A Little Background#

The webmaster subscribed to GPT Plus. Besides happily using Luna and asking Astra to make plans, the web app itself is something I really ought to make use of. So far, besides everyday chatting and searching, I have thought of a few things I can do with it:

Use chat mode to help me study for make-up exams. Although, honestly, AI is mostly useful for recommending videos and finding study materials. If you ask it to write practice questions, it absolutely loves making easy questions in formats that will never appear on the exam. That can make you far too confident right before the real exam cuts you down. Don’t ask how I know QAQ
Give me a daily briefing on fields I care about—the classic newspaper-reading experience.
Recommend one offline challenge or activity every week, so I can shed my social anxiety and evolve into an absolute beast.
Check my online services and help operate my servers every week, although I use Codex for that.
Recommend some long-form articles every week. I realized that I have not sat down and read a book properly for a long time. I bought books about game design and other subjects, but I only really read them back in high school. Since starting university, it honestly feels as though there is no time to read at all. So I decided to begin with long articles and rebuild the habit of sustained reading.

If you have other ideas for scheduled tasks like these, feel free to leave them in the comments!

And that brings us to the first weekend reading note. For now, I plan to update this series every Thursday afternoon, since that happens to be an empty slot in my class schedule!

The Articles#

Let’s start with two this time!

‘Ruin and decay bring a chance of rebirth’: Hidetaka Miyazaki on uniting players in a damaged world

feeling-sad-about-ai/

Reading Notes#

’Ruin and decay bring a chance of rebirth’: Hidetaka Miyazaki on uniting players in a damaged world#

“There’s no intentional FromSoftware [mantra] – ‘this is the way we make our games’ – because if you start to do that it becomes stale and conservative, and you start having to adhere to a set of rules and philosophies, and it just doesn’t become interesting, you can’t make anything that’s as fun,” Miyazaki says.

Miyazaki says FromSoftware has no doctrine declaring, “This is how we make games.” Once you begin following a fixed philosophy, the work becomes stale, conservative, and uninteresting.

My own understanding is that perhaps I should first make the gameplay I enjoy for my Steam game, and only then try adding things on top of it. (Laughs.) If I decide everything in advance only to discover during development that it is not fun, I will have wasted even more time.

“It’s so incredibly hard to do in video games, which are technical systems … But Miyazaki somehow creates an atmosphere of wonder. Even when finishing, you are left ruminating on fun, murky elements that feel half understood. It’s a delicious sense of intrigue.”

The worlds of the Souls games are not built by spelling out an entire setting bible. Instead, enemies, locations, and item descriptions leave clues for players to piece together themselves. Even after finishing the game, players remain in a state of “half understanding,” so they keep thinking about it.

As expected, leaving things unsaid is a very sophisticated writing technique. It works in all kinds of creative work: just the right amount of empty space can hook the audience’s curiosity, and once they become curious, they are already not far from falling in completely.

Feeling Sad About AI#

The author argues that money and time which could have gone toward developing clean energy have instead been spent on AI. Emmm, it is a fairly classic anti-AI argument… but of course that was not the only part that stayed with me.

What struck me most was the author’s love for programming as a craft in itself. I used to know that feeling very well. Back when I knew nothing, all I could do was follow videos step by step. When I only half understood a concept in a video, I had to sit there and puzzle over it, watch other similar videos, and try things myself until it finally made sense. Learning a new concept genuinely felt like a revelation. It was incredibly satisfying, and I kept leveling up and taking on new challenges that way.

Then AI arrived. Compared with moving forward one step at a time and overcoming every obstacle yourself, AI feels like a constantly upgraded, absurdly powerful lucky break. After trying it, I went through plenty of internal struggle. In the end, well, I kept riding that lucky break all the way to the present—and it really has been exhilarating. This blog itself is one example: from the domain and server to the self-hosted tools, I built nearly everything while asking AI questions and stumbling through the problems myself. Along the way, I joined all kinds of AI and technical communities and saw things I had never encountered or even imagined before… I barged through every new field I once would not have dared to dream about, and that also helped me accumulate a great deal of technical experience and infrastructure.

I cannot oppose AI as decisively as the author does… Yet AI really has made me less patient with programming, more inclined to reach for AI assistance and even outsource parts of my thinking… This very article is an example. Still, there is nothing inherently wrong with learning to command a tool, is there?

So in the end, perhaps the real issue is controlling the scope. In areas I do not consider important, using AI to get started and implement something quickly is naturally the most efficient choice. But in areas I consider important, or when I genuinely want to master a particular technology, I should not pursue implementation with efficiency as the only goal. Doing that would leave me with only a shallow understanding of the very thing I wanted to learn, until the accumulated problems eventually caused everything to collapse…

Afterword#

While writing this article, I was interrupted by a notification reminding me to complete a psychological assessment and comprehensive evaluation. Then I remembered GPT had been unstable lately and considered buying a yearly GLM Lite subscription. After talking it through, I decided not to buy it and to try the free quota first. Then I chatted with a friend about game development and career choices… From all that, I reached two conclusions:

Damn it, I really do need to figure out how to make some money. (Sigh... After all, the reason I have not done so many things is simply that I do not have enough money. But can I actually earn any...? Even I am not optimistic about myself (collapses)
Reality and the virtual world are tangled together after all. These days, blindly throwing yourself into one while neglecting the other will not end well…

I also discovered that this kind of article is not especially easy to write. Maybe I should switch to reading one chapter from a substantial book instead. I could return to The Art of Game Design… or review Level Up! The Guide to Great Video Game Design? Let’s do that! Still, recommending articles like these is interesting too, so I will keep it as an occasional side branch.

Oh, and if anyone wants to send me something they wrote, you can simply send me the URL! I would love to read all the different things you have been thinking about.