Weekly Science Headlines

[The Good Reverend Roger]

Much f'kin closer If the VIV ecosystem matures the way I think it could, I could see us getting there pretty rapid. Think price-comparison websites but applied to AI comparison AI's, nested meta-agents selecting ranked domain expertise. The utility is guaranteed off the charts and, sooner or later, some smart arse is going to come up with a good enough personality model to wrap it up in.

Remember, this is all based on my theory that it takes significantly less than human-level intellect to fool a human into thinking it is. (like a fraction of a fraction of 1%) If I'm wrong in this then I'll revise my guess from imminent to - shortly before the heat death of the universe. 

A while ago, I suggested making an AI that isn't trying to emulate a human being.  Fuck the Turing test.  If that's what we need, we can just squat out more humans.

Plus, it would scare the poop out of everyone with a functional arachnid response.

[Junkenstein]

NASA discovers >1000 more planets

Yeah, I've been talking about that.  Or at least what lead up to it.

More related to the election thread, but I'm guessing neither are talking much about this or NASA in general? Unless I missed trump saying something about a great big dome too.

[Vanadium Gryllz]

NASA discovers >1000 more planets

Yeah, I've been talking about that.  Or at least what lead up to it.

Yeah interesting stuff. Are they gonna continue looking or use the telescopes for something else now? What do they actually do where you work?

[P3nT4gR4m]

Much f'kin closer If the VIV ecosystem matures the way I think it could, I could see us getting there pretty rapid. Think price-comparison websites but applied to AI comparison AI's, nested meta-agents selecting ranked domain expertise. The utility is guaranteed off the charts and, sooner or later, some smart arse is going to come up with a good enough personality model to wrap it up in.

Remember, this is all based on my theory that it takes significantly less than human-level intellect to fool a human into thinking it is. (like a fraction of a fraction of 1%) If I'm wrong in this then I'll revise my guess from imminent to - shortly before the heat death of the universe. 

A while ago, I suggested making an AI that isn't trying to emulate a human being.  Fuck the Turing test.  If that's what we need, we can just squat out more humans.

Plus, it would scare the poop out of everyone with a functional arachnid response.

Totally. There's a ton of voices in the AI community who think "Human-level AGI" will be some kind of worthwhile and notable milestone. I can only attribute this obsession to hubris, all the while yelling "Why do we want to produce a simulated retard?" to anyone who'll listen.

The main lesson from all the progress in narrow AI (or, as I've taken to call it - "Actual Intelligence") is that, for all we have a seriously heavy-duty neural net, human brains are pretty much the worst fucking tool for any problem you want to solve. Turns out all you get when you scale up to 100-billion neurons is a system that can't concentrate on even the simplest task for longer than a couple of nanoseconds. Scale it back to a couple of million, however and it can actually drive a car without killing everyone in the vehicle cos it wanted to check Facebook.

[The Good Reverend Roger]

NASA discovers >1000 more planets

Yeah, I've been talking about that.  Or at least what lead up to it.

Yeah interesting stuff. Are they gonna continue looking or use the telescopes for something else now? What do they actually do where you work?

I work at the Kitt Peak National Observatory.  Most of what we are doing these days is related to exoplanets.  I was thinking of writing a bit of a summary for geeks.

We also look for gigantic rocks that are aimed at us in the SARA dome, but sadly there don't seem to be any.

[The Good Reverend Roger]

Much f'kin closer If the VIV ecosystem matures the way I think it could, I could see us getting there pretty rapid. Think price-comparison websites but applied to AI comparison AI's, nested meta-agents selecting ranked domain expertise. The utility is guaranteed off the charts and, sooner or later, some smart arse is going to come up with a good enough personality model to wrap it up in.

Remember, this is all based on my theory that it takes significantly less than human-level intellect to fool a human into thinking it is. (like a fraction of a fraction of 1%) If I'm wrong in this then I'll revise my guess from imminent to - shortly before the heat death of the universe. 

A while ago, I suggested making an AI that isn't trying to emulate a human being.  Fuck the Turing test.  If that's what we need, we can just squat out more humans.

Plus, it would scare the poop out of everyone with a functional arachnid response.

Totally. There's a ton of voices in the AI community who think "Human-level AGI" will be some kind of worthwhile and notable milestone. I can only attribute this obsession to hubris, all the while yelling "Why do we want to produce a simulated retard?" to anyone who'll listen.

The main lesson from all the progress in narrow AI (or, as I've taken to call it - "Actual Intelligence") is that, for all we have a seriously heavy-duty neural net, human brains are pretty much the worst fucking tool for any problem you want to solve. Turns out all you get when you scale up to 100-billion neurons is a system that can't concentrate on even the simplest task for longer than a couple of nanoseconds. Scale it back to a couple of million, however and it can actually drive a car without killing everyone in the vehicle cos it wanted to check Facebook.

Thing is, even if it's aware, getting it to do shit for us wouldn't be a huge deal.  It would be like having someone come by once a century and asking you to mow the lawn, which has only a week's worth of growth.

[Vanadium Gryllz]

We also look for gigantic rocks that are aimed at us in the SARA dome, but sadly there don't seem to be any.

Sadly because you don't therefore get to blow them up with high powered space lazers?

[LMNO]

We also look for gigantic rocks that are aimed at us in the SARA dome, but sadly there don't seem to be any.

Sadly because you don't therefore get to blow them up with high powered space lazers?

Sad because he's got a Honda full of silver, and nothing to do with it.

[The Good Reverend Roger]

We also look for gigantic rocks that are aimed at us in the SARA dome, but sadly there don't seem to be any.

Sadly because you don't therefore get to blow them up with high powered space lazers?

Sadly because in 3 years, mostly everyone will still be here.

[The Good Reverend Roger]

We also look for gigantic rocks that are aimed at us in the SARA dome, but sadly there don't seem to be any.

Sadly because you don't therefore get to blow them up with high powered space lazers?

Sad because he's got a Honda full of silver, and nothing to do with it.

  Also this.

[Junkenstein]

This may be worth a look:
http://www.bbc.co.uk/news/world-middle-east-36498234

A huge monument has been discovered buried under the sands at the Petra World Heritage site in southern Jordan.
Archaeologists used satellite images, drone photography and ground surveys to locate the find, according to the study published in the American Schools of Oriental Research.
The large platform is about as long as an Olympic swimming pool and twice as wide.

There's the history side of making/moving monuments of this scale which is always pretty fucking impressive. The other side is the location. If you're not familiar with Petra, it's worth a google. Things weren't done on a small scale so it's quite likely to lead to more too.

[Brother Mythos]

'Scientists observe first signs of healing in the Antarctic ozone layer'

Here are two quotes from the article:

'Scientists at MIT and elsewhere have identified the "first fingerprints of healing" of the Antarctic ozone layer, published today in the journal Science.'

'"What's exciting for me personally is, this brings so much of my own work over 30 years full circle," says Solomon, whose research into chlorine and ozone spurred the Montreal Protocol. "Science was helpful in showing the path, diplomats and countries and industry were incredibly able in charting a pathway out of these molecules, and now we've actually seen the planet starting to get better. It's a wonderful thing."'

Here's the link:   http://phys.org/news/2016-06-scientists-antarctic-ozone-layer.html

I tried linking to the original 'Science' article, but full articles on that site are only available to paid subscribers.

So, for a change, there's a little good news concerning Planet Earth.

[Mesozoic Mister Nigel]

Nice! Now we just need to make some quicker progress on reducing carbon emissions before the icecaps are gone.

[P3nT4gR4m]

If this shit isn't exactly what it sounds like I will kill a mofo 

Cos it sounds like we can now build cyborg fish. You ever hear something that's so fucking cool, you plain don't have the heart to fact check it?

[Mesozoic Mister Nigel]

If this shit isn't exactly what it sounds like I will kill a mofo 

Cos it sounds like we can now build cyborg fish. You ever hear something that's so fucking cool, you plain don't have the heart to fact check it?

We covered a teeny bit of biomimetic engineering in the muscle and movement section of neurophysiology, and artificial swimming meat fish are totally a real thing.

Found the paper, it's 100% legit: http://scholar.google.com/scholar_url?url=http://dspace.mit.edu/openaccess-disseminate/1721.1/79685&hl=en&sa=X&scisig=AAGBfm3AhsgQSFlgCphGBJBghvurTo4M3A&nossl=1&oi=scholarr