[goaty]

 
.
Hows your creecha coming along?    Have you got your actuator down pat your going to use for your legs?

[HS]

Dude! I thought summer was going to include more free time.  hehe
nope...
So not much progress. I got a nice solar panel for the top of it.  Letss see... what else.. I got a nice box with compartments for batteries, motors, and neurons, got some motors, but I hope my neural net can figure out how to run brushless, (cause I'd have to buy a bigger box for all the paraphernalia required for three phase.) At this rate it's gonna end up looking (and probably acting)  like the luggage from Discworld...




Following that train of thought on testing AI, you could know the strength of an AI by the constraints on the tests which provide useful feedback from said AI. But you’d have to compare different tests, the more the better. They will end up specialized to various types of AI that the people have to deal with, in order to extract the best data, not one size fits all. So, I guess we should sit back, let everyone try to figure it out, and then develop a test for testing tests! Then we’ll have a true comparison on a single metric.

[goaty]

Love the picture, there were plenty of little monsters in those novels, fun-fun.

Ill send a picture back over->


Did u do a bit of carpentry for this "compartmentalized" box?

Ive got my system to 4 main compartments now,  power, actuators, eye and brain.  =)

Run a BIG solar panel,  if it can lift it, then itll be a stronger robot the bigger its panel is,  otherwise your playing catch-ups, with a sleeping cycle, and it ends up asleep more than its awake more often than not.

Running brushless - maybe if you train a net on where it thinks its eye is supposed to be, after so much electricity goes into its motors.   If you managed that its more adaptive than absolute positioning,  but still I like absolute positioning, I think im going to have it in mine, all my prior thought went into me having it, but it would be weird if its simple to adapt around it with a.i. instead, and its just as good.  (and more adaptive to error in the system.)

And keeping things simple, I KNOW It only takes a few metrics to create a very lively and varied activity.   

.

[HS]

Lolz that diagram! It's like someone trying to calm down Lovecraft.

Did u do a bit of carpentry for this "compartmentalized" box?

I rummaged around in some old boxes and found the perfect thing. Don’t think my attempt at carpentry would produce anything… liftable. I’ll just attach a thin piece of plywood to the base an it should be strong enough.

Run a BIG solar panel,  if it can lift it, then itll be a stronger robot the bigger its panel is,  otherwise your playing catch-ups, with a sleeping cycle, and it ends up asleep more than its awake more often than not.

The solar panel is thin flexible about the size of a medium laptop monitor. 12 volts. Perfect to charge my 9-volt lithium batteries (7 in parallel). Don’t see a way around a sleep cycle though, especially at night. I mean, it could hang out under streetlights, theoretically… But the police might try to break into it to find out what it was selling.  →

Running brushless - maybe if you train a net on where it thinks its eye is supposed to be, after so much electricity goes into its motors.   If you managed that its more adaptive than absolute positioning,  but still I like absolute positioning, I think im going to have it in mine, all my prior thought went into me having it, but it would be weird if its simple to adapt around it with a.i. instead, and its just as good.  (and more adaptive to error in the system.)

Yeah, I like the brushless motor idea; it would be a useful and simple starter skill to work on with the neural net.

[goaty]

Selling items,   a portable automatic middle man,  sounds like a nifty idea.
If you could make it purchase supply itself, then you could just sit back and watch the money roll in without getting out of your arm chair, but ppl would have to want to deal with it for your idea to work...  and some ppl don't like robos…

[patchon]

My bots im not planning on involving them with people, because I think it can actually get a bit negative, bots pretending to have a personality doesn't cut it for me, so I guess others might be similar...  they are just going to do jobs for me in secret,  like hunt for rabbits, go fish like bears for me, clean out my toilet and house, maybe they can even prepare the meat, and definitely tend to a garden and maybe some kind of revenue, but I probably wont need money, depends on how good I get it.     

[yotamarker]

cope, they need to surpass humans, so they must have human skills.

[goaty]

Surpassing humans involves what?  Using human skills? No.  It has to have something that humans HAVENT GOT.

Its hard to think about what is the metric of success,  so we don't even know WHAT EVEN IS MORE SUCCESSFUL to EVEN MAKE IT DO IT!!!

Inventing more technology is not it...  we are good enough at that ourselves at making cheesy crap.

If there is something that surpasses us, I cant think of it,  and just because it can beat you in a fight and take over the worlds armies and take the world hostage DOES NOT MEAN IT SURPASSES US. 

.

[Korrelan]

It has to have something that humans HAVENT GOT.

Or more of it… take a moment to consider all the different species and brain structures in the animal kingdom, that we know to exhibit our definition of intelligence.

Intelligence/ sentience/ consciousness are derived from a scalable law of nature… humans have the most… but by no means the maximum.

 

[HS]

Intelligence/ sentience/ consciousness are derived from a scalable law of nature.

But even simple things get emergent properties as you zoom out. I think intelligence is more like a series of thresholds, like wave forms on a string. With enough activation energy you can snap into another level, and get a new reaction going.