Artificial Intelligence: The Human Machine

My favorite American entrepreneur, Elon Musk, who is very much the Thomas Edison of our time, has helped start a new enterprise aimed at advancing the field of artificial intelligence. It's called "OpenAI". On their website they state:

OpenAI is a non-profit artificial intelligence research company. Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return.

Since our research is free from financial obligations, we can better focus on a positive human impact. We believe AI should be an extension of individual human wills and, in the spirit of liberty, as broadly and evenly distributed as possible.

An Ivory Tower?

The phrase "unconstrained by a need to generate financial return" is troubling. I'm all for non-profit. I really am. But the question arises, what is going to motivate this organization to produce anything? There's a very real risk of ivory-towerism.

Which would be a shame. So much talent and resources wasted in that case. It seems that their first priority needs to be to set measurable goals that they can evaluate themselves against. A couple of things come to mind:

• Number and quality of peer-reviewed papers.
• Number and quality of users of their planned open source software.

What is the question?

Before you can come up with an answer, you need to know the question. Before Einstein could come up with special relativity, he needed a clear statement of the problem that he was trying to solve. In retrospect, coming up with the question was perhaps the hardest part and took many people working many years. Without the Michelson-Morley experiment in 1887, for example, Einstein would not have known what the problem was he was trying to solve in 1905.

So instead of trying to come up with answers before knowing the questions, perhaps the focus needs to be more on coming up with the questions. In the context of AI, this paper from the Proceedings of Aslib Informatics provides some insight.

The human machine.

One possible way to generate the questions as well as to test progress would be to build what may be called a "human machine". This would be a platform consisting of networked software that lets a group of humans work together to interact with a human interrogator. Someone who wants to chat would go to a web site, and the other side of the conversation would be handled by the platform.

This is a difficult problem in itself, even with no AI plugged in. In one possible architecture, a single person would manage the platform side of the conversation, and would create sub-problems to be distributed to other humans to solve. How you do that, and what is the architecture, even with no AI involved, is a hard problem. Once you had such a platform, you could begin adding AI features to make it faster and better, with the goal of ultimately replacing all the humans.

The point of such an experiment would be to focus on the appropriate questions as well as to serve as a measure of progress, in terms of how satisfying the interaction is to users. In the absence of financial or military motivations, some method is needed to map the space being researched and to know if you're going anywhere, whether it's a "human machine" or something else.

Securing Your Subversive Communications

There's a lot in the news lately about encryption. If you're engaged in subversive activities, perhaps environmental activism or journalism for example, you know how important it can be to keep your communications secret.

Those of us of a certain age are familiar with the Nixon presidency and Watergate, and these days we have the Snowden revelations, but it's not just government we have to worry about. Howard Hughes, the famous aviator, aerospace engineer, movie producer and womanizer, provided financial support for right-wing spying on liberals in his day. The Koch brothers do the same today.

But keeping your communications secure is much more than encryption. Here are some of the things you need to consider:

• Finding secure services. Probably services like proton mail or tor instant messaging are secure, but how can you be sure? One of the Snowden revelations was that the NSA plants saboteurs among the developers of such software to weaken its security in subtle ways. And besides, to be secure, you need to be perfect, and no one's perfect. Still, this is probably not your biggest problem.
• Using a secure operating system. It does no good to use a secure service if your operating system is broken. Windows is hopeless and even Linux has myriad security problems, though it can be made better by using "hardening" procedures.
• Using a secure device. The hardware itself is subject to tampering. It's not paranoid to imagine that your home could be broken into and your motherboard replaced with one that is in some way more friendly to attackers. Smaller, simpler devices, such as Raspberry Pi, are easier to secure, so long as you keep them as "bare bones" as possible. And there are "hardening" practices recommended for these devices as well. To be safe, you need to epoxy the case closed and paint it with distinctive designs to prevent tampering with or swapping the device.
• Protecting meta information. Even if the content of your messages is secure, knowledge about who is communicating with whom, how much and when, can be a great help to your enemies. You have to assume that they know who you are and are targeting you. They can plant a hardware device in your internet connection or infiltrate your internet provider. So you need to use random internet connections, at cyber cafes and the like, rather than your home or work connections.
• Generating noise. One of the best ways to protect yourself is to create a lot of noisy but unimportant communications using insecure means. This will keep your watchers busy and they may be too lazy to look further. It's useless to try to hide from them. They know who you are. If you try to conceal all your activities, they will just keep looking that much harder.

Assuming you have achieved secure communication, you still have the problem of "social engineering" attacks. If you have much interest in spy stories, these will be familiar to you:

• Moles. Your enemies may plant people in your organization who are disloyal. They may be people who seem the most gung-ho and the most ideologically pure. They are likely to be newcomers to your organization, possibly with a fabricated history of support for your cause. You need to be suspicious of anyone who doesn't have verifiable history with people you know.
• Traitors. Your people can be bought, seduced or blackmailed. Look for changes in behavior, especially if they seem to be pressing for information. Their handlers will be impatient and push them to find out as much as possible as quickly as possible. This will give them away if you're looking out for it.
• Disruption. This is the flip side of the problem of moles and traitors. Your enemies can try to convince you that loyal people are actually disloyal and create confusion and conflict in your organization. There's no solution to this conundrum except to be smarter than your enemy.

With all the news that's focused on encryption lately, it's easy to forget all the other things you need consider to manage a successful subversive operation. Don't be seduced into thinking that encryption is the only thing you need. There's a lot more to the game than that.

Moral Hazard

The rescue of the financial giant AIG by the Federal Reserve during the financial crisis of 2008 probably prevented a global financial catastrophe and probably prevented the "Great Recession" from becoming another "Great Depression". As bad as it was, with the assets of the U.S. middle class being all but wiped out and millions of people losing their jobs, it could have been a lot worse and lasted a lot longer.

The financial crisis was caused by poor management of large international banks and other businesses that aren't technically banks but function like banks. They took huge risks. When the bubble burst, they were in danger of going bust and causing everyone else to go bust along with them.

The Dodd-Frank law that Congress passed after the financial crisis is supposed to fix that by prohibiting the banks from taking excessive risk. Large banks that are big enough to cause another financial crisis get increased attention from the Federal Reserve. But one part of the Dodd-Frank law prohibits the Federal Reserve from ever again rescuing large banks and financial institutions. The reason given for this change is the danger of "moral hazard", which means, if the banks think they will be rescued in a financial crisis, then they won't care how reckless they are.

The argument, however, ignores the fact that banks and bank-like companies are run by executives. It assumes that these executives care if the bank fails or is rescued by the Federal Reserve. They don't, not very much.

These executives are fairly wealthy. They are clever about money and keep some of their money in gold and other things that guarantee they and their families will be OK in any financial crisis. This leaves them free to gamble with the bank's money without much risk to themselves. If they are lucky, their gambling will pay off and they will make even more money. If they are unlucky, the bank will go broke, but they won't suffer much as a result. This is called "heads I win, tails you lose", and it's the real "moral hazard". Dodd-Frank does nothing to change this.

So we are doomed to endure another financial crisis sooner or later. The next time, though, the Federal Reserve will not have the power to rescue us and we will all end up beggars.

The Nuerosurgeon's Health Care Plan

All the Republican candidates for President are for reducing Medicare and Medicaid. Dr. Ben Carson, a retired neurosurgeon, is currently leading in the polls. He has a plan to replace Medicare and Medicaid with a plan that covers all Americans, young and old. Every person would receive $2000 a year from the government to pay for health care, with $700 to go towards health insurance and $1300 available to pay routine medical expenses. If you don't use the $1300, it accumulates in your health savings account.

It's a great plan for everyone who is not currently on Medicare or Medicaid. Quite simply, they get $2000 from the government that they are not currently receiving. For a family of four, that's $8000 a year. Very nice. Working people would still be able to get health insurance through their employer, so for them, this is just free money. If I were 18, I'd vote for that.

What about people on Medicaid? Some would be better off but some would not be better off. Those with low medical bills would get free money from the government. For those with high medical bills, the $2000 a year would fall far short.

But for people like myself, retired and on Medicare, it's a complete disaster. First of all, there is currently no private health insurance in the United States for persons 65 and older, and even if there were, it's unlikely that I would be able to get it for the $60 a month that Carson's plan provides. For someone born when the plan is in effect, they might be able to save enough from the $1300 a year that they have plenty in their health savings account by the time the get older, but that doesn't work if you're already retired when the plan goes into effect. I've had no medical expenses this year, but last year was an ordeal involving three surgeries and the medical bills were almost $150,000.

Total yearly medical expenditures in the U.S. are about $3 trillion, of which about one-third is Medicare and Medicaid. There are 320 million Americans and that times $2,000 is $640 billion a year. So the Carson plan pays less than one-fifth of total medical expenditures. Those on Medicare and Medicaid, whose expenses are now pretty much fully funded, get hit with an 80% reduction in the amount that the government pays for their care. 

There is a sense in which the Carson plan is more fair, of course. It treats everyone equally, rich or poor, young or old, healthy or sick. That's great for the rich, the young, and the healthy, but it leaves the poor, the old and the sick in a hopeless position. In particular, it pulls the rug out from under retired people who have paid for Medicare all their lives through payroll taxes and were counting on it in their old age.

My AI Program

This morning, the Loebner Prize is being contested at Bletchley Park, the location outside of London where Alan Turing and other British code breakers successfully cracked Nazi encryption.  The Loebner Prize is described on their web site as follows:

The Loebner Prize for artificial intelligence ( AI ) is the first formal instantiation of a Turing Test. The test is named after Alan Turing the brilliant British mathematician. Among his many accomplishments was basic research in computing science. In 1950, in the article Computing Machinery and Intelligence which appeared in the philosophy journal Mind, Alan Turing asked the question "Can a Machine Think?" He answered in the affirmative, but a central question was: "If a computer could think, how could we tell?" Turing's suggestion was, that if the responses from the computer were indistinguishable from that of a human,the computer could be said to be thinking. This field is generally known as natural language processing.

In 1990 Hugh Loebner agreed with The Cambridge Center for Behavioral Studies to underwrite a contest designed to implement the Turing Test. Dr. Loebner pledged a Grand Prize of $100,000 and a Gold Medal (pictured above) for the first computer whose responses were indistinguishable from a human's. Such a computer can be said "to think." Each year an annual cash prize and a bronze medal is awarded to the most human-like computer. The winner of the annual contest is the best entry relative to other entries that year, irrespective of how good it is in an absolute sense.

I've been interested in artificial intelligence for many years and at one point I wrote a modest program designed to give human-like responses. While my program is not sophisticated enough to be entered in the competition for the Loebner Prize, I thought that, in honor of the Loebner Prize, I might try it out this past week, using my neighbor here in Florida as the human who would give answers to my questions that I could then compare with answers from my computer program. The results are shown below. I've labelled the responses "A" and "B". Can you tell which is the human and which is the computer?

Question: Why is Donald Trump doing so well in the polls?

A: He will make America great again.
B: He is cynically telling stupid people what they want to hear.

Question: Is President Obama a Muslim?
A: Yes of course.
B: No, he is a Baptist Christian.

Question: Where was President Obama born?
A: Kenya.
B: Honolulu, Hawaii.

Question: What should we do about the problem of illegal immigration?
A: Build a wall.
B: Grant young illegal immigrants the right to work in the U.S.

Question: What should we do about Putin?
A: Try to get along well with him.
B: I don't understand the question.

If you guessed that A was the human and B the computer program, you're right, of course. This is just my admittedly simplistic program, and the programs competing for the Loebner prize are much better. Even so, as you can see, computer programs have a long way to go before they can match human intelligence.

Trump, The Bully

Donald Trump is many things--billionaire, misogynist, raging id, narcissist, genius. But there is one thing he is above all. He is a bully.

Bullying is out of favor these days, but here's the thing: bullying works. Just ask Vladimir Putin, or go to your favorite psychic medium and ask Adolph Hitler or Genghis Khan. I wish we lived in a world in which we did not tolerate throwing your weight around, but in my lifetime, nothing much has changed.

Trump's bullying came out in full force over his rough treatment by Roger Ailes in the recent Republican candidates debate. He charged back Hell-bent on gore. Everyone who, like myself, is voting for the Democrat for President next year no matter what, cringed at this pre-adolescent tirade. Half of those who are voting for the Republican next year no matter what, cheered, stomped and hollered. Those who will actually decide the election next year, the "undecideds", aren't paying any attention at all to this and won't until a few weeks before the election.

There's a good chance that Trump will bully his way to the Republican nomination for President. He's right that the other Republican candidates are weak, except maybe Ted Cruz. His threat to spoil the Republican chances by running as a third party candidate must be taken seriously. He has let everyone know that he will do anything to take revenge on his enemies. I believe him. In the end, the party will have to give in to his bullying and give him the nomination.

If it comes to Hillary vs. Donald, it will be the most amazing Presidential election in a very long time. Donald would have a chance to win, except for one thing. As he never tires of telling us, he is no one's puppet. Rich people, though, need to have a puppet as President. They were happy enough with Bill Clinton. They won't mind Hillary. But they will stop Trump. I hope.

Birthers Got It All Wrong

"Birthers" are those, like Donald Trump, who believe that President Obama wasn't born in Hawaii as he claims, but in Africa. They've got it all wrong. Obama wasn't born in Africa, but he wasn't born in Hawaii either. He was born on Alpha Centauri Bb, an extrasolar planet orbiting the star Alpha Centauri B which is about four light-years from Earth in the southern constellation of Centaurus, affectionately called "lava world" by its inhabitants because of its high level of volcanic activity. I know because I'm from there too and I came to earth about the same time as he did. Both our families are originally from the Skitze clan of the Alphane moon, which accounts for our tendency to blather.

I don't remember him well. He wasn't so interested in politics then. Mostly he was known for being wicked good at playing the video game Space Invaders, which on Bb is similar to but way different from the game played here on earth. For one thing, it's played with real space invaders, who invariable get the worst of it. Earthlings would do best to consider that if they ever think of invading Bb.

The year on Bb is only 3 earth days, which makes him (and me) really old in Bb years. There are no days on Bb because the same side of the planet always points to its sun. On Bb, when you want to go out for the evening, you have to travel to the dark side. It was rumoured that Obama spent way too much time there. Most Bbans stay mostly on the day side of the planet, where the work day is infinitely long and there is no weekend because there are no days. This can get to be a drag. It's why I came to earth and I'm sure it had something to do with Obama's migrating here as well.

Since it was discovered by earth scientists in 2012, there has been some scepticism here as to the existence of our home planet. I don't take offense, though. There are still some on Bb who don't believe in earth, despite the large number of Bbans who have migrated here over the years. Once we get here, there's a tendency to forget about Bb and not write home, and that may have something to do with it. My own birth parents on Bb started denying my existence soon after I left, but it's only because of a wicked law on Bb that makes parents liable for their children's exit tax if they leave without paying. The tax is really high, roughly comparable to the gross domestic product of Haiti, and no one pays it. If you reporters want to ask Obama an awkward question, ask him whether he paid the exit tax.

All taxes on Bb are really high, which is why most people there feel they need to work infinitely long days without weekends. That's probably why Obama feels that Americans don't have it so bad tax-wise. We really don't. Bbans do have really good government mandated health insurance though. The excellent health system helps people to keep working those infinitely long workdays for an infinitely long lifetime, which ensures that they'll keep paying taxes and keep the whole thing going. Do the math.

So, Donald Trump, get a clue. Don't go making wild accusations about things you don't know nothing about.

NSA Slide Shows

The latest revelation from Snowden is that NSA and British intelligence stole the encryption keys for billions of cell phone SIM cards, allowing them to listen in on conversations without the assistance of the telephone companies involved merely be setting up radio receiving equipment near where the cell phone is being used.

That fact that NSA did this is not too surprising. It's safe to assume that spy agencies do everything you can think of and a lot that you can't think of. What strikes me, though, about this and the other Snowden revelations, is the form in which this information appeared within NSA and which was stolen by Snowden.  It was a "slide show", that is, a PowerPoint presentation.

I remember somebody telling me once, I don't remember who or when, maybe my parents, that if you tell someone something, it's not a secret any more. A slide show is of course specifically created to inform a group of people, presumably people who don't have much knowledge of what is going on to begin with. My point is that, if you tell too many people about it, it's not going to stay secret for very long. It seems that NSA hires, I don't know, 2,000 people maybe? 20,000? And then swears them all to secrecy and tells them everything and expects that everything will stay secret. The truly shocking thing about Snowden's revelations is that NSA seems to be spreading around a lot of information about its operations to people who don't really need it to do their jobs. It's problematic to understand the motivation for this. It may just be braggadocio. It may be to justify funding or advance careers.

In spook films like "Mission: Impossible", spy secrets are kept in vaults with exotic electronic defenses. In reality, it seems that the nation's secrets are scattered haphazardly all over God's creation. If you need more evidence of this, consider the leak to a news reporter of CIA operative Valerie Plame's identity by Richard Armitage in 2003. As far as I can tell, there was no reason for Richard Armitage to know that Valerie Plame was a CIA agent. It was never explained how he knew that in the first place. Yes, he was a Deputy Secretary of State, but even so he would not need to know the identity of secret agents and shouldn't know it. The fact that this question never even came up in the scandal and grand jury investigation tells you that this is business as usual.

Then there's the question of how Snowden himself had access to all this. Well, OK, he was a system administrator. But it's not that hard to encrypt files to keep system administrators from being able to read stuff they're not supposed to read. I could do it using simple tools and I'm not much of a security expert. NSA on the other hand are supposed to be encryption experts. I can only conclude that our security agencies are bloated bureaucracies that are not very focused on their mission.

Some forty years ago I expressed my dismay to a friend about how inept U.S. intelligence agencies seemed to be. He expressed the thought that they might put out disinformation to make everyone believe they were dumb so that our enemies would underestimate them.

I wish it were true.