Updated the gig as there is a ready-made bot kit.

182 rules and 104 synonyms. Always there for you.

And yes, before you ask, we can do "boyfriend bots" too :-D

(Not the same as a real one, but what's wrong with some fun and games?)

theory validity {
  // a human weighs between 80 and 400 pounds
  $x is a human
    & $x weighs ($y pounds)
    => $y >= 80 & $y <= 400
   
  contradiction => say (That can't be right)
}

Yes, this works already. Full program and some responses.

So yes, the speedrun definitely continues—just without videos right now.

Creating a very nice "girlfriend" bot with a friend too.

I code until it's done.

Bot receives facts:

RAM has a size
I have RAM

And asks:

What is the size of my RAM?

To achieve this, we invent and implement a whole new language which mixes English and Python. Complete program: 299 lines.

2:45:07 mega-stream.

It is a good idea after all.

Our new bot plays MKLab FM, a very nice (and free) electronic dance music radio, in your Discord server.

1000 € are a good number and will help this project a lot :)

It can either be a gift, an investment or a bot order, depending on your wishes...

Let's just talk about it.

I introduce a new concept to JavaX—coding on the meta-level. Just write meta {} and put your code-making code inside the brackets.

OK the concept is not completely new, but it does work exceptionally well in JavaX.

In fact, we are creating and using a custom language directly within this example. Oh, and there is now a shortened version with only 33 lines.

Spends the time, pays the bills.

Another functioning JavaX keyword invented in a few minutes.

class Me {
  selfType hello() { print("hello!"); this; }
}

new Me().hello().hello();

selfType just refers to the enclosing class. Avoid repetition. It really shines in combination with includes.

Let's say a program is something that describes itself, tests itself and improves itself.

I have a special reconnaissance operation that can be performed in less than 2 nanoseconds (sic, see benchmark to the right) on any image stored in a special smart format. Combining a few of these operations allows us to find things in that image very quickly.

So in total, I expect a time of maybe, say, a millisecond?—for recognizing a typical image's major features. We still need to invent some smart algorithms for this to happen, but this is the goal.

And what that cornerstone operation is? Welll... stay tuned is all I can say.

...I extend it. (Have I said that before? Deja vu!)

Let's say I am annoyed by the inability to nest a function inside a function in Java. So I invent the "embedded" keyword.

sS bla() {
  S x = "hello";
  embedded S calc() { ret x + " " + x; }
  ret calc() + " " + calc();
}

In this useless but instructive example, I embed the function calc in the function bla with access to the local variable x.

How is it translated? Using a function of a mere 21 lines and an anonymous inner class. (Nota bene: It's currently more or less limited to a single embedded function per "mother".)

Example with translation to Java.

This is such a wonder. A compiler written in Java compiling a slightly weird, but interesting Logo-inspired special purpose language into actual DOOM levels. Our AI could produce DOOM levels for us from just a simple description.

Oh, ah... link.

But a good one.

Question-dot.

c ?. setDoubleBuffered(b);

=>

if (c != null) c.setDoubleBuffered(c);