天空вℓσи∂ [Deprecated edition] @skyblond@social.cleverlibre.org

Start wondering if I should use some managed database.

I set up my current database using a LAMP script 5 years ago. And I never touch it again. Of course, no backup too. (LOL)

In the past, I have a lot of time and ENERGY to do server stuff, while now I just want to deploy and forget, I don't want to touch the maintenance thing anymore. I mean, setup is pretty straightforward, but try not to forget what you have done on a server is hard, especially if it's an all-in-one server and without docker. Maybe a few years ago a random setting I changed during debug will screw me up today.

Finished this quick weekend project:
https://github.com/hurui200320/GuliKitAPG

Some simple proprietary file format "reverse engineering" (I would say try and guess in plain English).

I thought it can replace an ESP 32 in Splatoon auto drawing, but the limitation is too bad. However, if you don't really need an ESP 32 and are willing to worst thing in automation, aka repeatedly reloading the controller, this might work. At least it's a controller, you can still play games with it. For an ESP 32 with normal people who don't want to step into this developing, it's just another E-waste.

@freemo SEND TU DE?

LTT‘s channel is back!
----
My Channel Was Deleted Last Night
by Linus Tech Tips
https://www.youtube.com/watch?v=yGXaAWbzl5A

Wow
----
We Built a BULLETPROOF John Wick Suit in Real Life!
by Hacksmith Industries
https://www.youtube.com/watch?v=Eeb4aZObp-0

maybe I should run a quick test to make sure a single cnn layer can do what I'm expected. Or do I just waste my time and gamble on that genetic algorithm?

Also, got some progress on my side project.

The cnn based auto encoder works better without the dense layer (or full connect layer), I don't know why, but it just works.

Here are some examples, the green pixel means the autoencoder thought there should be a living cell, and it is. Red pixel means the autoencoder think it's alive, but it's not. White means the autoencoder missed, it though it's dead, but it's alive.

Next step: try to build a neural network that can give necessary addition to make next step as close as possible to a given target pattern. (So I can display the bad apple in Conway's game of life). And to make things less boring, I decide to use genetic algorithm to "train" the network. Wish me luck 🤣

I have never owned a Wii or 3ds, but I do saw my classmate playing it.

I sincerely hope there can be a torrent file that accidentally contains those data, maybe along with some simulators, and f*ck those copyright limitations.

Donating those data to a history is a good choice. But in my (aggressive) opinion, it's useless unless the public can get their hands on those data and make use of them.

Anyway, this man is a real hero.
----
I bought EVERY Nintendo Wii U & 3DS game before the Nintendo eShop closes https://youtu.be/ujHUMG0Uovs

I saw ltt's review on GuliKit's joystick replacement kit for joycon, and recently my switch pro controller's left joystick is heavy drifted, which eventually turned itself into an E-waste.

So I bought their controller (they don't offer replacement kit for the pro controller). I still prefer the official controller because of the surface and vibration. Each time this controller vibrate, it just scared me and make me want to throw it away.

But I did spend sometime with it (not playing game), I found out it offers a way to update it's firmware! And it can record your input sequence and replay it, or export to and import from your PC. And you know what? I think I can generate script for painting the profile picture in Splatoon 3.