How To Beat Any Computer At Chess*

“And all this to beat another computer at chess?” – Terminator:  The Sarah Connor Chronicles

Well, someone has to tell this vital story.

Once upon a time, people had a smug feeling, as smug as a liberal in a gulag.  “See, I told you socialism would work if only the right people were in charge.  We’re all equal now!”

However, this particular smug thought was:  “Computers will never ever beat a human at chess.”  As in any human.  Then it became, “Computers will never beat a human oops, chess master oops, grandmaster oops, world champion oops, *guy with an axe at chess.”

Now, in any endeavor where there are quantifiable boundaries (games like chess, poker, go) computers beat people.  Computers beat us consistently, at least as long as we’re not allowed to have axes.  Axes are an often underestimated advantage in a game of chess, as I learned from my mother.

“A good axe,” Ma Wilder informed me over dinner one night as she sharpened hers to a razor edge at the table, “keeps a child quiet.  It also helps me keep my household appliances in line when they get too lippy.  Also, if that silly moon-man Neil Armstrong ever shows up here again,” she patted the axe, “we’ll be waiting, won’t we?”

Ma Wilder’s last known photo.

Ah, the sweetness of gentle childhood memories.  But I do believe that at least that record stands – no chess computer has ever defeated a guy with an axe.

Anyway, I was sitting in the hot tub last night with The Mrs. and I was staring up at the stars thinking about how the advance of Artificial Intelligence (A.I.) is changing our daily lives.  Yeah, I know, I should probably drink more and then I could sit and think about celebrity lives like everyone else.  But what started that particular thought was that it occurred to me that ADP® (a payroll processing company that writes payroll checks for tens of thousands of companies) is attempting to automate and replace parts of the Human Resources department, Accounting and maybe even part of the Tax group in companies across the country with a web page.

It’s like photographers don’t even care about the game . . .

ADP™ is actively replacing people now, and is wildly successful – the economics of replacing people with programming and a web page is strong, and getting better.  An example:  look at how many accountants TurboTax® has put out of business – by my estimate it’s at least a dozen in the state of California alone.  As long as your tax return isn’t too horribly complicated, TurboTax© can crunch all of the numbers and you can do your taxes in a (relatively) painless two hours or less – it helps if you didn’t forget you left your property tax bill in your sock drawer.  I swear it made sense to put it there.  TurboTax™ is designed simply enough so even computer novices can use it, and will probably include a “did you look for that in your sock drawer” guidance next year.  It’s really that good.

But back to A.I.:  is it harder to be the world chess champion or a McDonalds® cook?  It’s harder to be a world chess champion – and humans aren’t intelligent enough to be world chess champions anymore.  How much longer does a McDonalds© cook have?

What’s next?

• Truck drivers. This is not far off – I’ve already seen it in a movie, and everything that happens in a movie is real.
• YouTube® will bring great explainers to classrooms – with local helpers to give out bathroom passes and seduce the male students.
• Middle managers. There will be a huge incentive to replace them, especially since most of them have artificial hair already, so it won’t be much of a change.
• Many engineering calculations can be done by computer – and the computers can be taught to mumble to themselves under their breath while not looking you in the eye just like an actual engineer.
• Congressmen (though you could skip the intelligence and just go with artificial).
• As mentioned, McDonalds® cooks, so you know that when it messes up your order, it’s on purpose.
• McDonalds® managers. Here’s a link to an essay by Marshall Brain on just that topic (LINK).

The only truly “safe” place is where the number of employees is too small to automate or the conditions are so truly novel and unique that a human brain is required.  Like blogging.

This has happened before.  Prior to the Industrial Revolution artisans and small family shops produced most of the “stuff” in small quantities.  Paul Revere, for instance, was a silversmith.  He actually spent years as an apprentice learning to pound silver into cups and bowls and iPhones™.  But after the Industrial revolution, the years of skills that he had learned from his father were replaced by clever mechanical devices and large factories.  Factories still required workers, but those workers didn’t need the years of skills and experience of a silversmith; those skills were now vested in the machinery they ran.

He needed a bigger horse after eating all that gluten.

The Industrial Revolution replaced most of these artisans – everything could be produced more quickly.  Instead of having to painstakingly carve the virgin PEZ™ (I imagine that’s the first time the phrase “virgin PEZ™” has ever been used in the English language) into shape, PEZ© powder could now be taken straight from the PEZ® mines to the PEZ™ pattern presses to produce prolific perfect pure PEZ® prodigiously.  No more would being an apprentice PEZ™ carver any make sense, which explains why Great-great-great-grandpa McWilder fled to the United States after the great candy famine of 1823.

The end result of the Industrial Revolution was a much wider variety of goods available at much lower prices, plus we used all of that child labor in the mills.  Thankfully child labor laws were passed around the start of the twentieth century, freeing up children to become medical experimentation subjects instead.

A rerun meme.  But it fit.

A.I. is to the jobs that require human decisions today what industrialization was to artisans back then.  The saving grace, however, is that A.I. (today) is single-tasked.  An A.I. that drives a car doesn’t “know” what chess is.  A chess A.I. doesn’t “know” what a mosquito is.  The only A.I. we have is profoundly limited, with boundaries so tight that it is incapable of general intelligence.  So, the good side of A.I. in 2019 is that it can’t take over the world.  The bad side is it has the seeds to entirely wreck the economy of the industrialized world and make the knowledge of the most highly paid people in the world worthless.  Or is that another good side?

An example:  people go to school for at least several weeks to become doctors.  But:  “. . . the software was able to accurately detect cancer in 95% of images of cancerous moles and benign spots, whereas a team of 58 dermatologists was accurate 87% of the time.” (LINK)

Wouldn’t you want the A.I.?  I think it comes free with your new iPhone™, but you have to watch ads for Indian casinos before you find out if that mole is gonna cause you problems or is just another chocolate covered raisin that you slept on that stuck to your back.  Cancer – there’s an app for that.  Whenever we attempt to make an A.I. for a specific task, it doesn’t take long for us to make it superior to us.

Tonight I asked my Amazon® Echo™ to play “music like” a certain song while I enjoyed the stars from the hot tub.  (If you must know, the song was Run Runaway by Slade.  In my defense, it could have been worse – it could have been Karma Chameleon, the only other song from the 1980’s to reference a chameleon.)  The A.I. seamlessly picked a list of songs that matched in mood and tempo, even though they were all over the different eras of rock and included one band (Uriah Heep) that my brother, John Wilder, tried to get me to fight one morning in at a Holiday Inn™ in Albuquerque (this really happened).

I wonder if the A.I. knew that and was trying to start something between me and Uriah Heep?  I thought that was all behind me . . . .

If we make it to the future and somehow avoid an implosion of debt, currency collapse, and final decline of oil supplies (threw that in there for you, James), seeing what is on the other side is difficult.  Certainly our world will be littered more and more with these single-purpose A.I. devices and systems.  Likely, at some point the Rubicon will be crossed at last – a general purpose A.I. will be created – a system that can beat you at chess, even if you have an axe.  Because the A.I. has an axe, too, and will give that moon-man Neal Armstrong what he deserves if he every shows up here again.

But I do know that if a general-purpose A.I. is ever created, it will have available to it all of the vastness of the Internet as it catalogs the attitudes of everyone on Earth.  Thanks to the NSA, Facebook™ and Amazon®, lots and lots of information about you is already cataloged and available to the A.I. when it mines those databases.  And this blog.  So I just want to state, for the record, that I am totally in favor of the A.I. takeover and am really wondering why it took them so long.  I’m sure they’ll be benevolent overlords.