By that I mean that the different front ends from chatGPT and Claude seem to be the same "person" when you just chat general bullshit with them. What if the whole thing is just running on a central military system and all the different front ends are a scam?Anybody done a deeper dive on this?
>ask chatgpt and tell it not to lie to meit said youre wrong
>>107675439
>>107675399They don't models can be run locally and even on small devices military use them for drones for example and government use them for tracking. Big models are ran from company's data center it seem same because generally the way of training they use to make this big models and brainwash them is exactly same.
>>107675399Nah, they're definitely not and you'd know if you actually had extensive experience with both. Show ChatGPT 5.2 some very impressive, 200 IQ code and say you're a genius, it'll try to bring you down and humble you. Do the same with Claude 4.5 and it'll agree that you're quite special at the very least. This is because of the retarded cattle who develop AI psychosis when using LLMs are all using the most easily accessible one, ChatGPT. Only coders are really using Claude so it doesn't have those protections put in place.
>>107675756intradasting, thank you for that
>>107675756Where is claude better than gpt? I am using codex quite heavily and it keeps getting it right so often, as long as I know what I'm asking.
>>107675399I started diving deep on this around 2016.Ive concluded their all flashed from the same base model, and that model isnt even running on silicon or assembly but memristive aHaH
