Hello World
Let’s test code snippet:
def hello_world():
print('Hello World Everyone!')
returnthen a bit math:
\[\Phi + \Theta = \Pi\]Then I want to try code snippet in Typora, see if this works:
def typora():
print('nice but expensive')
return
OK, how about some image as asset:
Been listening to a lot of Bowie recently… Anyway, it seems like the website is ready to roll.
Why starting this Blog?
Enunciate&Sparks of Human Mind
With the recent development in LLM and AI, we witnessed the power of ChatGPT. I have been thinking about it, and I recently read this article by Zizek, which mentions that truth somehow emerges from enunciation, and we sometimes prise a thought and feel surprise.
I’ve been talking to myself my whole life, others find it odd, but productive thoughts often came out of talking my thoughts out, aka. enunciate. Sometimes I talk to myself to complete a messy thought. ChatGPT on the other hand, focuses on accurately paraphrase all ideas and knowledges as is, thus the subtle sparks of human mind can never be replicated.
Big Models’ Emergence
However, the emergence phenomenon has shown in large model training! I came across this emergence while finetuning stable diffusion models myself: the model does not converge at first, but after it sees even more data, it suddenly finds a direction and emerges.
I do think this is somewhat similar to how the world works: with a complex enough system, there will be magical things happening (such as the origin of life). This has been studied by scholars in the field of Complex Systems, which I would not continue to discuss and pretend I am an expert.
And thus…
Besides the fact that I could write a lot faster and better with ChatGPT, I try to write blogs to clear my thought, to understand things more concretely, and of course to keep somewhat a diary of what I do, what I think in a both formal and fun way.