Everything online is paid for, if you have a decently modern computer, its worth learning offline stuff like comfyui is pretty good for image/video generation stuff.

Apparently you can even download some offline LLM's and load them into something like lm studio, but I havent actually got that to work yet. When I tried the install for lllama on facehugging I think its called, it was 200 gigs and I cancelled thinking I screwed something up, but no I think thats what it takes to have a actual "offline" llm experience. When using something like chatgpt's model in the lm studio, its not running the entire model, its just running the weights and input that feed into it, which is handy if you dont want to login/pay to use/store chatgpt chats, but you can only input/output so much before you run out of resources. I read recently chatgpt is going to start putting in ads into their chats and I wont be using it anymore at that point.

I really hate how everything is censored too. Ive been thinking about trying to use veracrypt to store some important documents(like the work stuff for example), but in the end I decided its to important I cant risk a random muon coming through and locking the entire contents. Just easier to offline it.

Chatgpt absolutely refused to elaborate on veracrypt support, no amount of rephrasing would make it talk, the government must really hate that shit.

Everything online is paid for, if you have a decently modern computer, its worth learning offline stuff like comfyui is pretty good for image/video generation stuff.

Apparently you can even download some offline LLM's and load them into something like lm studio, but I havent actually got that to work yet. When I tried the install for lllama on facehugging I think its called, it was 200 gigs and I cancelled thinking I screwed something up, but no I think thats what it takes to have a actual "offline" llm experience. When using something like chatgpt's model in the lm studio, its not running the entire model, its just running the weights and input that feed into it, which is handy if you dont want to login/pay to use/store chatgpt chats, but you can only input/output so much before you run out of resources. I read recently chatgpt is going to start putting in ads into their chats and I wont be using it anymore at that point.

I really hate how everything is censored too. Ive been thinking about trying to use veracrypt to store some important documents(like the work stuff for example), but in the end I decided its to important I cant risk a random muon coming through and locking the entire contents. Just easier to offline it.

Chatgpt absolutely refused to elaborate on veracrypt support, the government must really hate that shit.

Everything online is paid for, if you have a decently modern computer, its worth learning offline stuff like comfyui is pretty good for image/video generation stuff.

Apparently you can even download some offline LLM's and load them into something like lm studio, but I havent actually got that to work yet. When I tried the install for lllama on facehugging I think its called, it was 200 gigs and I cancelled thinking I screwed something up, but no I think thats what it takes to have a actual "offline" llm experience. When using something like chatgpt's model in the lm studio, its not running the entire model, its just running the weights and input that feed into it, which is handy if you dont want to login/pay to use/store chatgpt chats, but you can only input/output so much before you run out of resources. I read recently chatgpt is going to start putting in ads into their chats and I wont be using it anymore at that point.

Everything online is paid for, if you have a decently modern computer, its worth learning offline stuff like comfyui is pretty good for image/video generation stuff.

Apparently you can even download some offline LLM's and load them into something like lm studio, but I havent actually got that to work yet. When I tried the install was 200 gigs and I cancelled thinking I screwed something up, but no I think thats what it takes to have a actual "offline" llm experience. When using something like chatgpt's model in the lm studio, its not running the entire model, its just running the weights and input that feed into it, which is handy if you dont want to login/pay to use/store chatgpt chats, but you can only input/output so much before you run out of resources. I read recently chatgpt is going to start putting in ads into their chats and I wont be using it anymore at that point.