Saturday, January 30, 2016

The (Incredibly Cool) Future of Personal Computing

Shhh. They don't actually 'exist'...
When you use a website or web application, the processes that make it work are not running on anything you'd normally recognize as a 'computer' or a server. The infrastructure that hosts the modern web is a disparate collection of bits scattered around the data center. It's impossible to walk in the door, point to one rack of hardware and say "that's the AirBnB servers". Most compute hardware today is virtualized, delivered not as a box of hardware, but as a set of services at a defined service level. Just as in the early days of electrification, if you wanted electricity, you had to generate it yourself, in the early days of the internet every server represented a capital investment of thousands of dollars, and the thing that killed so many early internet business was growth. Adding capacity took time and a huge amount of money.

Later, electricity was produced in large-scale facilities and was available for purchase off the grid. You didn't have to have infrastructure, and you only paid for what you needed. That is the nature of compute resources today. You specify a system - some processor cores, some memory, some storage, a network fabric - and you supply or specify an operating "image" - the OS and software stack you want to run. Then you tell your cloud provider you want five hundred of these virtual servers. When demand runs high, you easily add another thousand, and when you no longer need them, you shut them down, minimizing your operating overhead and producing profits that would have been impossible a few years ago.

This model has changed everything, and made computers much more powerful tools. When processing power requires no up-front capital investment, all of that capital can be shifted to software development. And as a result, we've seen huge advances in machine intelligence that have resulted in incredible technologies. At the high end, that means no more multi-million dollar super computers. Just hand over a credit card and spin up 50,000 cores on AWS, run them for 20 or 30 or 50 hours, get your results and shut them down. Anyone can do it, and they can do it NOW. At the low end you get Siri, and OK Google, and real-time navigation and traffic, at the high end genomics research and climate models on-demand.

But so much more is coming. The simple fact is you no longer need an actual computer. They are just another resource, and you can use one with the power (and cost) appropriate for the job. When you're surfing the web and and chatting on Facebook, you can use a tiny VM that costs virtually nothing. When you want to play a more demanding game or use a higher end piece of software, you seamlessly switch to a more powerful platform, and only pay for it while you use it. If you want to use a Mac for internet and a Windows PC for games, you don't have to buy the OS - the image comes provisioned.  Instead of the computer you are using today, you merely need a box that connects you with your resource provider, and a display.  Even your handheld and wearable devices gain access to much more powerful processors and you are never again constrained by what hardware platform you chose to buy, say, two years ago. You'll always have the latest version (or a specific older version if you prefer) and all the available security patches.

And once we are freed from the costs, complication and compromises of hardware purchases, access to the global digital infrastructure will truly be ubiquitous. Capacity will be whatever we need, whenever we need it, and even dedicated platforms like game consoles and Bitcoin miners can be virtualized.

THAT'S the next step, and it's coming to a screen near you.

Friday, January 29, 2016

Sagittarius A*

Never forget to be awestruck by what we can do
What the heck is Sagittarius A*, you ask? Is it a band? An album? A novel? An indie movie? Nope, none of those things. Sgr A* (pronounced Sagittarius A Star), as the scientists like to write it, is an intense radio source in the center of our home galaxy, the Milky Way. Scientists have long believed that at the center of every large spiral galaxy there resides a 'supermassive black hole', with a mass of several million suns providing the 'anchor' around which gravity could build the galaxy itself. Proving this perfectly reasonable assumption, however, has been problematic for a number of reasons.

First, the very concept of a black hole makes it impossible to 'see'. A black hole emits no energy at any wavelength (well, except for something called 'Hawking Radiation' that we can skip over for now), so none of our observational tools can measure it directly. And second, the center of the galaxy is a very dense, busy, messy place full of huge clouds of interstellar dust and gas - essentially the same stuff planet earth is made of. These clouds effectively block a broad portion of the EM spectrum, including visible light, which prevents us from seeing much of anything going on in the entire central region.

But all is not lost. We have radio telescopes, which capture the longer wavelength energy that passes easily through the obscuring interstellar medium, and we have IR - infrared radiation that is generated by heat, that can be detected behind and through the clouds of dust and gas. And what's that we see? Stars orbiting the black hole in small, tight orbits. The bell cow was a star they called S-2, and by measuring the mass of that star and analyzing its orbital dynamics, scientists could reach some conclusions about the structure of the black hole we call Sgr A*.  And the numbers are beyond mind boggling. With the mass of nearly four and a half MILLION times that of our sun, it's only 13 million miles across. The density of an object like that is unimaginable. And sitting way out here, we figured all that out.

Is this a new breakthrough? Actually, far from it. Sgr A* was first discovered in 1974, and named ten years later. You'll notice on the attached image that we've been tracking the orbits of surrounding objects for more than 20 years. Also note that the key objects, S-2 and S-102 together encompass a tiny bit of sky less than half an arc-second across. This is precision astronomy of the first order. And here we are, a species of intelligent primates evolved all on our own on a dusty little rock in a distant rural corner of our galaxy, looking at the enormity and complexity of our universe, and coming to understand it. If I take any comfort in my own mortality, it is rooted in that accumulated knowledge, and in our drive and cleverness that allows us to keep accumulating more.

Saturday, January 23, 2016

The Twitter Paradox

One of these is not like the other
Twitter and Facebook are often considered a kind of bookends, the quintessential social media platforms. And it's true that, in general, they both serve as a kind of global digital communication platform, albeit serving altogether different purposes and constituencies. And as a result of this tendency to view them as two pieces of the same business model, their financial and operational performance is often compared. When, industry watchers wonder, will Twitter become profitable? Why has their growth stalled while Facebook continues to add users and subsidiaries? And yet, the comparison is odd - Facebook is by far the industry juggernaut, the second most viewed site in the world with a billion and a half users and a market capitalization of $40 billion. Twitter, on the other hand, has struggled to acquire 330 million users and has a rather hard-to-justify market cap of $12 billion.

Considering that neither platform will ever have the option of charging their users any kind of subscription fee, both companies have had to be somewhat innovative in how to monetize a gigantic global digital communication service. Where Facebook has been mostly successful, Twitter continues to try to figure out a model that would make the service economically viable and financially sustainable. From where we stand here in January 2016, the most likely forecast will be that, at some point, Twitter will be subsumed into a larger ecosystem that can afford to subsidize its operating costs forever. The alternative is that it merely sinks beneath the waves, to be 'replaced' - as much as possible - by a large number of smaller services.

But here's the paradox. Of all the social media tools and services available today, the only one that is absolutely irreplaceable is Twitter. The world could simply not do without a Twitter, and the human and social costs if it were to disappear are hard to even imagine. Through Twitter, the way we see the world, the way we understand events in far-flung places, the way we participate and interact has changed forever. Long before the first headlines break on the news sites, those of us on Twitter are getting eyewitness reports, rumors, live video, cries of anguish and sobs of joy. With Twitter, we can now be everywhere at once, meeting the actual people who are doing the actual things that change the world. From Fukushima to Maidan, from Tunisia to Cairo, from Wall Street to Ferguson, we learned about the world first hand, and in real time. And we were able to make our voices heard.

So this is the paradox, and the challenge. Twitter may very well not be something that can be operated profitably. But as a people, as a society, as a species, we must never lose access to these kinds of global, instant communications. If Twitter were to fail as a business, it would be necessary to continue to operate it as some kind of global NGO - ideally under the auspices of the UN, but at any rate in a way that continues to permit the world to have unfettered access to this kind of real-time conversation. There's just no going back.

Saturday, January 16, 2016

...And Expecting Different Results

A long list of things he doesn't have the systemic
Authority to implement
In the Democratic primaries, I really don't have much of a dog in the fight¹. Primarily because it is so utterly crystal clear that a Sanders presidency and a Clinton presidency would look essentially identical. But it's true that I do sometimes get snippy with Sanders supporters. I don't ask for much, but I do like my political conversations grounded in observed and observable reality. No, if you clap harder, you won't get a pony, and no, if you click your heels you won't go home. The American system of governance has rules and friction and veto points - many, many veto points - and merely being the President isn't enough to overcome them.

Many of Sanders' supporters have gone way over the top, speaking breathlessly about a 'revolution' in American governance, a new era in social justice. It's gotten so bad they are even starting to predict support from the right. For Bernie Sanders. As President of the United States.

Of course, the Sanders campaign is ultimately responsible for much of this hyperventilation. They have felt no need to explain what they'd do or how they'd do it - instead of policies, we get bullet points. Instead of plans, we get a rehash of his ten years in the Senate.

He'd break up the banks. But, of course, the President CAN'T break up the banks, and he gives us no sense of how he'd try to get that kind of legislation through congress and the courts. (ProTip - he can't.) He'd reduce income inequality. Ok, that would mean some kind of income redistribution - how would he do it? Tax policy? No, the President doesn't make tax policy. Government spending programs? No, the President has no power to allocate funds. Education? Well, to work on tuition reform he needs to work through fifty different state governments along with hundreds of private institutions - good luck with that. Student Loan reform? Um, once again, that's done through legislation and approved by the courts.

In the case of the Clinton campaign, we know a great deal more about her policy agenda, and we know the Republicans will just as gleefully block and derail it at every opportunity. The only difference is she is more likely to get the support of the more 'moderate' Democratic legislators - but with the huge Republican majority in the House, that's unlikely to change anything to any significant degree.

Remember back in January of 2009, the outpouring of joy as we inaugurated Barack Obama, ending the foul, ugly GW Bush administration? Remember the lofty rhetoric? He'd close the illegal prison at Guantanamo Bay. He'd end the wars in Afghanistan and Iraq. He'd prosecute the criminals who destroyed the economy with their criminal greed. And remember the bitter anger and recriminations from the American political left as reality set in, and we discovered that not only could he NOT do these things, there was also assassinations and massive domestic surveillance and an outright war on whistle blowers?

Now, imagine it's the late spring in 2018. The Sanders administration² is well into its second year. And nothing - NOTHING - that he promised in his campaign has been accomplished. The Republicans in congress have gleefully blocked his every move. Many of his proposals are too far left for a lot of Democratic congressmen, who join with their Republican colleagues to derail them. An overwhelmingly conservative Judiciary blocks and obstructs his executive actions. Can't you just hear the shrieks of anger and frustration from liberals over his 'betrayal' - the very people that helped forge the narrative of unworkable promises that got him elected in the first place.

1 - Actually, I very much prefer Hillary Clinton as the Democratic nominee, not because of any significant differences in effective policy, but strictly due to electability. The Democrats have a structural advantage in the national popular vote of between 3 and 8%. Sanders gets the far left wing of the party - but so does Clinton. They're not going to vote for the Republican. But the 'moderate' wing of the party - rural and blue collar voters is NEVER going to vote for a self described Socialist. 

2 - The reality is, despite what Sanders supporters who haven't bothered to learn how the American primary election system works want desperately to believe, that Clinton has a HUGE structural advantage in the only thing that matters - delegates. The way the Democratic party allocates delegates - proportionally rather than winner take all, along with 800 party elite "Superdelegates" - strongly favors the establishment candidate. Numerically, it's almost impossible to conceive of a Sanders nomination.

Friday, January 15, 2016

Reaping the Whirlwind

Thank you, Wolfram Alpha. Yes. We're doomed
In 2005, the price of crude oil hit $70 a barrel. And we simply knew things. We knew that peak oil was upon us. We knew that growing demand in the developing economies, along with the impending exhaustion of the 'easiest' reserves to extract had placed the world in an economically precarious condition. We knew that demand was almost exactly commensurate with the available supply, and every time the economy boomed it was likely to cause a spike in oil prices. We also knew that this was essentially an irreversible condition. That energy producers would have to exploit more difficult petroleum resources, and as production costs exceeded $50 per barrel, a kind of Rubicon was crossed. By 2008 the price rose above a hundred dollars, and except for a brief collapse after the global economic meltdown, it stayed in that triple digit range.

Today crude oil prices will close under $30/bbl, dragging equities markets deeply into the red. What could possibly have caused such a dramatic plunge so quickly? Two factors worked in tandem, a kind of negative feedback loop that eventually blew up the entire system. First, demand was rising throughout the world, and there was no reason to presume that would end. As emerging economies developed an urban middle class of consumers of their own, the growth of automobile travel would skyrocket. And modern cities and factories would consume massive amounts of energy. In the energy sector, particularly in North America, oil executives saw this permanent increase in demand, and particularly how it was bumping up against supply constraints. With middle east wars and sanctions, the world could barely pump enough oil to keep up with average daily demand. Any spike, however small, would create at least local shortages.

But these oilmen knew of other sources. They knew that using methods like hydraulic fracturing and horizontal drilling they could access huge additional reserves, far away from the chaos of the Gulf. Oil like this is more expensive to extract and refine, but with prices holding well above a hundred dollars, it was immensely profitable. By 2015 the US was once again the largest Oil and Gas producer in the world.

While all this was happening, nations around the world were waking up to the realities of carbon pollution and climate change. First came significant increases in energy efficiency, and then large and growing deployments of alternative energy sources, primarily solar and wind power.

By 2014 the global economic slowdown was in its fifth year. China was contracting, and demand for oil throughout the developing world was in free fall. Meanwhile, there was much more oil on the market, and the adoption of alternative energy sources was reducing demand in the developed world. A vicious cycle that had only one solution. In time past, Saudi Arabia would simply reduce production volumes to create an artificial shortage, and the price would return to optimal levels. But now the Kingdom felt the economic challenge from US and Canadian oil production, and, locked in a cold war with Iran, chose to defend her economic position by keeping the taps wide open. The gamble was that House of Saud and the Emirs would be the last man standing after everybody else quit the market.

That particular wager hasn't played itself out yet, but the Saudi economy is in fairly bad shape, running unprecedented deficits and reducing the energy subsidies it had always provided for its citizens.

And there you have it. A modern-day trillion dollar morality tale of the wages of greed in a global zero sum game. But before you spend too much of your schadenfreude budget gleefully considering the fate of the oil barons, remember that the global economy is teetering, and economies from Russia to Saudi Arabia to Iran to Nigeria are teetering with it. On net, this is not a positive contribution to peace in our time.

Wednesday, January 13, 2016

That's Just About Enough...

They're clowns. But they're dangerous clowns
I know, I don't usually do this kind of brief, timely commentary here, but I'm starting to get seriously pissed off. I realize that we, as a people, have a deeply ingrained structural hypocrisy built into the most basic functioning of our system. That's not arguable. But we're supposed to have limits. We need some kind of a fig leaf that allows us to pretend we are a nation not only of laws, but of laws that apply to all people equally. We need to have a narrative that justifies the murder of young black men by white police officers, the racial and social imbalances in criminal justice and incarceration, the endless legal roadblocks placed between poor Americans and the voting booth - we tell the stories, and we argue about them. But the stories are there - the fig leaf covering the obscenity.

Which brings us to Burns, Oregon. What is the narrative that prevents our law enforcement agencies from acting? Sure, I hear the argument - just wait them out, arrest them when they quit, no need to escalate the violence. But if that was really what we were doing, we'd have a perimeter around their 'occupation'. We'd arrest anyone who came out. We'd turn off the power and the water, and interdict any supplies going in. In short, we'd be driving an end to the standoff without escalating the violence.

Now we know they've accessed federal files. They've destroyed federal property. They've trespassed, vandalized, threatened, intimidated and incited. The fact that no charges were ever brought after the Bundy Ranch adventure of 2014 is the single salient event that led us to where we are today. If you have a national policy that you don't negotiate with terrorists because you want to discourage terrorism, how can you not slap down these kinds of acts of domestic terror for the same reason?

Recess is over. It's time for federal law enforcement to start bringing this thing to a close - one that ends hard and ugly for the likes of Aamon Bundy and Jon Ritzheimer. To continue to encourage these kinds of activities with government passivity and inaction  - particularly in the face of the violent police response in Ferguson and Baltimore - is an abomination that merely guarantees more armed takeovers and leads us in a dangerous and violent direction. Unless and until these people understand that there is a very heavy price to be paid for armed insurrection, we'll have more events like this, and they'll get increasingly violent.

Friday, January 8, 2016

A-Bomb, H-Bomb, Potato, Po-tah-to

That right there is where the action is
Earlier this week, North Korea conducted another nuclear weapons test in the far Northeastern hinterlands at the Punggye-ri nuclear test site. While this is categorically NOT a good thing, and drives a lot of hard, real-world policy questions, the overwhelming buzz is simply around the Kim government's claim that the device was a so-called "H Bomb". As long as we are confronted with this pointless distraction, let's actually think about what it would mean in practical and policy terms if it WAS such a device.

First, what is a Hydrogen Bomb? We discussed that at some length here, so I'll just summarize. The design in use around the world today is called the Teller-Ulam design, named after the scientists that developed it. It's essentially a two stage device that uses a standard fission device to trigger some additional yield through a fusion reaction. It's worth noting that even with a well designed Teller Ulam device, the lions share of the energy released is through fission, not fusion. Why build a device like this? Well, the breathless reporting is around the tremendously huge yields the design can produce. Ten, twenty even FIFTY megaton city killers can, and have, been built. But nuclear strategy has evolved, and as the weapons have gotten smaller and the delivery systems more accurate, most warheads are of the variable yield type, with the maximum yields around 250-300 kilotons. The real value of an effective Teller-Ulam physics package is that it allows  the use of much less fissile material, resulting in more, smaller, more deliverable, more efficient weapons.

Ok, that makes sense for nations with large stockpiles. Why would North Korea want a bomb like this? The answer is they wouldn't. The design is difficult to get right - remember that the detonation is initiated by an atomic bomb. EVERYTHING that happens after that has to happen in the microseconds before the bomb breaks apart and is vaporized by that initial blast. Get a tiny detail wrong, you get the energy release of the primary and nothing more. In the nuclear weapons argot, this is called a 'fizzle'. There would be no reason for North Korea to add that level of complexity and reliability problems to their tiny stockpile when plain old fission or boosted fission weapons would serve them just as well. Indeed, Israel has never developed a two-stage design, preferring to hew to the tried and true boosted fission designs they know they can count on.

So why would they make the claim that they tested an H Bomb? Well, first the Kim dynasty has a long history of making outrageous claims. It's primarily for their domestic constituency, who tend to worship the generational leadership with something approaching religious fervor. It's important that they see a modern, advanced, technologically competitive government rather than the impoverished backwater criminals we know them to be. But it is also true that these sorts of claims tend to resonate in the international diplomatic community at a level far out of proportion to their reality. Once again, we're all talking about the type of device rather than the policy ramifications of another nuclear test in Northeast Asia.

In the end, the important reality isn't the weapons at all - it's geography. Just 90 seconds flight time from the border along the 38th parallel is the South Korean capital of Seoul, a huge, modern metropolis of ten million souls. Whether with nuclear weapons or massed conventional artillery, the North knows they hold that city at perpetual risk. In any outbreak of hostilities, the Kim government might well fall, but the cost to the south in lives and treasure would be atrocious. Seoul is their deterrent, and everything else is just propaganda and positioning.

Thursday, January 7, 2016

What is the "Mainstream Media"?

Times Change - Some Things are No Longer True
When the foul and odious GW Bush was 'elected' President in 2000, the media landscape in the United States was wildly different from what we see today. In April of 2001, 6% of American households had a broadband connection, while 41% accessed the internet through a dial-up connection such as AOL. Television, print and radio represented the media in the early years of the 21st century, and they had begun to break apart into ideologically specific silos. Fox News was five years old at the time of the 9/11 terror attacks, and Rush Limbaugh's radio show had been operating pretty much in its current iteration since 1988. Internet news sites were primitive and not terribly useful, but they were immediately recognized for their ability to drive and reinforce ideologically specific narratives in a way that television, with their sponsors' sensitivities, could not. Right wing pundits such as Ann Coulter and Rush Limbaugh began to regularly denigrate what they called 'the mainstream media' for their unwillingness to use the same kind of outrageous and over-the-top rhetoric.

During the Bush years and even into the 2008 election cycle, this concept of a 'mainstream' media channel mostly delivered by television, radio and newspaper, and a second media channel, delivered by the internet that existed outside of that mainstream was quite valid. The real question being asked by consumers of media was 'how was the ideological worldview and market positioning of a given outlet influencing what they covered and how they covered it?' Although they disagreed radically on who the specific culprits and malefactors were, both left and right were in agreement that this division existed, and even on what it was intended to accomplish.

But now it's 2016. Internet delivered media content is every bit as mainstream as any other delivery medium. It's equally accessible to most, and the production quality and attention to detail is competitive with cable news. Aggregators like Google and Yahoo News sit alongside services like Reuters, CSM and McClatchy, while MSNBC, CNN and Fox have equal presences across both delivery platforms. The question that needs to be asked is not which outlets are mainstream and which are upstart truth-tellers, but rather which can be trusted to tell the truth no matter what - and, of course, if we even want to be told the objective truth. Movement Conservatives in America have made a point of shutting themselves off from any information they find contradicts their pre-existing worldview. But liberals often bristle when someone from their 'side' asks hard questions or even simply considers a heterodox position. Bright, talented writers and thinkers like Matt Yglesias and Jon Chait have fallen victim to this narrowing of the discussion parameters time and time again.

At this point, I think we can dispense with the whole 'mainstream media' construct. The question we need to ask should be applied to all media resources equally: Can what they say be trusted to reflect an objective reality, or are they starting from a given worldview and building their content in such a way to support that initial conclusion? And be well advised - people aren't going to agree on the answer. We are all conditioned to believe that information that supports our beliefs, and reject that which does not. The only answer is to avoid dependency on any one set of sources, but to examine every viewpoint in terms of that which we can determine is objectively true, to avoid conflating opinion with reporting, and to seek authoritative, primary sources whenever possible.