File: 1683581616364.png (64.9 KB, 1231x387, gpthread.png)
No.590782
File: 1683582701472.gif (292.93 KB, 640x360, 1419650607787.gif)
I just tried GPT for the first time last night, because I found this proxy site that doesn't require an account
https://chatgptproxy.me/#/I argued with it about evolution and the age of the Earth, but it's no use, it's been hard programmed on this subject to just keep repeating the same copy/paste response argument from authority.
It's verbosity and silver-tongue methods of presenting wrong information got me kinda mad tbh.
No.590786
File: 1683583971293.png (1.49 MB, 2400x7040, whatmin.png)
I am an incomparable homo
No.591055
File: 1684180100746.png (2.02 MB, 1000x1247, 1684106611296789.png)
Hey if you guys want a to run an uncensored large language model on your own machine it's actually easy now.
1) llama.cpp is a portable C++ program with no dependencies that compiles by just running "make" on any Unix system
2) Run it and load a model .bin file, you can find models on HuggingFace
You only need 16GB of RAM and any half-decent CPU, no GPU required. This model is probably the best available right now
https://huggingface.co/TheBloke/Wizard-Vicuna-13B-Uncensored-GGMLIt's pretty smart, not as smart as ChatGPT but it gets close, it's leagues better than anything available even a few months ago before the data/model from Myspace called "llama" leaked and sparked new interest in open-source AI langauge models.
You can tell it to write up a story about whatever you want.
No.592274
File: 1686164472422.png (24.88 KB, 1364x484, 2023-06-07-150040_1364x484….png)