Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[shadenight123]

I did some more power optimizing and managed to halve the need Uristwatts again. I built the walkways and gears for something around a 23k Uristwatt windfarm a while ago, and it's looking like the power requirement for the machinery at rest (minus the gears and axles supporting the windmills) is only going to be around 1500. 1500! And probably about 2000 when you're using it. Madness.

So i'm almost done linking. I may even finish and be able to get it up and running sometime monday or tuesday. Before I upload it however i'll need to clean some things up and make it fit for public usage. Build self refilling water channels and things like that.

wait, so it's also ECO COMPATIBLE!?
oh sheesh.

[Mego]

An eco-friendly AI that will eventually take over the world.

Al Gore would be pleased.

[tahujdt]

If you read "I, Robot", the A.I.s still follow the Three Laws.

[MagmaMcFry]

The Three Laws are poorly worded, they use high-level abstract moral-dependent conflicting words such as "harm", "obey" and "protect" (thereby enforcing very high A.I.Q. standards to even have robots "understand" them) and are therefore susceptible to craploads of moral dilemmas. A robot can very easily be put into a situation in which it cannot uphold the first law, and after that the robot's behaviour is undefined by the laws, so it is free to do whatever it wants after it is forced to break a law. Also, why do robots have to follow these laws? What happens to them if they break a law, and who can enforce the punishment?

We need better Laws.

[shadenight123]

The Three Laws are poorly worded, they use high-level abstract moral-dependent conflicting words such as "harm", "obey" and "protect" (thereby enforcing very high A.I.Q. standards to even have robots "understand" them) and are therefore susceptible to craploads of moral dilemmas. A robot can very easily be put into a situation in which it cannot uphold the first law, and after that the robot's behaviour is undefined by the laws, so it is free to do whatever it wants after it is forced to break a law. Also, why do robots have to follow these laws? What happens to them if they break a law, and who can enforce the punishment?

We need better Laws.

indeed.
and the new laws will be:
No laws at all!!!
let them act without laws.
thus they'll be jobless and eventually turn themselves off.

[Kaleb702 Games]

Also, as a side effect of the laws, each and every bit of harm done by anyone is the AI's fault. It could've done an action to stop it. Therefore, law 1 is always broken.
Better laws would be:
1) An AI must stop humans from dying to the best of its capabilities.
2) An AI must do what a human says for it to do to the best of its abilities, unless this would violate law 1.
3) An AI must stop itself from dying to the best of its abilities, unless this would violate law 1 or 2.

There's still loopholes, but these are still better, because the AI won't always be breaking law 1 / 2 by being given an impossible order / not stopping a murderer on the other side of the planet.

[tahujdt]

The way Asimov's robots are programmed, a robot would break down before it could break a law.

[Necro910]

Also, as a side effect of the laws, each and every bit of harm done by anyone is the AI's fault. It could've done an action to stop it. Therefore, law 1 is always broken.
Better laws would be:
1) An AI must stop humans from dying to the best of its capabilities.
2) An AI must do what a human says for it to do to the best of its abilities, unless this would violate law 1.
3) An AI must stop itself from dying to the best of its abilities, unless this would violate law 1 or 2.

There's still loopholes, but these are still better, because the AI won't always be breaking law 1 / 2 by being given an impossible order / not stopping a murderer on the other side of the planet.

"I, Robot" would still happen.

[MagmaMcFry]

My opinion on the rules: Rule 1 is impractical, Rule 2 is obvious, and Rule 3 is unnecessary.

[Necro910]

My opinion on the rules: Rule 1 is impractical, Rule 2 is obvious, and Rule 3 is unnecessary.

Rule 1 both prevents and mandates human death, rule 2 can lead to some problems but not as much, rule 3 prevents robots from shooting themselves.

Also, a robot can say "I am not a robot, I am an artificial intelligence", leading to him becoming unbound 

[BloodBeard]

My plan backfired... staring at all the microcline gears is giving me a headache. The good news however is i'm done linking!

Spoiler (click to show/hide)

*fireworks*

The base power requirement minus windmill stuff stands at 1524. When it's running things are turning on and off all over so i'll have to wait and see its peak power consumption when I put it through its paces.

As far as all the other numbers that'll have to wait untill I eliminate everything not being used. It's at this point that the project is going to split into two, one will continue on to program the AI and the other i'm going to customize to focus on text displaying which I will upload. Right now i'm just going to try to get it working and see if everything is properly linked.

[shadenight123]

if you stare long enough, it's like entering meth/drugs.
i think.
...
it's gotten smarter already, it's using COLOURS as a mean to distract it's enemies!

[Talvieno]

The way Asimov's robots are programmed, a robot would break down before it could break a law.

Imagine the headaches this would give BloodBeard (even just during debugging) if his DFAI followed this... 

[BloodBeard]

*drumroll*

Success. A windfarm was built, the channels filled with water and I was able to type this with the lever keyboard:

Spoiler (click to show/hide)

 

Backspace works, clear works, no errors encountered. It took a while to type out that sentence, I was inspecting everything at the same time. For the version i'll be uploading I think i'll have a type of indicator by the keyboard that lets you know when you can type something because you need to wait untill certain gears are turned off between lever pulls

I typed random characters to fill the whole 32 tile display to make sure they all worked and pulled the clear lever. It was beautiful... Slow, but beautiful. My computer isn't the greatest, I only get around 20 FPS and clearing it all took about 5 minutes (fully automated).

So happy to see some results after all this time... and without errors. I'm still going to go through it a bunch more times but it looks good. It'll be a while before i'm ready to upload it, if you've been following this thread since the beginning i'm sure you have the patience to wait a little longer. 

[DS]

Glorious.

A note on speed - if you make an active record of something at an extremely slow FPS and load it up to the DFFD, it will appear to run much faster. Such a recording would show the clear lever functioning at a much, much faster speed.