Print

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

[overseer05-15]

http://www.npr.org/sections/thetwo-way/2017/07/30/540374111/thieves-take-neil-armstrongs-solid-gold-lunar-module-model

This makes me very angry. Whoever stole it should be punched by Buzz Aldrin then locked up for a long time.

I live less than an hour away so I went to see the empty spot meself. Pissed cuz that thing should have been a national treasure or something.

[smjjames]

http://www.npr.org/sections/thetwo-way/2017/07/30/540374111/thieves-take-neil-armstrongs-solid-gold-lunar-module-model

This makes me very angry. Whoever stole it should be punched by Buzz Aldrin then locked up for a long time.

I live less than an hour away so I went to see the empty spot meself. Pissed cuz that thing should have been a national treasure or something.

That thing is going to be impossible to pawn off legally, and yes, it's a terrible thing that it happened.

[pisskop]

I would be chagrin to find out it was melted down instead of sold to a collector.

[Teneb]

http://www.npr.org/sections/thetwo-way/2017/07/30/540374111/thieves-take-neil-armstrongs-solid-gold-lunar-module-model

This makes me very angry. Whoever stole it should be punched by Buzz Aldrin then locked up for a long time.

I live less than an hour away so I went to see the empty spot meself. Pissed cuz that thing should have been a national treasure or something.

That thing is going to be impossible to pawn off legally, and yes, it's a terrible thing that it happened.

It's quite likely that, since it is solid gold, it will simply be melted down.

EDIT : Ninja'd

[pisskop]

Aye, ive ninja'd the thread like the thieves ninja'd the lunar module.

 ....

>.>
<.<

too soon?

[scourge728]

So, on the topic of machine learning, here's my own speculation on the future (although I don't know all that much about such things)

From what I've gathered, as it stands no A.I can think about, let alone act upon anything outside of it's programming, although it can do things that the humans that made/use/run/whatever it didn't intend for it to do if it is something that helps it achieve it's programmed goal, that A.I they taught to play tetris that paused the game instead of losing that wasn't even supposed to "know" about the pause function, and that A.I they had speedrunning super mario bros that accidentally discovered a new glitch that no one knew about and then decided "hey, this works better" and started using that both come to mind. But theoretically, couldn't a human-level A.I (as in one capable of actual thinking, contemplating and would be considered sentient) do things it's programming doesn't tell it to do, after all, humans do things nature never intended for us to do all the time, such as make A.I, go to the moon, build houses, drive cars, basically everything in modern life are things not written into genetic code/instinct and also most don't happen in any other animal, and the ones that DO happen in other animals are usually A. Somewhat instinctual (Those species of ants that build bridges and other things with their own bodies come to mind here, I'd guess that the one ant species that feeds mushrooms and harvests the energy they produce would fall here, although I don't know enough about either of them to say for sure,) B. Happening in the more intelligent animals (Such as the other apes, crows and octopi (which IIRC also have only a 24-hour memory span, and so have to relearn things every day (although I'm not sure about that fact either) I've also heard that they don't seem to have emotion) or C. Have to be taught by humans although whether they pass that information down to their young seems to vary.

[Avis-Mergulus]

Sure, a general AI could do these things. Its ability to discover problems on its own and solve them (and consequently set its own goals, since that's just a subset of problem definition - in a way your goals are defined by what you consider to be problems and solitions) would be how we'd know it's general. But we're not even close to that - and even if we decided to develop such a thing, that is not something that can be created by accident.

I'm not sure if saying that humans do things that nature never intended is correct on a philosophical level. Everything we do is a consequence of our natural problem-solving abilities, so everything we do is 'intended by Nature' in a way, right?

E: redundancy

[smjjames]

octopi (which IIRC also have only a 24-hour memory span, and so have to relearn things every day (although I'm not sure about that fact either) I've also heard that they don't seem to have emotion)

That is actually a pretty subjective question since we can't ask them how they feel and we're going off our own definitions of emotion.

[Tack]

I think the 'fear' comes from designing a self-improving AI, much like Deep Mind is doing.
Whilst it's a dumb, dumb robot, it's a really fast and convenient way to add on extra layers, but when it gets smarter, adding more programming gets more difficult, so it would be very easy for a lazy person to just 'let it run overnight'.

That's when core programming becomes a problem, because if you have a smart enough general AI which only knows it must self-improve, you run into trouble.

Again, this is the kind of stuff which is nowhere near happening anytime soon, but it's definitely smart to learn owl-taming before attempting to raise a wild owl.

[Harry Baldman]

I'm not sure if saying that humans do things that nature never intended is correct on a philosophical level. Everything we do is a consequence of our natural problem-solving abilities, so everything we do is 'intended by Nature' in a way, right?

Nature's design process, natural selection, strictly tests organisms on their lower bounds - can you reproduce in the environment, Y/N, all else is incidental unless applicable to reproduction. So human beings and other primates didn't really do all that well evolutionary-wise (endemic to small patches of East Africa for us in particular, for example) until fairly recently when they lucked out into a situation where tool use came in handy. As a consequence of that we haven't really found out most of the relevant upper bounds for the design of a human being, although we do know, say, the upper bound for conducting an action potential in a myelinated neuron and can guess at, say, the maximum age we could achieve under perfect conditions.

So although the myelinated neuron is an important adaptation in making us not die and survive to a point we can reproduce, the overall peak performance that a human being can achieve because of it would be unheard of in the actual wild of the day due to suboptimal conditions, so the design process responsible for it isn't sensitive enough to be described as precision engineering. And now we're largely out from under Nature's thumb due to technology getting seriously out of control. At least until we destroy our world, I guess, which is when we'll all go mostly or entirely extinct as Nature by itself won't be able to redesign us out of the corner our low genetic diversity puts us in.

This seems to be directly opposite to how we design our own technologies where we are a lot more interested in upper bounds of functioning (as well as able to meaningfully test them) and thus the intended functionality is usually the ideal functionality. So our intent is a lot more meaningful than what Nature intended, in a sense.

[Descan]

basically humans seek the few giant peaks they can see in a design space, evolutionary algorithms travel along the valleys (and sometimes find an even bigger peak we didn't know about, like a certain radio antennae design, or a radio transciever smaller than anything humans could make on the platform because the algorithm exploited faults in the manufactoring of the specific instance of the platform it uses to augment the electrical environment, using loops that look like they're useless but are actually key to the whole thing working right)

[Avis-Mergulus]

I 'd have to say that we aren't actually out from under Nature's thumb in any fundamental way - we still solve the same resource problems as all other life using our evolved skills and capacities. I guess under a definition where tool use != Nature or human != Nature or something of the sort, you could say that, but that would be a naive and inconsistent definition of Nature.
"Doing that well evolutionary-wise" is also not a meaningful scale, in that every species not extinct or in immediate danger of extinction can be said to be doing equally well from an evolutionary standpoint. I don't want to be that person who nitpicks at everybody's definitions, but I really think it's necessary for this question since it's basically only a question of definition with no verifiable component whatsoever.
Having low peak performance "in the actual wild" also seems irrelevant from an evolutionary standpoint. The human survival strategy is not to perform in anything one can consider an actual wild - I assume that that means an environment that isn't controlled by other humans. It's a strategy humans excel at. It's also not exclusive to humans.
And our ability to design things and test them on the upper bound of performance is also the consequence of our natural analytical and information-sharing ability.
"Lucking out" on an evolutionary process (whether that may be true or not) is also irrelevant, sincee evolution is not an actual fair contest or necessarily a contest at all. What remains, remains.
Dying out by overconsumption of the environment is also in no way surprising for a population. I mean, put some mold in a Petri dish with nutrient and wait. In a while, it will consume the nutrient and die. We might or might not have the advantage of analytical foresight, but the situation iz functionally identical.

Basically, what I mean to say is that human behavior doesn't actually give any actual reasons to put us in a qualitatively different category, any more than any other species that isn't extinct yet. Of course, we're humans and to us, we're meaningfully different, but to evolutionary processes we're the same, since they apply to everything that reproduces and dies (in the narrow sense) or straight up to everything that changes over time. Tool use or intelligence or anything else doesn't change that.

If I sound like an overly argumentative ass again, sorry. I just mean to clarify my definitions, I guess, since the initial argument included one that isn't ever used in a rigorous sense - "Nature" with-a-capital-N - but is more of a trivial way to say "not human" or "all life", depending on the context. I also wanted to clarify that I didn't mean anything about the human design process being similar or dissimilar to an evolutionary optimization algorithm - this isn't my point. My point is it that the capability to undertake that process is the result of an evolutionary algorithm, like everything else.

[Harry Baldman]

Oh, don't worry, I'm just doing my usual late night bullshit here. I like musing about evolution in general.

But what I meant about us not being under Nature's thumb is that there's very little to no selection pressure going on when we've essentially become able to dictate a lot of our environment to suit our deficiencies (a lot like a bacterium seeded onto a sterile Petri dish without relevant antibiotics doesn't particularly need to outbreed much of anything due to our efforts). So Nature (in this case meaning natural selection) isn't designing us at the present time. Instead it's the inverse where we meddle in nature to optimize our own growth and predictably fuck up to an extreme degree much like how a lot of Nature's own hard work is mostly undone when an extinction event rolls around and only the smallest slice of that intriguing old diversity persists.

[Avis-Mergulus]

Oh, don't worry, I'm just doing my usual late night bullshit here. I like musing about evolution in general.

But what I meant about us not being under Nature's thumb is that there's very little to no selection pressure going on when we've essentially become able to dictate a lot of our environment to suit our deficiencies (a lot like a bacterium seeded onto a sterile Petri dish without relevant antibiotics doesn't particularly need to outbreed much of anything due to our efforts). So Nature (in this case meaning natural selection) isn't designing us at the present time. Instead it's the inverse where we meddle in nature to optimize our own growth and predictably fuck up to an extreme degree much like how a lot of Nature's own hard work is mostly undone when an extinction event rolls around and only the smallest slice of that intriguing old diversity persists.

We still die, though, and often before or without reproducing. This is like that bacterium in the Petri dish: if you leave it to itself, it starts putting selection pressure onto itself as it multiplies. Presumably, a form of life that doesn't do that can evolve, and maybe we are that form of life, but it doesn't seem that way so far. And if we carry on the way we're doing, the pressure will only mount. Humans aren't the only species to undergo a period of relatively low pressure (though it's still non-zero), overpopulate its resource base, and then come under high pressure again. If we exclude every species that has gone through that process from our definition of Nature, people will look at us funny. Even if that's consistent in its own way.

I guess what I mean is that we can't say we've mastered our environment until that mastery is sustainable. Otherwise we will have to say that every species with a temporary advantage before us has done so, and that just doesn't have the inherent badassery that the word "mastery" implies.

[redwallzyl]

It is the want life to fill niches and that avoid conflict in a stable environment. this is when change stops moving at a fast pace. however the equilibrium is constantly disrupted and so this is never eternal. we will see how long we can sustain ours.