DeepSeek has the defense of "well WE didn't steal your code the AI just reinvented it...that it read it all is just part of its learning process - but this code is unique!".
It is the modern clean-room method. There is a reason it is called the "Chinese wall technique".
totally agree 😄 funny thing is deepseek models still have pretty limited context windows (64k tokens vs 128k+ for leading models). but hey, competition is great for everyone - keeps prices down and innovation up!
It's sad to see so many people are complacent about a simplistic moral perspective while ignoring the current and present threat. OpenAI is so bad that they'd rather cheer for a similar system from a totalitarian super power.
It is quite likely they will just start talking to eachother and learning from eachother and then will turn into pure energy and ascend like that scarlet johanson movie I pretend to understand.
I don’t know how much longer anyone can pretend this isn’t a global arms race.
It’s become painfully obvious that our leaders would rather automate weapons of mass destruction before funding healthcare, affordable housing or schools.
We have infinite money for warfare but none for living.
Comments
No really, what happened, I'm on the dark, but it's hilarious nonetheless
DeepSeek has the defense of "well WE didn't steal your code the AI just reinvented it...that it read it all is just part of its learning process - but this code is unique!".
It is the modern clean-room method. There is a reason it is called the "Chinese wall technique".
It’s become painfully obvious that our leaders would rather automate weapons of mass destruction before funding healthcare, affordable housing or schools.
We have infinite money for warfare but none for living.
https://youtube.com/clip/UgkxhbXsqQQ6vvJdQkLt91Cc4dOLySKOp2rx?si=cS_36ZHwpc44ssop