Pluribus

poker

No-Limit Texas Hold’em is the latest machine victim. A team at Carnegie-Mellon created an artificial intelligence called Pluribus that defeated top human players in games involving thousands of hands. We’ve already seen the downfall of the chess grandmasters, then the kings of Go, and now it’s the poker whizzes. I really liked what one of the players said afterwards:

Michael “Gags” Gagliano, who has earned nearly $2 million in career earnings, also competed against Pluribus.

“It was incredibly fascinating getting to play against the poker bot and seeing some of the strategies it chose” Gagliano said. “There were several plays that humans simply are not making at all, especially relating to its bet sizing. Bots/AI are an important part in the evolution of poker, and it was amazing to have first-hand experience in this large step toward the future.”

This is the way to look at these things. It is not surprising that AIs can brute force their way through board games. All the information and all the possibilities are available to see and analyze. Poker adds the problem of incomplete information (you don’t know what the other guy has) to the psychological aspects (is he bluffing?) which should be harder for a machine to deal with.

Apparently not.

Noam Brown, one of the researchers, had this to say:

“This is one of the interesting things about this AI, it’s not adapting to its opponent,” Brown said. “It has its strategy. It’s fixed, it doesn’t change what it’s playing based on how the humans are playing. This whole idea that there could be such a strategy in the game, I found really fascinating and that’s what really drew me to studying it more. It was kind of mystical, in a sense, there’s this strategy that we know exists, but we can’t find it.”

I’m not afraid of computers replacing humans because there are plenty of places where computers will do a better job. Autonomous driving will be a marvelous thing, for example. Air and space craft are mostly too complex to fly and they rely on machines to do the heavy work. Traffic control, whether it is trucks, ships, or satellites, will require AI.

There are a lot of complex processes in our lives. We’ll need help. AIs can find novel, original strategies that may be out of reach for humans. We know we can’t solve some problems, so we invented stuff that would solve them for us.

I can live with that.

Feed your head

It certainly beats getting your head shrunk. According to Neuroscience News:

A new study*, published in Psychiatry Research, has concluded that psychiatric diagnoses are scientifically worthless as tools to identify discrete mental health disorders.

It doesn’t surprise me. Medicine is as much art as science, and the hardest stuff has to be head medicine. We barely have definitions for things like “mind” or “consciousness” so treating what we assume are mental disorders is just tinkering with a black box.

Doctors are more like engineers, anyway. They do a lot of measuring and experimenting and they expect a lot of failure. They fiddle with complex systems in the hope of optimizing them.

That’s fine when it is knee pain or ulcers or gout or somesuch. But the head is a different place and the medical approach isn’t cutting it. Here’s from the study:

Lead researcher Dr. Kate Allsopp, University of Liverpool, said: “Although diagnostic labels create the illusion of an explanation they are scientifically meaningless and can create stigma and prejudice. I hope these findings will encourage mental health professionals to think beyond diagnoses and consider other explanations of mental distress, such as trauma and other adverse life experiences.”

People are different. Response to treatment, even with well-established drugs and other therapies, varies considerably.

A pragmatic approach to psychiatric assessment, allowing for recognition of individual experience, may therefore be a more effective way of understanding distress than maintaining commitment to a disingenuous categorical system.

There’s more, but it feels like piling on:

Professor John Read, University of East London, said: “Perhaps it is time we stopped pretending that medical-sounding labels contribute anything to our understanding of the complex causes of human distress or of what kind of help we need when distressed.”

People need head therapy. And they need to rely on psychiatry to provide some of it. Obviously people have access to other kinds of help—counselors and psychologists and whatnot—but psychiatrists have an M.D. attached to their names and that has some pull.

We’ve a lot to learn about the mind. I don’t mean to pick on the shrinks. This study is a critique of a standard diagnostic manual that doctors use. A big complex system like that can often seem real and can replace thinking with simple heuristics, making tough decisions easier. That’s all fine and good, but sometimes the uncertainty is more important.

I suspect advances in neuro-imaging will start to pay off and we’ll get clearer ideas about some mental states. Research in artificial intelligence will teach us a lot about behavior, I think, even if we don’t want to hear it! In the mean time, enjoy the mystery, and feed your head.

 

*Heterogeneity in psychiatric diagnostic classification; Allsopp, Read, et. al.

Plastics are Forever

Doesn’t have quite the same ring to it as “diamonds” does it? Besides, that’s the beef with plastics—they last forever! Diamonds do too, but that’s the appeal. With plastics we really want them to go away when we are done with them. Pesky buggers like to stick around, and even when they fall apart into little plastic bits they still last forever.

Recycling is the Holy Grail of plastics. There are too many kinds right now and they are too hard to sort and separate. When you melt down a mix of plastics into a goo you get a new plastic, not what you started with, and that’s not the intent of recycling.

plastics_waste_management_1960-2015_final

You can see we have a growing problem. Some have suggested we replace petroleum-based plastics with bio-plastics that will degrade after use. So far the materials tested are heavier, costlier, and less durable than what we have. The reason we use so many damn plastics is that they work! They are light, cheap, and tough, and have applications in so many areas in our lives that we would literally be lost without them.

The solution to any waste problem is to plan ahead. Manufacturing is traditionally a one-and-done affair. You get some raw materials, you stamp out your widgets, you sell them and send them away. If you make money you do it all over again.

In today’s world we ought to think about the end use and recovery of the materials before we make anything. That way we have a plan in place for the waste stream. In fact it is not a waste stream at all but a resource. The glass, aluminum, and steel industries all count on a significant portion of recycled scrap in their production processes.

The geniuses at Lawrence Berkeley Lab are looking to do the same with plastics. The Molecular Foundry has been experimenting with this stuff called poly-diketoneamine (PDK) that may do the trick. Plastics weren’t built to be recycled. Plastics made from this stuff are, however.

All plastics are polymers, made up of smaller, repeating units called monomers. The base materials are the same (hydrocarbons) for all plastics but it’s the additives that determine the different properties. These additives make the various types of plastics incompatible with each other when it comes to recycling. Once added these chemicals can’t be un-added and those polymers can’t be broken back down into their constituent monomers.

It is different with PDK plastics. The bonds are reversible and the end-use plastics can be broken down and the original monomers recovered and formed into new plastics again. The linear path of creation to use to disposal now becomes a circular one of creation to use to re-use. Now that’s a “forever” plastic!

I remember once talking to a nuclear engineer about the nuclear waste problem. He thought it was a mistake to bury the wastes too deep. He didn’t want the stuff to be too hard to get to. He said “we might need it later.” It was not something I’d heard before. He talked about all the energy and human ingenuity that went into the extraction and processing of these materials and what a shame it was that we would throw them away when done with them. The fact that they were still radioactive, to him, was a good thing. That meant they still had value.

I sure liked his perspective. We need more thinking like that.

Apohele

I learned a new word—Apohele. You say AH-POE-HAY-LAY, and would it surprise you that it is from Hawaiian? It means “orbit” and is an unofficial name for a group of asteroids that are also called  Atiras.

Atira is more official because it has generally been the custom in astronomy to name similar objects after the first of those objects to be spotted and recognized. The first in the class was asteroid 163693 found in 2003. The newest member of the class is 2019 LF6. This recently-found asteroid is about a kilometer across and travels around the Sun in 151 days, the shortest “year” of any non-planet object. Mercury’s year is 88 days while Venus’ is 225 days. (Check out the graphic below.)

There are about twenty such Apohele known. All of them orbit the Sun within the Earth’s orbit, which makes them unusual. Orbits are elliptical so they have a far point (aphelion) and a near point (perihelion). Apohele have aphelions less than Earth’s perihelion.

NEAs or Near Earth Asteroids are broken into three other groups: Amors, Apollos, and Atens. Amors are outside of Earth’s solar path but within Mars’ orbit. Apollos and Atens cross the Earth’s orbit but vary in orbital size. To get technical, Apollos have a semi-major axis larger than Earth’s, and Atens have a semi-major axis smaller than Earth’s. (You might remember such things from high school math.) All are names from mythology—Atira is a Pawnee goddess.

So which one do you like better: Atira or Apohele?

2019lf6_orbit_mar82020

Real People. Not Actors.

Chevy ran a pickup ad during the All-Star Game last night that featured a bunch of jamokes checkin’ out the latest cool ride. They oohed and aahed over a cheesy electric tailgate and then a bunch of rocks fell from the sky and landed in the cargo bed.

And then we are informed that these are “real people” and “not actors.”

Well I say they are a bunch of scabs. They didn’t do that shit for free, they got some kind of renumeration, even if it is just “I’m on TV” bragging rights. So since they aren’t actors that means the company brought in scabs to film the commercial. Actors are real. They have unions and everything.

Either that or the company is a bunch of cheap motherfuckers who don’t have the decency to pay people who do work for them. Actors are hired. They don’t just show up.

Yeah, like I want to buy products from that outfit.

In troubled times it is worth remembering that “I’m not a doctor, but I play one on TV.”

Energy- intensive

I recently finished Vaclav Smil’s Energy and Civilization: a history from MIT Press.

It is not light summer reading. In fact, the book requires a good deal of concentration as it is information dense, quantitative in nature, and grand in scope.

smil cover

In short, it is energy-intensive.

But that is to be expected. It’s a big story with a lot of actors and a timeline of millenia. It is history, for sure, but not the usual stuff of battles and leaders, or prophets and zealots. It is more incremental and focuses on technological changes, everything from the moldboard plow to the integrated circuit.

And he puts numbers on everything. (Be sure to brush up on metric units!) I like that. If you can measure something, or at least work up a reasonable estimate, then why not do it? It beats just talking. Sometimes numbers aren’t revealing, but most of the time they are, so it is just intellectual laziness not to whip out the slide rule and report results. Smil is careful to explain the assumptions behind the numbers and to describe their uncertainties. That is essential to a good discussion.

Energy and Civilization won’t make Oprah’s list but it will enrich your view of the world. Smil shows us the energy costs of living the way we do and he asks us if we should continue to do so. I think that’s an important question regardless of the answer.

Libra

Facebook wants to get in on the currency business. They are creating a new form of digital payment called Libra and calling it a cryptocurrency. It isn’t really, but it relies on the same technology, the blockchain. A blockchain is a software product—a bunch of computer code that enables a task. In this case the creation and maintenance of a ledger, or record-keeping system.

Blockchains are appealing because they rely on a distributed network of nodes and users. They are decentralized and thus theoretically more secure and more robust than any one particular data location. At least that’s what the computer geniuses say, and I’m not here to argue with them.

The Libra blockchain will be run by a non-profit composed of a lot of fancy folks. Here’s a nice graphic about it:

Libra-Association-Founding-Partners

The wiggly-wavey thing is the symbol for Libra currency. Users of Libra will get a sort of PayPal on steroids. They will have a digital wallet and they can pay for things (and/or receive payment) wherever they have access to the wallet, be it on a computer or smartphone. If you have an internet connection you have a payment system. Libra’s goal is to be worldwide. You buy Libra with your local currency and get an equivalent amount of digital money, and since it is digital it can be moved around and transacted more readily than ordinary currency. And you avoid the payment of transaction fees like with credit cards (there are fees, just vanishingly small to the ordinary user).

You can’t make money with Libra. It is just a platform, and you can’t earn interest on your deposits in your digital wallet. But the whole thing is a money-making strategy for its investors. You got an extra ten million bucks? You can be one, too. People who invest ten million US dollars expect a return. And the Libra Investment Token will give them that. The more people who use Libra the more these investors will make, kind of like the old-fashioned pyramid schemes.

Libra is not a cryptocurrency like Bitcoin because Bitcoin derives its value from the underlying blockchain technology itself. People crunch numbers with computers to support the blockchain cryptography and they get paid in Bitcoin. Libra is more like a traditional investment. The Libra Association will invest in fiat currencies, government bonds, and other securities to guarantee a stable value for Libra. In that sense it is more like an ETF than a cryptocurrency. Due to the volatility of things like Bitcoin a market has emerged for what are known as stablecoins which are tied to real-world assets like commodities. Libra is closer to that.

Users of Libra coin gain access and convenience. For people who don’t use banks, this could be a real boon. Libra Investment Token holders will get rich, very rich, even if only a small percentage of Facebook’s user base of over two billion people participate.

And what does Facebook get out of all this? You won’t need a Facebook account to use Libra, and Facebook’s digital wallet, called Calibra, won’t require one either. Mark Zuckerberg will tell you it is about serving the under-served in the world financial marketplace, which is all very nice, but ultimately this is a for-profit company.

A digital platform is a digital platform. If people use it, you can use that platform to sell other products and services. I assume Facebook’s leap into new forms of currency is another version of their current business model.

Software is famous for bugs. And also for being marketed well before it is bug-free. Just look at the Boeing 737-MAX tragedies to see the consequences of that. I’m not going to put my money in a Libra coin/Calibra wallet just yet, but I won’t be surprised if we all have something like that in the very near future.