Suuper's Sewing and Crafts

[Suuper-san]

Ahaha just YouTube videos!
I love working from scratch so working from nature is an automatic favourite, like making huts in the middle of the forest kind of thing (not done yet but would love to try :P)
Mostly I Google tutorials for everything I do but if I see there is a large variation with different tutorials then I just take all the information and make up my own mashed version XD

I like trying out simple craft projects, especially stuff that doesn't require anything specialist, so just paper, card and glue that sort of thing.

[Suuper-san]

Finally, finally, finally.
I made paper!!!!!
A6 was the size my method made.



Used pretty much the same method that's everywhere on Youtube, blended some scrap paper in water, tipped it into a bowl, and then used a deckle (a wannabe flat sieve) to scoop a layer out and then tip it on to felt to dry out. I used an oven at very low temperature because air drying can take days, so in 2 hours start to finish I had paper.

I actually made 2 batches and improved on the second batch hugely.
I don't really have a use for it but it's good to know that it can be done now.

[legomaestro]

Of course you made paper suuper hahahaha. Man it's always cool to see what you're up to. My pa used to make firewood out of pulp paper so I know about that in general, but never seen them folded out neatly like that. Are they good to write on? I definitely have enough scrap to work with.

Man you really can make paper out of anything.

[Suuper-san]

well it helps using old paper as it's already processed fibres just waiting to be reused so in that sense it's quite easy. I'd love to make paper from scratch i.e. cutting down a tree XD

I was able to write on the paper fairly well, but overall the texture was closer to a tough sort of tissue rather than paper. The problem was that the paper was not getting compacted enough, and I did a second run today using a rolling pin to squash it and the result was a very good quality paper that I could write on with no problem.
The only large issue I have left is drying the paper, the oven works well but the paper distorts and doesn't remain flat. So it needs to be held flat while drying.
Also a bulk method would be nice to make more than just a few sheets an hour.

Here's two versions of paper with writing tests. Top is older tissue feel, bottom is much more paper feel. You can clearly see the texture difference too.


ahhh the logs from paper thing! I've been wanting to try that as well but I need to make a little press mold to make them. Hopefully before winter hits I'll manage it otherwise gotta wait until next winter to burn them haha

I'm after making a loom as well in other crafty business :P

[legomaestro]

xD And ah no i used the term wrong my parents just rolled them into balls and used it for fire, so no fancy log-shaped press.

And damn I thought the second one was a comparision with a regular sheet of A4 paper haha.

[Suuper-san]

haha awesome glad you thought the quality looked good :P
ahhh cool paper balls
so they made them from paper pulp and then had to dry them out?
I've actually still got one drying on my shelf from the first batch of paper pulp I made lol

[Suuper-san]

I'm also now using this thread for my programming related things, since that's a form of craft, from a certain point of view.

Basically I've finally managed to program my first neural net AI. If you dont know what that is, it's basically a form of artificial intelligence, a program that can learn information without the data having to be put piece by piece into the system directly. In it's simplest form, it's basically a statistics engine, something that can take a lot of data and work out the relationship between them all, and make predictions when it's given new data. Such as object recognition which many people use with apps such as Google Lens.

So I have made a simple object recognition AI, one that can recognize handwritten digits (0-9). You give it a bunch of data to learn, and then test it on a separate bunch of new data and see if it has learnt how to recognize the digits.

In my case I got 82% accuracy recognition, so 1 in 5 digits it guessed wrong. But that's pretty cool for data that looks like this:
I actually used my own hand written data because that was easier to work with than downloading other datasets which are available online XD


All it took was a scary bit of math:


And that's it!
Who knows what AI will achieve in the future????
It's amazing to understand just the surface of how they work, and to have produced my own tiny AI from scratch has been a great learning experience.

[Coryn]

Guh, the biggest hurdle between knowing how to program and not learning in the first place has got to be the math, lol

Any bigger plans for this thing? Or just a side interest?

[Suuper-san]

To be honest, a LOT of coding doesn't involve more than just simple addition, subtraction, multiplication and division etc. AI is a supremely different beast in that regard. You also have to understand the logic behind which the system works, which is beyond a lot of people because of so many key-words that you have to know just to stand at the entrance-way of understanding.

But it's totally just a side interest, simply because it would such up too much time and not go anywhere if I did invest more time into it, as much as I would like to. I'm just a bit of a programming geek in that regard.

However, if I were to make a game, which I have always promised on and off, then a true responsive learning AI would make for an impressive addition to the system, assuming I could get it to work within the game's constraints.
Such as a sparring practice partner that always increases their skill as you improve your level, as a possible option. Or just enemy AI with impossible to read attack patterns.

[Coryn]

I guess it just depends on what you're going for. Impossible to read enemies don't necessarily make for an enjoyable game, just a hard one.

[Suuper-san]

I meant impossible to read as in no predefined or obvious attack patterns, but you're totally right if the enemy is too hard to beat then it's not a fun game. has to be just the right amount of challenging to be rewarding.
Game balance is its own category of problems lol
I've watched quite a few Youtube videos about different famous games and how they struggled to get the right balance with certain weapons or levels. It's quite informative.
Although I suspect any games I make won't be complex enough to need any sort of real balancing tbh. They're more likely to be tetris clones than anything else :P