Search Results
7/2/2025, 8:46:10 AM
>>529620706
(Continue) is good enough, though. That's literally all the prompt you need once a few things are set in motion. You can also train Deepseek to write it's own instructions based on your general instruction format, or end each generation with a break and something like a "In the next section, Clarence the fuckbuddy will get titslapped by the Fox Mother he besmirched earlier..." as a guiding DNA after each generation. I also find Deepseek V3 actually is pretty good at past context recall, and critically, it has an insanely high max context and is cheap enough that there's basically no sweating cranking that shit to maximum.
>>529620856
I asked you to define co-writing, I never said it was important or exactly what it is. Why is any of your methodology more 'proper'?
>>529620857
Is your argument entirely that the more effort it takes, the better it is? Does someone who drives a car only want to put in minimal effort to get somewhere? I mean, yeah? Isn't that why we're using AI tools?
>>529621409
I used Dragon well before I found /aids/ and started posting here regularly and sharing prompts. I've always been a Dragon detractor though. It was fun, but not that great. I actually liked a lot of NovelAI's early models more than Dragon, and I had quite a lot of Dragon experience.
(Continue) is good enough, though. That's literally all the prompt you need once a few things are set in motion. You can also train Deepseek to write it's own instructions based on your general instruction format, or end each generation with a break and something like a "In the next section, Clarence the fuckbuddy will get titslapped by the Fox Mother he besmirched earlier..." as a guiding DNA after each generation. I also find Deepseek V3 actually is pretty good at past context recall, and critically, it has an insanely high max context and is cheap enough that there's basically no sweating cranking that shit to maximum.
>>529620856
I asked you to define co-writing, I never said it was important or exactly what it is. Why is any of your methodology more 'proper'?
>>529620857
Is your argument entirely that the more effort it takes, the better it is? Does someone who drives a car only want to put in minimal effort to get somewhere? I mean, yeah? Isn't that why we're using AI tools?
>>529621409
I used Dragon well before I found /aids/ and started posting here regularly and sharing prompts. I've always been a Dragon detractor though. It was fun, but not that great. I actually liked a lot of NovelAI's early models more than Dragon, and I had quite a lot of Dragon experience.
Page 1