through network and local file transfers unnoticed.
A replicating genAI model is like a semi-truck trying to slip through that tiny back alley next to your house at night while you can see all the fumes coming out from its exhaust and occasionally hear their obnoxious reverse alarm going off.