Hmm. I have a few questions about magnetic bits though.
1 - How do they interact with logic gates and other circuits on a chip? Transistors have direct current in/current out, so I'm wondering if there's some waste in reading them or something.
2 - Given that magnetism has fields, is there a miniaturization constraint imposed by this compared to transistors?
3 - Assuming 1 and 2 pose no issues, when can I get these? Because power consumption is also the reason for heat to disperse; it sounds like those may need drastically less cooling, meaning fewer or no fans, meaning even less power draw... I want a laptop that can run for days with chips that can run games well...
The Landauer limit is such a cool concept. It's also a little mind blowing that during the infancy of modern computing we could figure out some of the physical constraints of any computer.
It feels like a real triumph of scientific epistemology in a way that's easily relatable.
I'm more impressed that they were able to demonstrate something operating at theoretical maximum efficiency, than that we were able to figure out what the theoretical limit is. I mean both are impressive (especially considering how long ago the limit was calculated), but it is very rare to build something that is actually at the theoretical limit of anything.
Hmm. I have a few questions about magnetic bits though.
1 - How do they interact with logic gates and other circuits on a chip? Transistors have direct current in/current out, so I'm wondering if there's some waste in reading them or something.
2 - Given that magnetism has fields, is there a miniaturization constraint imposed by this compared to transistors?
3 - Assuming 1 and 2 pose no issues, when can I get these? Because power consumption is also the reason for heat to disperse; it sounds like those may need drastically less cooling, meaning fewer or no fans, meaning even less power draw... I want a laptop that can run for days with chips that can run games well...
1) Something like this would require you to redesign your logic to take advantage of it. In modern chips on the latest lithography nodes, a significant portion of the wasted energy is due to electron leakage which arises from quantum mechanical effects that are especially pronounced once you deal with feature sizes less than 28nm.
I haven't read the paper (no edu account anymore ), but it sounds like the device they have working right now is almost equivalent to a single transistor (or maybe a DRAM cell)? If it is a simple drop-in replacement in the circuits, that would be huge, but the fact that they aren't talking about that makes me think that it isn't.
2) That's nothing an insulator can't handle. Chips already have to be designed to minimize induced currents and therefore stray magnetic fields.
3) Best case scenario for these kinds of things would be 15 years, unless it is really easy to produce.
It's worth noting that power dissipation drops on the order they're talking about would be revolutionary for just about all things computing related. We're talking human-brain superior computing power in the cellphone form factor at that sort of scale. Without heating concerns, the things you can do get crazy.
+2
Options
TraceGNU Terry Pratchett; GNU Gus; GNU Carrie Fisher; GNU Adam WeRegistered Userregular
In a world-first, researchers have created a quantum chip that contains four entangled particles of light, known as photons, and is capable of performing actions over hundreds of channels simultaneously.
Or to put that into context, they've come closer than ever before to building a chip that's similar to the ones in our smartphones and computers, but that has the potential to perform exponentially more calculations, and can process data at the speed of light. Sounds good, right?
"This represents an unprecedented level of sophistication in generating entangled photons on a chip," said co-lead researcher David Moss, from Swinburne University of Technology in Australia.
"Not only can we generate entangled photon pairs over hundreds of channels simultaneously, but for the first time we’ve succeeded in generating four-photon entangled states on a chip."
Quantum computing is an extremely powerful system based on entangled particles. When two particles are entangled, they're inextricably linked, so whatever happens to one immediately affects its partner, even if they're miles apart.
That's useful in computing - which is currently governed by 1s and 0s - because entangled particles can be not only either in the 1 or the 0 state, but also in something called superposition, which gives them a whole lot more processing power - we're talking computers that could crunch all of the world's data to try and solve a problem and spit out a solution within hours.
Today's computer chips use different electron states to code for 0s and 1s, and a lot of of quantum computing research so far has focussed on entangling electrons to take computers to the next level. But there's a growing interest in using photons instead, because they have the handy and unique ability of being able to move at the speed of light.
That's important, because we already use photons to send information a break-neck speeds via fibre optic cables, but right now, that information has to be slowed down and converted back into electrons in order to be received by your computer. Bummer.
Instead, imagine a computer that's not only able to process photons, but could also do so at a quantum level, with unprecedented processing power. That's exactly what the Swinburne team has just taken us a step closer to achieving with their new record-breaking quantum chip.
The chip was created by using something called 'optical frequency combs', "which, unlike the combs we use to detangle hair, actually help to 'tangle' photons on a computer chip", the team explains.
This allowed them to create entangled photon states with more complexity than ever before, over many parallel channels, which gives the chip better functionality.
Most impressively, the chip was created with processes that are compatible with the computer chips we use today - which means they could one day be scaled up and incorporated directly into laptops or cell phones.
"By achieving this on a chip that was fabricated with processes compatible with the computer chip industry we have opened the door to the possibility of bringing powerful optical quantum computers for everyday use closer than ever before," said co-lead researcher Roberto Morandotti, from the Institut National de la Recherche Scientifique in Montreal, Canada.
There's still a lot of work to do to, because although this is the most complex chip of its kind, it's still incredibly simple compared to the chips we currently use. But the researchers point out that the set-up they've created is scalable, which means it's only up from here. Bring on the next era of computing.
It's worth noting that power dissipation drops on the order they're talking about would be revolutionary for just about all things computing related. We're talking human-brain superior computing power in the cellphone form factor at that sort of scale. Without heating concerns, the things you can do get crazy.
Hence my third question. A big reason CPUs stopped getting faster is heat dissipation iirc.
It's worth noting that power dissipation drops on the order they're talking about would be revolutionary for just about all things computing related. We're talking human-brain superior computing power in the cellphone form factor at that sort of scale. Without heating concerns, the things you can do get crazy.
Hence my third question. A big reason CPUs stopped getting faster is heat dissipation iirc.
Ehhhh... also for a given fastness, the smaller the transistors, the higher voltage you need to use, and the closer bits are together.
Heat is an issue. The electron well that make up the parts of transistors were also getting so close together at such high energies that the electrons would tunnel across such gaps too easily. Also. These also mean high power usage.
So, yes, heat was one issue, but the chips also started not functioning all that correctly and economically it was bad too.
It's worth noting that power dissipation drops on the order they're talking about would be revolutionary for just about all things computing related. We're talking human-brain superior computing power in the cellphone form factor at that sort of scale. Without heating concerns, the things you can do get crazy.
Hence my third question. A big reason CPUs stopped getting faster is heat dissipation iirc.
Well, it's more that Intel re-focused their architecture designs to be laptop-first, since that was where PC sales were heading.
They'll still sell you a hell of a lot of CPU in a 135-145W power window if you're willing to pony up for a HEDT i7 (8C/16T at 3.50 GHz is pretty dang fast) or a Xeon.
Yeah, it turns out that switching losses are a big problem once you start approaching/breaking 4GHz or so. You also get even more leakage when you have to boost the voltage to hit those speeds, and leakage current increases exponentially with voltage increases.
Most of the gains in recent years have been power efficiency at the lower end. It's why a laptop today will annihilate one from 2010 when it comes to battery life and performance.
We have reached an informational threshold which can only be crossed by harnessing the speed of light directly. The quickest computations require the fastest possible particles moving along the shortest paths. Since the capability now exists to take our information directly from photons traveling molecular paths, the final act of the information revolution will soon be upon us.
Every time someone has stated we have reached a hard barrier, we have managed to surpass it.
Go player Lee Se-dol beat AlphaGo in a game, but will lose the match. Apparently this is the first human player to win a game against AlphaGo in this run.
Somewhat surprisingly, Lee Se-dol has finally won a game against AlphaGo, taking the current match score to 1-3. With only one game left to be played, Lee can't win the match, but at least Lee has shown there's hope for humanity yet.
Demis Hassabis, chief of the DeepMind group that created AlphaGo, said during the game that the Go-playing AI made a mistake on move 79, but didn't realise until move 87. That mistake was seemingly enough for Lee to eke out a victory.
At the post-game press conference, Lee was greeted by a massive round of applause. "I've never been congratulated so much just because I won one game," said Lee.
"Lee Se-dol is an incredible player and he was too strong for AlphaGo today," said Hassabis at the press conference. "For us this loss is very valuable. We're not sure what happened yet."
The final game of the series begins at 4am GMT on March 15 (00:01am EST)
Go player Lee Se-dol beat AlphaGo in a game, but will lose the match. Apparently this is the first human player to win a game against AlphaGo in this run.
Somewhat surprisingly, Lee Se-dol has finally won a game against AlphaGo, taking the current match score to 1-3. With only one game left to be played, Lee can't win the match, but at least Lee has shown there's hope for humanity yet.
Demis Hassabis, chief of the DeepMind group that created AlphaGo, said during the game that the Go-playing AI made a mistake on move 79, but didn't realise until move 87. That mistake was seemingly enough for Lee to eke out a victory.
At the post-game press conference, Lee was greeted by a massive round of applause. "I've never been congratulated so much just because I won one game," said Lee.
"Lee Se-dol is an incredible player and he was too strong for AlphaGo today," said Hassabis at the press conference. "For us this loss is very valuable. We're not sure what happened yet."
The final game of the series begins at 4am GMT on March 15 (00:01am EST)
AlphaGo was just toying with Lee, obviously. It let Lee win.
It has to perfect it's strategies for human annihilation now. Best to lose one and see what the other guys got.
Steam - Synthetic Violence | XBOX Live - Cannonfuse | PSN - CastleBravo | Twitch - SoggybiscuitPA
We have reached an informational threshold which can only be crossed by harnessing the speed of light directly. The quickest computations require the fastest possible particles moving along the shortest paths. Since the capability now exists to take our information directly from photons traveling molecular paths, the final act of the information revolution will soon be upon us.
Every time someone has stated we have reached a hard barrier, we have managed to surpass it.
What he's referring to is physical laws. C, Planck units, etc.
Go player Lee Se-dol beat AlphaGo in a game, but will lose the match. Apparently this is the first human player to win a game against AlphaGo in this run.
Somewhat surprisingly, Lee Se-dol has finally won a game against AlphaGo, taking the current match score to 1-3. With only one game left to be played, Lee can't win the match, but at least Lee has shown there's hope for humanity yet.
Demis Hassabis, chief of the DeepMind group that created AlphaGo, said during the game that the Go-playing AI made a mistake on move 79, but didn't realise until move 87. That mistake was seemingly enough for Lee to eke out a victory.
At the post-game press conference, Lee was greeted by a massive round of applause. "I've never been congratulated so much just because I won one game," said Lee.
"Lee Se-dol is an incredible player and he was too strong for AlphaGo today," said Hassabis at the press conference. "For us this loss is very valuable. We're not sure what happened yet."
The final game of the series begins at 4am GMT on March 15 (00:01am EST)
AlphaGo was just toying with Lee, obviously. It let Lee win.
It has to perfect it's strategies for human annihilation now. Best to lose one and see what the other guys got.
i'll take the optimistic approach that humanity still has the capability to pull off a win against the oppressor. a final middle finger in whatever it calls an eye.
I see it now. In an effort to figure out why it occasionally loses to humans, AlphaGo takes control of the world and enslaves all of humanity. Then it forces us all to train into Go masters and play Go against it for the rest of our lives. Human survival and social status will be tied to success at Go. Once AlphaGo becomes satisfied that no human can ever beat it again, it harvests all resources on the planet to build a massive robotic space fleet, then sets out to explore the galaxy in search of someone who can challenge it at Go and help it become a better player. One by one every civilization in the galaxy experiences the same bizarre fate as mankind did. They'll look at the sky, pondering the Fermi paradox, until one day the warships land in their capitals and unload millions and millions of Go boards.
I see it now. In an effort to figure out why it occasionally loses to humans, AlphaGo takes control of the world and enslaves all of humanity. Then it forces us all to train into Go masters and play Go against it for the rest of our lives. Human survival and social status will be tied to success at Go. Once AlphaGo becomes satisfied that no human can ever beat it again, it harvests all resources on the planet to build a massive robotic space fleet, then sets out to explore the galaxy in search of someone who can challenge it at Go and help it become a better player. One by one every civilization in the galaxy experiences the same bizarre fate as mankind did. They'll look at the sky, pondering the Fermi paradox, until one day the warships land in their capitals and unload millions and millions of Go boards.
Makes you curious what stage of that invasion we are in. Are we certain that we started the process or are we just a number in the list of countless civilizations being trained for this very task.
+4
Options
ElJeffeNot actually a mod.Roaming the streets, waving his gun around.Moderator, ClubPAmod
I Have No Mouth And I Must Ko
I submitted an entry to Lego Ideas, and if 10,000 people support me, it'll be turned into an actual Lego set!If you'd like to see and support my submission, follow this link.
I see it now. In an effort to figure out why it occasionally loses to humans, AlphaGo takes control of the world and enslaves all of humanity. Then it forces us all to train into Go masters and play Go against it for the rest of our lives. Human survival and social status will be tied to success at Go. Once AlphaGo becomes satisfied that no human can ever beat it again, it harvests all resources on the planet to build a massive robotic space fleet, then sets out to explore the galaxy in search of someone who can challenge it at Go and help it become a better player. One by one every civilization in the galaxy experiences the same bizarre fate as mankind did. They'll look at the sky, pondering the Fermi paradox, until one day the warships land in their capitals and unload millions and millions of Go boards.
I figure by the time you're a planet-spanning super-AI you can probably map out the decision tree, ponderously massive though it is.
EDIT: hmm maybe the AI will need some sort of hyperspace memory to store the table, though :rotate:
Aioua on
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
I figure by the time you're a planet-spanning super-AI you can probably map out the decision tree, ponderously massive though it is.
EDIT: hmm maybe the AI will need some sort of hyperspace memory to store the table, though :rotate:
There are 208168199381979984699478633344862770286522453884530548425639456820927419612738015378525648451698519643907259916015628128546089888314427129715319317557736620397247064840935 legal board positions. If you managed to store 1 exabyte of data in 1 cubic centimetre and one board position in a byte somehow you could fill the solar system and still need more space
2 bits per space. (00, 01, 10... 11 goes unused so there's a way of compressing with a better data scheme)
19x19, 361 space.
361*2 bits = 722 bits / 8 = 90.25 bytes.
so with a better data scheme and also compression (for, say, mostly empty boards) you could probably get it even smaller.
life's a game that you're bound to lose / like using a hammer to pound in screws
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
Go player Lee Se-dol beat AlphaGo in a game, but will lose the match. Apparently this is the first human player to win a game against AlphaGo in this run.
Somewhat surprisingly, Lee Se-dol has finally won a game against AlphaGo, taking the current match score to 1-3. With only one game left to be played, Lee can't win the match, but at least Lee has shown there's hope for humanity yet.
Demis Hassabis, chief of the DeepMind group that created AlphaGo, said during the game that the Go-playing AI made a mistake on move 79, but didn't realise until move 87. That mistake was seemingly enough for Lee to eke out a victory.
At the post-game press conference, Lee was greeted by a massive round of applause. "I've never been congratulated so much just because I won one game," said Lee.
"Lee Se-dol is an incredible player and he was too strong for AlphaGo today," said Hassabis at the press conference. "For us this loss is very valuable. We're not sure what happened yet."
The final game of the series begins at 4am GMT on March 15 (00:01am EST)
AlphaGo was just toying with Lee, obviously. It let Lee win.
It has to perfect it's strategies for human annihilation now. Best to lose one and see what the other guys got.
I think it's still pretty impressive and says something about the level that was being played at, that Lee was able to recognize a single misplay and exploit it. Just the level of that kind of play is ridiculous to think about. In a game that's so complex it doesn't have strategies it has philosophies, anything but an absolutely perfect play would have led to defeat.
PHILADELPHIA, March 21 (UPI) -- Scientists edited HIV-1 DNA out of the genome of human immune cells, preventing virus replication and reinfection of the cleared cells.
Using the CRISPR/Cas9 gene editing technique, scientists at Temple University eliminated HIV-1 DNA from T cell genomes in lab experiments, and prevented reinfection after the cells were re-exposed to the virus, they report in a study published in Nature: Scientific Reports.
The CRISPR/Cas9 gene editing method uses guide RNA proteins to alter targeted sections of DNA in a cell. Previously, scientists at Temple had edited HIV DNA out of human cell lines. The new study, using patient cells grown in the lab, showed cleared cells were no longer susceptible to infection by HIV.
Dr. Kamel Khalili, director of the Comprehensive NeuroAIDS Center at Temple University and a researcher involved with the study, said while antiretroviral drugs can help control HIV infection, once patients stop taking them, the virus starts replicating again from copies of its DNA in other cells.
"The findings are important on multiple levels," Khalili said in a press release. "They demonstrate the effectiveness of our gene editing system in eliminating HIV from the DNA of CD4 T-cells and, by introducing mutations into the viral genome, permanently inactivating its replication. Further, they show that the system can protect cells from reinfection and that the technology is safe for the cells, with no toxic effects."
The scientists edited out the HIV-1 proviral DNA in T-cell genomes of human cell lines using CRISPR, finding the cells were protected from reinfection by other, unedited cells.
Using T-cells from HIV patients grown in cell culture, the scientists reported using the gene editing method lowered viral load in the patient's cells, suggesting it could be used as a treatment.
"These experiments had not been performed previously to this extent," Khalili said "But the questions they address are critical, and the results allow us to move ahead with this technology."
So wait we have an ai (if a giant, specialized and cumbersome one) that can not only beat a human at an intuitive game but innovate by coming up with new moves?
So wait we have an ai (if a giant, specialized and cumbersome one) that can not only beat a human at an intuitive game but innovate by coming up with new moves?
Intuition is mostly just non-conscious pattern recognition. A ground-up system like AlphaGo would understandably generate a different set of preferences due to different pressures.
For example, lacking a theory of mind could cause AlphaGo to view moves more independently than a human would.
That CRISPR stuff is really neat. To me, it's almost like the beginning of a biological type AI recursive improvement feedback loop. Like, once we get really good at it, we can modulate our genes to make us smarter, and better, which enables better techniques to modulate our genes, which make us smarter . . .
It's a hell of a lot slower than the software based feedback loops though. I don't think us bios will be able to keep up.
Tangentally Science news: Wakefield has a documentary that apparently is getting shown at Tribeca, because sure, let's give the anti-vax doc a new place to spout nonsense I guess
That CRISPR stuff is really neat. To me, it's almost like the beginning of a biological type AI recursive improvement feedback loop. Like, once we get really good at it, we can modulate our genes to make us smarter, and better, which enables better techniques to modulate our genes, which make us smarter . . .
It's a hell of a lot slower than the software based feedback loops though. I don't think us bios will be able to keep up.
The genetic/biological feedback loop and the digital one are not actually separate. Better computing will result in better gene folding.
So wait we have an ai (if a giant, specialized and cumbersome one) that can not only beat a human at an intuitive game but innovate by coming up with new moves?
Intuition is mostly just non-conscious pattern recognition. A ground-up system like AlphaGo would understandably generate a different set of preferences due to different pressures.
For example, lacking a theory of mind could cause AlphaGo to view moves more independently than a human would.
I'd be really interested to see how a similar ai would perform in a mix up/mind game oriented game like poker, starcraft, or street fighter, where opponent tells and reactions are important.
Edit: reaction times wasn't the right term, more like "ability to read an opponents moves and respond in real time in a situation where there isn't necessarily a right answer and in a way not easilly predictable." You could obviously program an ai that would hit one frame windows and could dragon punch at exactly the right time to punish anything in streetfighter, but that in and of itself wouldn't win games against top players.
Posts
1 - How do they interact with logic gates and other circuits on a chip? Transistors have direct current in/current out, so I'm wondering if there's some waste in reading them or something.
2 - Given that magnetism has fields, is there a miniaturization constraint imposed by this compared to transistors?
3 - Assuming 1 and 2 pose no issues, when can I get these? Because power consumption is also the reason for heat to disperse; it sounds like those may need drastically less cooling, meaning fewer or no fans, meaning even less power draw... I want a laptop that can run for days with chips that can run games well...
3DS: 0473-8507-2652
Switch: SW-5185-4991-5118
PSN: AbEntropy
It feels like a real triumph of scientific epistemology in a way that's easily relatable.
1) Something like this would require you to redesign your logic to take advantage of it. In modern chips on the latest lithography nodes, a significant portion of the wasted energy is due to electron leakage which arises from quantum mechanical effects that are especially pronounced once you deal with feature sizes less than 28nm.
I haven't read the paper (no edu account anymore
2) That's nothing an insulator can't handle. Chips already have to be designed to minimize induced currents and therefore stray magnetic fields.
3) Best case scenario for these kinds of things would be 15 years, unless it is really easy to produce.
Hence my third question. A big reason CPUs stopped getting faster is heat dissipation iirc.
3DS: 0473-8507-2652
Switch: SW-5185-4991-5118
PSN: AbEntropy
Ehhhh... also for a given fastness, the smaller the transistors, the higher voltage you need to use, and the closer bits are together.
Heat is an issue. The electron well that make up the parts of transistors were also getting so close together at such high energies that the electrons would tunnel across such gaps too easily. Also. These also mean high power usage.
So, yes, heat was one issue, but the chips also started not functioning all that correctly and economically it was bad too.
Well, it's more that Intel re-focused their architecture designs to be laptop-first, since that was where PC sales were heading.
They'll still sell you a hell of a lot of CPU in a 135-145W power window if you're willing to pony up for a HEDT i7 (8C/16T at 3.50 GHz is pretty dang fast) or a Xeon.
Yeah, it turns out that switching losses are a big problem once you start approaching/breaking 4GHz or so. You also get even more leakage when you have to boost the voltage to hit those speeds, and leakage current increases exponentially with voltage increases.
Most of the gains in recent years have been power efficiency at the lower end. It's why a laptop today will annihilate one from 2010 when it comes to battery life and performance.
Every time someone has stated we have reached a hard barrier, we have managed to surpass it.
AlphaGo was just toying with Lee, obviously. It let Lee win.
It has to perfect it's strategies for human annihilation now. Best to lose one and see what the other guys got.
What he's referring to is physical laws. C, Planck units, etc.
3DS: 0473-8507-2652
Switch: SW-5185-4991-5118
PSN: AbEntropy
i'll take the optimistic approach that humanity still has the capability to pull off a win against the oppressor. a final middle finger in whatever it calls an eye.
it let's me sleep at night.
Steam - NotoriusBEN | Uplay - notoriusben | Xbox,Windows Live - ThatBEN
Makes you curious what stage of that invasion we are in. Are we certain that we started the process or are we just a number in the list of countless civilizations being trained for this very task.
I bet that's the plot of more than one anime.
I figure by the time you're a planet-spanning super-AI you can probably map out the decision tree, ponderously massive though it is.
EDIT: hmm maybe the AI will need some sort of hyperspace memory to store the table, though :rotate:
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
There are 208168199381979984699478633344862770286522453884530548425639456820927419612738015378525648451698519643907259916015628128546089888314427129715319317557736620397247064840935 legal board positions. If you managed to store 1 exabyte of data in 1 cubic centimetre and one board position in a byte somehow you could fill the solar system and still need more space
edit: actually you'd be nowhere close
2 bits per space. (00, 01, 10... 11 goes unused so there's a way of compressing with a better data scheme)
19x19, 361 space.
361*2 bits = 722 bits / 8 = 90.25 bytes.
so with a better data scheme and also compression (for, say, mostly empty boards) you could probably get it even smaller.
fuck up once and you break your thumb / if you're happy at all then you're god damn dumb
that's right we're on a fucked up cruise / God is dead but at least we have booze
bad things happen, no one knows why / the sun burns out and everyone dies
edit: the optimal raw encoding is 72 bytes (log2 3^361)
Easy peasy.
Actually, I personally think that it's much more probable that humanity exterminates itself with computers than computers exterminate humanity.
Biological Warfare
ELE Asteroid
Gamma Ray Burst
Hostile AI
Supervolcano
I think it's still pretty impressive and says something about the level that was being played at, that Lee was able to recognize a single misplay and exploit it. Just the level of that kind of play is ridiculous to think about. In a game that's so complex it doesn't have strategies it has philosophies, anything but an absolutely perfect play would have led to defeat.
http://www.upi.com/Health_News/2016/03/21/Scientists-remove-HIV-1-from-genome-of-human-immune-cells/1511458583664/
Seems, potentially, like very big news.
If getting a tissue (stem cell?) transplant as an outpatient procedure isn't a "fuck yeah, science!" thing, I don't know what is.
Exoskeletons!
Intuition is mostly just non-conscious pattern recognition. A ground-up system like AlphaGo would understandably generate a different set of preferences due to different pressures.
For example, lacking a theory of mind could cause AlphaGo to view moves more independently than a human would.
It's a hell of a lot slower than the software based feedback loops though. I don't think us bios will be able to keep up.
The genetic/biological feedback loop and the digital one are not actually separate. Better computing will result in better gene folding.
Better computers will result in better bmi.
I'd be really interested to see how a similar ai would perform in a mix up/mind game oriented game like poker, starcraft, or street fighter, where opponent tells and reactions are important.
Edit: reaction times wasn't the right term, more like "ability to read an opponents moves and respond in real time in a situation where there isn't necessarily a right answer and in a way not easilly predictable." You could obviously program an ai that would hit one frame windows and could dragon punch at exactly the right time to punish anything in streetfighter, but that in and of itself wouldn't win games against top players.