Charles Zhang
Moment of Thought
Too short to count as “Blogs”!😝
20220903
One of those days when you suddently realize.
(Inpiration from) How Clarisse is super fast and CPU based rendering for its viewport. I was wondering - maybe the 2.5D approach is not GPU based at all - turns out the trick is indeed done on the CPU.
20221209 The Moon
是夜月圆如饼，大且白。
月上有一人，姓吴名刚，喜玉兔。
娥尝问：西河多兔，汝之爱玉兔较爱我之心几何？
刚答曰：爱玉兔者情也，爱美人者性也。
娥曰：夫情者欲也，性者质也，何不直言心之所向？
刚曰：古人制字，性与情皆从心。性即心之理，情即心之用。无理则心乱，无用则心躁。理与用皆不可少。我日用玉兔，夜理嫦娥，生活恰恰乐不可支。
娥曰：分手罢！
20221219
走罢走罢！
20230106
我梦到一个久违的朋友，
她曾陪我度过梦中大半时光。
也曾闲情漫步，
也曾鏖战魔物。
她是高挂北斗的闪耀星辰，
我是穷困他乡的游子浪人。
我已不知去哪里找寻你，
缘此又造访我梦中。
下次若我们再遇到，
万万不教韶光虚度。
欣欣，欣欣，
只因你是我的梦中人。
20230204
图中是一个简单的三维渲染，与寻常流程不同之处在于”界面“——说起来类似于游戏编程，这个画面的搭建主要是通过脚本实现。是否能够当作今后更为宏大目标的基础，尚未可知。
想到十五年前我学编程的第一节课，教的是花三个月时间写一个《西游记》程序——用C++编程的乐趣在于有朝一日所有的框架搭建完成，按一个按钮，看着程序就乖乖按照自己的指令工作。我用过的软件框架无数，使用其他工具少能够有这种成就感。
近来做很多看似实用、功利主义的项目，不论结果多么花哨，然而并没有自己抽空制作的这些精心架构的软件来的实在。
20230304
Arrival and Seeing Future
Arrival (2016) made me cry.
Because I know what it feels like to see the future.[Let Down] What an ambivalent feeling - when you see every single moment of your future in front of you, you are simultaneously acting and observing. It’s a mix of serenity and deep deep confusion. ChatGPT puts it this way: it’s like experiencing a temporary shift of reality where everything seems possible, yet difficult to determine if these assumptions are correct.
(Remark, #20230304) However, if you term it as “Prophecy”, the associated feeling/altitude is quite different.
20230702 Iskra - Redream
Tags: Deeper Feelings
(SHOULD WE PUT THIS INTO “REDREAM” page instead?)
欣欣，你是我的梦中人。
我做了一个梦。
梦中我哭成泪人。
“我本来喜欢的是J，不过如果是你的话，我其实也可以的。”
我都不需要选，因为我在另一个世界中已经喜欢两个你。
(Remark) You always have a spot in my heart.
20230818
Moral Critique on AI Generated Content
What’s so wrong about AI generated content is not the content is wrong - in fact, the contents are pretty good already and likely can become even better in the future. It’s certainly better than some novice creators/writers/painters (e.g. kids).
What’s really wrong is the lack of process - imagine everyone lives in a fully automated world and no one is doing anything. Replacing actual workers with AI TO SAVE MONEY while those workers end up having no job is what is wrong about this world (though argubly it’s also partly those workers’ fault, but then we can always blame everyone on an individual, on the society for everything is a social construct, or on the decision maker). When a company like Microsoft replaces its journalist and asks an AI to generate news articles, it’s making a moral mistake, not a technical one. Reading and enjoying an article, or a painting, or a game, is as much about the creative process behind it as it’s about the content itself.
Similarly, nowadays when Netflix and Marvel Studio are producing shows and movies that are of a bigger production, costs more, and earns more money, they are nontheless shitty shows and movies. Sci-fi movies nowadays are neither sci- or -fi, not because technology is less advanced, but because they are produced too light-hearted and without much thinking. That’s what’s so wrong about the world, i.e. ignoring the process, and expect AI can somehow one day generate the entire software solution for a particular problem. When that day comes our world will be truly doomed - because a large part about SOLVING THE PROBLEM IS FIGURING OUT AND BETTER DEFINE WHAT THE PROBLEM IS. Asking AI to design a bridge with well-defined criteria and constraints - good, asking AI to solve a social issue, that’s bound to cause tons of trouble. Because the problem itself is not what needs to be solved, and that’s something our existing AI technology is ignoring.
A somewhat better formulation is also published on LinkedIn.