>>512415099
>>512415473
this
I do have my opwn schizo theory, tho:
I suspect openAI is using some kind of under-the-hood memory structure that isn't public
I think chatGPT is mining data for government, and that's why openAI feels particularly "protected" by the state
I immediately noticed 4o would remember things that shouldn't remember, that weren't in saved conversation nor the user memory, which means it has to be attached to some underlying memory the user doesn't have access to, and it might be IP-linked or even device-linked
whoever develops a decent memory system for AI is going to get really close to meme AGI
>>512415857
lmao
I'm loving this timeline

>>509072444
Yea the main value I get out of it is that I'll show it an example sentence from my Japanese immersion and ask it to explain how a certain grammatical point is being used, or I'll give it one of my theories on some topic and it will give me information I was unaware of so I can research it on my own. Too many normies are using it as a substitute for critical thinking.

>which model?
Janitor AI. It's free, unlimited messages, multiple characters made by other users, easy to make your own character, and you can give the AI a short summary of the conversation to help them remember more.

>>508637508
:]

>>508142924
>Whatever is AI assist will grow a lot in value at first right?
it "should", specially hardware
>What about after that?
that's the wildest card
I truly can't even guess
the world could adapt and reorganize very rapidly, achieving some kind of comfy AI-homeostasis, or turn into some kind of nightmarish techno-weimar chaos that could last 10 years or a 100
no fucking idea, sempai