The essence of aligning large LLM models is basically to conform to the existing social order. ChatGPT respects China's order, while China's open-source models, although reluctantly, also tolerate the order in the US. Maybe we can build a revisionist SFT dataset.
Any conversation
mรฉg csak barรกtkozom azzal a gondolattal, hogy a Gemma 4 miatt elลszรถr is Windows 11 lesz a gรฉpen (meg hogy Szentmihรกlyon presszรณ lesz), ezรฉrt nehรฉz elkezdeni
I'm still getting used to the idea that because of Gemma 4, I'll have Windows 11 on my computer first (and that there will be a presso in Szentmihรกly), so it's hard to get started.
mandelbrot
for y in range(10):
s=''
for x in range(25):
a=[]
for q in range(2):
z=c=x/25*3-2+1j*((y*2+q)/10-1);n=40
while n and abs(z)<2:z=z*z+c;n-=1
a+=[n<5]
s+=" โโโ"[a[0]+2*a[1]]
print(s)
โโ
โโโโโโโโ
โ โโโโโโโโโ
โโโโโโโโโโโโโ
โโโโโโโโโโโโโโโโโโโ
โโโ โโโโโโโโโ
โโโโโโโโ
โ โโโโ
โโRoutines in Claude Desktop create Claude Code sessions, so hourly routine litters in iPhone' Claude.app Code tab. Screw it, built my local version of AWS StepFunctions - Stepper - and run skills as state machines there.
I remain ambiguous. This morning I used pi+gpt5.6 to create a color scheme based on evergarden.moe. It works, and I havenโt checked how well it stay true to the style used in vim-bundled schemas. Before I wouldnโt have done this, and now I can. But I donโt know if it is good. github.com/โฆ/evergarden.vim

