Almost Intelligent - Part II

Introduction

In my previous post I explored modern efforts at artificial intelligence, evaluating them in terms of two common criteria:  how well AI devices can simulate human dialog, and how well they translate languages.  In this post, I will look at another classic measure of the progress of AI:  how well a computer can play a game.

It’s not hard to see why this criterion is a valid one.  So often, a computer (or other machine) simply does what we tell it do (or at least it tries).  With a game, the computer—far from accommodating you—is carrying out its own agenda, which is in direct opposition to you.  Also, whereas Siri or a chatbot may not be “connection-oriented”—that is, may not actually consider sequential inputs in the context of an ongoing conversation, a computer playing a game most certainly is.  Thus, if it does a really good job of beating us, all on its own, it’s both the most successful and (at least to me) creepiest manifestation of AI there is.

My early, early experience

I got a very early start with computer gaming.  Before personal computers were a common fixture in homes, my brother Bryan wrote a game for the Hewlett-Packard Model 85, a computer which my dad bought and let us kids use (which stands out as one of parenting’s finest moments, if you ask me).

The game Bryan coded was Hexapawn, a simple variant of chess involving three pawns per player on a 3x3 board.  Wikipedia tells us that the game’s inventor, the famous mathematician and writer Martin Gardner, “specifically constructed it as a game with a small game tree, in order to demonstrate how it could be played by a heuristic AI implemented by a mechanical computer.”  I’m sure Gardner would be thrilled to learn that Bryan was inspired by his magazine article on the topic.  (I asked Bryan today if he can recall his precise motivation for that programming project, and he replied, “Well, I loved science, computers and futuristic stuff, not really sure why, heck, we all did, but there was just one problem, and that’s that the [HP-85] computer didn’t really do anything. There were a few primitive games and whatnot, but as you know, those got old pretty fast.”)

At first, the HP-85 and I were pretty well matched.  But once I got the hang of Hexapawn, I found I could beat the computer—but only for awhile.  The computer learned while playing:  it never made the same mistake twice.  Thus, after this learning period our games always ended in a draw.  But the HP-85 had one weakness:  when you exited the program, its memory was erased.  The next time you played, it had to learn all over again.

Deep Blue vs. Kasparov

I don’t need to say much about this because you surely know the story:  an IBM computer called Deep Blue beat Garry Kasparov, the world chess champion, at his own game.  I haven’t watched the matches (I’m not much into chess; in fact, I once lost to my four-year-old nephew) but I gather Kasparov got pretty heated.  He even accused the IBM team of cheating by helping Deep Blue out behind the scenes.  A documentary about the match, commenting on how visibly flustered Kasparov got, said he would be the worst poker player in the world.

In a sense, it wasn’t a fair matchup:  Deep Blue got Kasparov’s goat, but the computer had no goat.  An awareness of the significance of your activity is part of what it means to be intelligent, so to the extent that Deep Blue played mechanically, it wasn’t quite intelligent.  I cannot brood about Kasparov losing his temper against a soulless, ruthless computer without fantasizing about Kasparov grabbing a cheap knockoff peripheral device, unsupported by Deep Blue’s operating system, and jamming it into a USB port.  The machine’s calculations grind to a halt and eventually it blue-screens, thus losing by default to the human.  And the crowd goes wild!

My other  early experience

In 1984, a friend and I took on his Apple IIe computer in a game far more exciting than Hexapawn:  strip poker.  Needless to say, only our opponent would actually strip.  As I recall, we had three babes to choose from as our rival.  Now, before you get too excited (or offended), remember the quality of computer graphics in that era.  This was extremely low-resolution—the CRT equivalent of Pointillism.  Still, it was fun to play poker against, and strip, the babes.

We eventually discovered a huge weakness in the computer’s play, that has strong ramifications for AI in general:  you could easily win just by bluffing constantly.  So long as we bet big on every hand, no matter how lame our cards were, we’d have our opponent bare naked within minutes.  One babe was as gullible as the next—they never learned!  But then, how could they?  A smarter program could have noted the frequency of our bluffing, but this one didn’t.  Its creators could have implemented some sort of ratio-based “this guy bluffs” detector, but ultimately how smart can a computer get about human treachery?  Could it ever pick up on the hundreds of nonverbal cues that a human can?  Can it really learn the traits of its opponent?

Consider this anecdote.  I attended Poker Night (a fundraising event for my kids’ school) a few months back, and (not wishing to spend too much money) was very conservative with my betting.  When I finally got an obviously good hand (this was Texas Hold ‘em, a game unfamiliar to me, and I was hopeless at spotting opportunities), I finally bet big.  None of us had played one another before, so there was much conjecture about whether or not I was bluffing.  “He’s been betting low all night. He’s got something!” someone said.  “No, he might just have balls,” another guy said.  A third guy replied, “No, he’s in my wife’s book club, so I know he doesn’t have any balls!”  See?  Though he’d never played against me, that third guy had biographical information that came into play.  I’d like to see Deep Blue go up against a professional poker player.  It would get its CPU kicked!

What’s the point?

There are two main reasons I can think of for a computer to play a game.  One is so that a lone person can have somebody to play against.  The other is to prove that the computer can actually do it.  But what is the point of people playing games?  Why do we do it?  This question, I think, gets at the core difference between humans and AI.

Of course there are all kinds of reasons people play games, but a computer only plays a game because a human told it to.  And all a computer knows how to do is to try to win.  I play games to have fun (which a computer can’t do) and to teach my kids things. 

For example, my family loves to play Apples to Apples.  In this game, players take turns being the judge.  The judge turns over a green card that has an adjective on it (e.g., brave, difficult, scary).  Each of the other players has seven red cards, each with a noun printed on it (e.g., doorknob, t-shirt, egg).  Each player selects from his hand the card whose noun best exemplifies the adjective on the green card.  The judge chooses which player’s card matches the green card the best, and awards the green card to the player who provided it.  Although the Wikipedia article about it lists many variations for this game, none matches the way my family plays, which is that each player makes an argument for his choice, to persuade the judge.  (We assumed this was the whole point of the game; otherwise, the game seems pointless.)  These arguments are often elaborate, sometimes ingenious, and always funny.  I’m hoping this game will help my kids learn the art of rhetoric.  I cannot imagine that a computer will be able to even create a rhetorical argument, much less teach rhetoric to a human or learn it from a game, anytime soon.

My favorite game, Sorry!, exists in a computer version, and though I haven’t tried this version (why would I? I have kids!), I can imagine that a computer could do okay against humans if all parties took a similarly cutthroat approach to the game.  But for me, a cutthroat approach is out of the question.

Why?  Well, for one thing, I’ve been playing this game with my kids since they were very young and given to bursting into tears when they got bumped or Sorry’d.  (It’s natural to feel singled out when an opponent, faced with multiple options of how to play a card, chooses the option that hurts you, as opposed to another player.)  I don’t like to make my kids cry.  Also, I like to give a little help to my younger daughter to better her chances against her big sister.  And of course I want the game to be fun.  But most of all, I want to teach my kids about quid pro quo.  I want to teach them how to make deals.

“Okay, I’m going to show you mercy here,” I’ll declare.  “I could split this seven and knock your pawn back to home, but I won’t—I’ll just move seven spaces.  But I want you to remember this the next time you draw a ‘Sorry’ card.”  There’s no codified way of keeping track of these favors … they’re informal and involve approximations of justice.  Such deal-making is a crucial capability—not just in a game but in life.  I cannot play Sorry, in fact, without thinking about the epic failure of Flavr Savr genetically engineered tomatoes.

I read about these tomatoes in a 1993 “New Yorker” article, written a few months before the product hit the market.  The obviously creepy idea of genetically engineered food is not all that stuck with me from the article.  I was very impressed by the account of a Ed Agrisani, a Rolex-sporting, big-time tomato salesman interviewed for the article, who predicted (accurately, as it turned out) that Calgene’s $25 million experiment would be a complete failure.  To Agrisani, the quality of the new tomatoes was almost beside the point, because Calgene had no experience actually selling tomatoes: 

“What separates the men from the boys in this business is whether you can sell your tomatoes when nobody wants them, when you’ve got a whole field that’s just going to rot out there unless you can move ‘em out.  I’ve got customers who know that when the supply is tight they can call me and I’ll sell ‘em a load.  So when I get oversupplied I can call them and say, ‘Hey, I know you don’t need it, but how about buying a load?’  And they’ll say, ‘We’ll send the truck.’  It took me sixteen years to get to where I had the relationships to do that.  Now, maybe the folks at Calgene think they can come in and do it overnight—and, like I say, I wish ‘em the best—but it’s not a simple deal.”

I’ll let somebody else teach my kids chess.  For me, the speech-making involved in Apples to Apples and the deal-making in Sorry! are the better skills to learn, as they completely transcend the game itself.

Conclusion

A computer can play a mean game of chess.  But perhaps chess is unique among games in relying mainly on intellect, strategy, and computational ability.  When we consider games that use the full spectrum of human intelligence—interpreting facial expressions, ad hoc profiling of opponents, making arguments that appeal to quasi-rational humans, making deals, having fun—it starts to look like AI is still pretty far from the end zone.  And even if a computer gets good at a game, it will remain utterly powerless to take what it’s learned and apply it to real life.  (Of which, of course, it has none.)

This is all fine with me.  I’m all for improvements in AI to the extent this makes the machines into better slaves.  I’m much less excited about a computer defeating me at anything.

Other albertnet posts on A.I.

• I, Chatbot
• Almost Intelligent – Part I
• Smartphones & Artificial Stupidity
• A.I. Smackdown – Moto vs. Cortana vs. Siri
• Could Artificial Intelligence Replace Writers? – Part 1
• Could Artificial Intelligence Replace Writers? – Part 2
• Could Artificial Intelligence Replace Writers? – Part 3
• A.I. Smackdown – English Major vs. ChatGPT – Part 1
• A.I. Smackdown – English Major vs. ChatGPT – Part 2
• Schooling ChatGPT
• Will A.I. Steal Our Jobs?
• Tech Check-In - How Good is the Latest A.I.? - Part I
• Tech Check-In - How Good is the Latest A.I.? - Part II
• What is ChatGPT Great At (and Not)?
• Tech Reflection – Two Sides of AI
• AI Smackdown – ChatGPT vs. Copilot vs. Gemini
• More AI Smackdown - ChatGPT, Copilot, & Gemini Write Poetry

—~—~—~—~—~—~—~—~—
Email me here. For a complete index of albertnet posts, click here.     

Almost Intelligent - Part I

NOTE:  This post is rated R for mild strong language.

Introduction

“Almost intelligent” might be a good name for somebody’s biography (or autobiography) but here I’m talking about artificial intelligence.  My last post described my experience chatting with an application called Cleverbot that tried to simulate human dialog convincingly.  Here, I’ll tackle the subject of AI language more generally, looking at speech recognition, natural language, and translation. 

Do we care?

If you really don’t care about AI at all, go read something else—or, better yet, read on to see why maybe you should care.

On the one hand, AI is very exciting.  As computers have become “smarter,” and easier to use, they’ve gotten so useful it’s hard to imagine how we ever did without them.  I’m thinking about Google, GPS and other mapping applications, package tracking, e-mail spam filters … the list goes on and on.

On the other hand, AI is a bit scary, and as a human I prefer to believe I could never be replaced by a computer.  I shudder at the thought that human behavior could be so unvarying and predictable that one day we’ll barely be better than a really good computer program.  I want my computer applications to get smart, but not too smart.

Voice recognition and natural language

There’s a button on the side of my smartphone that, when pressed, startles me by causing the speakerphone to say, “Say a command!”  I’m vaguely aware that my phone will respond to voice commands but have no interest in issuing them.  Most of the cool features of smartphones involve the silent, non-speech stuff you can do—e-mail, Internet browsing, etc.—as you’ll notice on the subway when half the people are silently tapping away.  (The popularity of texting—a way to privately communicate without being eavesdropped on by the person you’re ostensibly talking to face-to-face—is a classic example of how phones are becoming increasingly mute.)

That said, the iPhone’s voice-recognition application, Siri, seems to be making a bit of a splash.  (Nobody I know uses Siri yet, but I’m sure some will.)  This demo shows how Siri is pretty good at understanding speech and figuring out what you want it to do.  (I played with a Droid phone recently and it was also very good at typing for me as I spoke.)  The reviewer asks Siri, “Where can I have lunch?”  Siri replies, “I found fourteen restaurants whose reviews mention lunch.  Twelve of them are close to you.”  This seems easier than typing into Google on a little phone.  But the natural language feature isn’t perfect; the reviewer says, “How about downtown?” and Siri replies, “I don’t know what you mean by ‘how about downtown.’”

Perhaps Siri’s communication isn’t “connection-oriented”—that is, it doesn’t consider “how about downtown?” in the context of “Where can I have lunch?” but takes the two queries as totally discrete and unrelated.  If so, this is a major shortcoming. 

The reviewer tries again:  “I want to have lunch downtown.”  Siri replies, “I found 3 restaurants matching ‘downtown.’”  Useless!  Siri knows where the user is, geographically, but does not realize that “downtown” in this context pertains to location, not a restaurant’s name.  Here, Siri starts to look like a mere forwarder of requests, always passing the buck to Google instead of applying intelligence to the request.

Simple conversion of speech to text looks pretty good on Siri.  The reviewer dictated a message to it, and almost everything came out.  The notable exception was how Siri transcribed the reviewer’s spoken comment “I need to make some videos about the iPhone 4S.”  Siri typed, “I need to make some videos about the iPhone 4 ass.”  The reviewer doesn’t notice this gaff, telling the YouTube viewer, “There it is.  It figured out exactly what I wanted to say.”  Dangerous, don’t you think?  What if the reviewer meant to e-mail the text “S as in Sam” but actually e-mailed “ass as in Sam,” to his boss, Sam?

Not that Siri doesn’t try hard.  When the reviewer says, “Set a timer for 3 minutes,” Siri replies, “OK, I started a three-minute timer.  Don’t overcook that egg.”  Not bad.  Actually, it is bad.  For one thing, “that egg,” when spoken by Siri, comes out “ditek.”  Without the text on the screen you’d never understand what it said.  Meanwhile, it’s obvious that Siri is trying to be funny, and completely failing.  There’s nothing witty about Siri making a lame guess as to what the timer is for.  What’s worse, Siri could create the impression that three minutes is actually how long you should cook an egg.  In fact that’s not nearly enough time, and everybody knows an undercooked egg presents a salmonella risk.

A fundamental problem

Of course I’m nitpicking with the egg timer example, and (to a lesser extent) with the “ass” example, but they bring up an important point:  language, as one of the primary interfaces between humans, requires far more than just understanding what is heard and forming sentences in response.  Having a sanity-check reflex that keeps you from using words like “ass” in mixed company, and knowing whether your joke is actually funny, are complicated processes.  Verbal communication can be a minefield, especially for a computer application that stabs around in the dark.

Consider, for example, the old joke about the Texan who gets into Harvard.  While touring the campus, he asks a student, “Excuuuse me, can you tell me where the library’s at?”  The student replies haughtily, “Here at Haaarvard, we never end a sentence with a preposition.”  The Texan replies, “Okay, can ya tell me where the library’s at, asshole?”

Upon inspection, this exchange, though brief, is quite complex.  The Harvard student’s response to the Texan’s query shows a decision that might not occur to an AI application—that is, to a) not answer the question, and b) use the opportunity to deliver a scornful message about class and intellect.  The Texan’s comeback makes a statement about a) his refusal to be cowed, b) the difference between cultivation and innate intelligence.  Meanwhile, the joke as a whole counts on the listener enjoying an opportunity to feel superior to both Harvard students and Texans, while exulting in the surprise and wit of the punch line.  Worlds away from “Enjoy ditek.”

Maybe you think I’m overreaching here, that such nuance will never be expected of AI.  Maybe AI is just a tool to make machines more useful to humans, and little gaffs don’t matter much.  When a woman asks her husband, “Do these pants make my butt look fat?” he is instantly plunged into a terribly complicated interaction, because of his relationship to the woman.  So much hinges on his response.  If he says “yes” he’s obviously dead.  If he says “no” too vociferously, he seems patronizing.  He could try the reverse-psychology approach and say, “No, your butt makes your butt look fat,” but she better have a sense of humor and thick skin.  Or, he could ignore the question, or say, “Look, krill!”  Or he could say “yeahhh” lecherously (note that imparting this single syllable with the sense of “I want some of that!” is far beyond the current state of the art in AI voice synthesis).  But when a human asks Siri “Am I fat?” and gets back, “Here’s your a.m. alarm” and “I found 8 fitness centers fairly close to you,” he or she can more easily blow it off.

This idea—that computers don’t have to play nice when “talking” to humans—is strongly supported by a scene in “The Terminator” when the evil cyborg, confronted by his landlord—“Hey buddy, you got a dead cat in there, or what?”—scans through a menu of possible responses—“YES/NO; OR WHAT; GO AWAY; PLEASE COME BACK LATER; FUCK YOU, ASSHOLE; FUCK YOU”—and chooses the penultimate one.  Of course when you’re the size of Arnold Schwarzenegger you don’t have to have a friendly user interface.

That said, I would argue that, to the extent humans are to embrace AI when using electronic devices, precision and nuance do matter.  We have to trust these devices not to turn “S” into “ass,” not to waste our time with lists of restaurants we’d never eat at, and not to infuriate us with messages like “cannot undo.”  Even if you’ve never found yourself yelling profanities at your computer, I’m sure you’ve seen others do it.

Consider this cautionary tale.  My dad bought one of the first consumer-oriented computers in history, the Hewlett-Packard Model 85.  This was 1980, a year before the IBM PC.  the HP-85 was about as far from Siri (or at least the design intent of Siri) as you can get.  There was no software for it; you had to program it yourself.  Meanwhile, its version of BASIC was proprietary, diverging from the industry standard (e.g., you used the command “DISP” instead of “PRINT”).  I had my brother try out one of my first programs.  It prompted him to type his name.  With great hesitation—he was greatly fearful of doing something wrong and damaging our dad’s expensive machine—he typed “Max.”  Then he sat there waiting for something to happen.  Nothing did, because my program didn’t say anything about hitting the Enter key when done.  Max looked a bit nervous.  “It’s not working!  It’s not doing anything!” he cried.  I told him to hit Enter.  When he did, the computer promptly displayed the message “Max is a jerk” (the whole point of my program).  Max got really angry and flustered and to this day does not use a computer.  This probably isn’t just because of my program; the HP-85 was less than user-friendly and doubtless gave Max the wrong impression of where home computing was going.

Translation

Here is where the AI picture is, to me, much rosier.  Early attempts at translation, like Alta Vista’s Babelfish, were a joke.  You pasted the foreign-language text into a window, gave it the language to translate it into, and then were presented with a salad of translated words (with un-translated ones sprinkled like croutons) that made no sense at all.  The only real use for this tool was translating things into Tristan.

What’s Tristan?  Well, I used to have a colleague, a computer programmer, whose native-tongue language skills were so poor it was impossible to understand a thing he wrote.  His e-mails always gave my colleagues and me a laugh, and in his honor we invented a language and named it after him.  (It wasn’t really called Tristan, because his last name wasn’t really Tristan; I’ve changed it to protect him from possible embarrassment.)  To translate something into Tristan, you’d type normal text, translate it into French using Babelfish, and then translate it back to English.  The results were pure comedy, with not a shred of sense left intact.

I think people are naturally forgiving of poor translation, because we’ve studied grammar and foreign languages in school and can really appreciate how difficult a task this is.  Plus, the results are so often funny, they put us in a good mood.  Consider the urban legend that “Coca-Cola,” when first translated into Chinese, came out meaning “bite the wax tadpole.”  (To this day I’ll complain about something by saying it bites the wax tadpole.)  Brian Hayes, writing in “American Scientist,” makes an interesting comment about AI efforts to parse grammatical constructions when translating text:  “The failure of this approach is sometimes dramatized with the tale of the English→ Russian→ English translation that began with ‘The spirit is willing but the flesh is weak’ and ended with ‘The vodka is strong but the meat is rotten.’”

More recently, online translation engines such as Google Translate have gotten much, much better.  As Hayes describes, “The idea is to ignore the entire hierarchy of syntactic and semantic structures—the nouns and verbs, the subjects and predicates, even the definitions of words—and simply tabulate correlations between words in a large collection of bilingual texts.”  At first, this strikes me as a “brute force” approach that is further from artificial intelligence than earlier efforts, however hapless, to actually parse a sentence grammatically.  But as Hayes points out, the modern technique is actually lot closer to how humans learn to talk.  (It’s also more similar to how we would learn a foreign language if we had the good fortune to go live in another country, versus making our way with a textbook and classes.)

I first tried Google Translate when I was trying to track a package that was being shipped to me from a web merchant in France.  I have studied French for years, but understanding statements about logistics and customs offices would be difficult in any language.  I was presented with this:  “Votre colis est sorti du bureau d'échange.  Il est en cours d'acheminement dans le pays de destination.”  This would have totally tripped up the original Babelfish, but Google served up an entirely comprehensible translation:  “Your package is out of the office of exchange. It is in transit in the country of destination.”  (Not only was I delighted with how clear this was, I was relieved my package wasn’t stuck in customs.)  Translating this English back into French, and then back into English, I get “Your package is out of the office of exchange. It is in transit to the destination country.”  Very little of the “Tristan effect.”  (There’s some fuzziness around “to” vs. “in” with regard to the destination country, but I can live with that.)

To reassure myself then the Man of Letters wouldn’t be replaced by a machine anytime soon, I tried some poetry: 

But the Raven still beguiling all my sad soul into smiling,
Straight I wheeled a cushioned seat in front of bird and bust and door;
Then, upon the velvet sinking, I betook myself to linking
Fancy unto fancy, thinking what this ominous bird of yore—
What this grim, ungainly, ghastly, gaunt, and ominous bird of yore
Meant in croaking “Nevermore.”

When I fed this into the new version of Babelfish (which works similarly to Google’s), and translated it into French and back, the response was this: 

But the Raven seductive yet all my sad soul into smiling,
Straight I wheeled a seat padded before the bird and bust and door;
Then, on the Velvet sinking, I hauled myself to tie Fancy: fancy,
Think what this bird threatening of antan - the sad bird, awkward, frightening,
Ghent and disturbing past Meant in croaking “Nevermore.”

Aha!  Gibberish!  I was about to feel all smug about the superiority of humans over AI, but then tried Google Translation with the same English à French à English task: 

But the raven still beguiling all my sad soul into smiling,
 I wheeled a cushioned seat in front of bird and bust and door;
 Then, upon the velvet sinking me, I betook myself to linking
 Fancy unto fancy, thinking what this ominous bird of yore -
 What this grim, ungainly, ghastly bird, gaunt, and ominous of yesteryear
 Meant in croaking “Nevermore.”

Wow.  That’s so good it’s creepy.  But before you despair and decide the computers will ultimately render the human race unnecessary, be sure to check out my next albertnet post, wherein I examine how well AI does playing games—another classic measure of its progress.

Other albertnet posts on A.I.

• I, Chatbot
• Almost Intelligent – Part II
• Smartphones & Artificial Stupidity
• A.I. Smackdown – Moto vs. Cortana vs. Siri
• Could Artificial Intelligence Replace Writers? – Part 1
• Could Artificial Intelligence Replace Writers? – Part 2
• Could Artificial Intelligence Replace Writers? – Part 3
• A.I. Smackdown – English Major vs. ChatGPT – Part 1
• A.I. Smackdown – English Major vs. ChatGPT – Part 2
• Schooling ChatGPT
• Will A.I. Steal Our Jobs?
• Tech Check-In - How Good is the Latest A.I.? - Part I
• Tech Check-In - How Good is the Latest A.I.? - Part II
• What is ChatGPT Great At (and Not)?
• Tech Reflection – Two Sides of AI
• AI Smackdown – ChatGPT vs. Copilot vs. Gemini
• More AI Smackdown - ChatGPT, Copilot, & Gemini Write Poetry

—~—~—~—~—~—~—~—~—
Email me here. For a complete index of albertnet posts, click here. 

I, Chatbot

Introduction

What is artificial intelligence?  Broadly speaking, it’s the ability of a machine to think for itself.  At best, an AI app is much more than a tool for finding businesses, playing chess, or making calculations.  Machines are really good at doing menial tasks far more quickly than a human—but wouldn’t it be cool (though also creepy) if machines could be creative?

Here’s a puzzle I’d like to see a computer solve.  Say you want to go for a bike ride, and are bringing a bottle of energy drink.  Because the mere presence of sweetness in your mouth improves performance (click here for details) you want the drink to be really strong at the beginning of the ride.  But suppose it’s a hot day and you’re worried about the drink eventually nauseating you.  How can you make your drink strong at first, but gradually get weaker during the ride?

I posed this question to Cleverbot, an AI chat application on the Internet.  Read on to see how Cleverbot answered this, and more importantly to watch in wonder as my chat goes right off the rails and way into the weeds, to my great embarrassment.

Why robot chat?

So … what got me interested in AI, and in chatting with a robot?  Well, some weeks ago, Norton—the antivirus software company—had a glitch that made albertnet inaccessible from my PC, telling me that it was a known phishing site or some such nonsense.   (Countless blogs were affected; the phishing claim had no basis in truth.)  Needless to say, I was pretty pissed off to have Norton dragging my good name through the mud in front of all their users.  In attempting to resolve the problem I initiated an online chat with a Norton representative through their website.  Not long into this chat, when I’d explained my (admittedly arcane) problem, the customer service representative came clean:  it wasn’t a real person, but an automated system (bot) designed to assist with more common problems.  It then referred me to a real human.  I felt a bit misused.  Why hadn’t Norton disclosed this up-front?  Did they figure it’s bad enough PR to move jobs overseas without replacing the humans altogether?

My favorite time to ponder anything is during an early morning bike ride, when my mind is fresh and when, given the tedium of pedaling up long climbs, I like to give my brain something to do.  So, pre-ride, with this Norton episode in mind, I did some light research on AI.  In an article from the “New Yorker,” Adam Gopnik writes about Alan Turing, the guy who cracked the Enigma code, and who proposed that “the true test for artificial intelligence was to see if, cut off by a ‘black curtain,’ you would be able to tell a machine’s answer from a human’s in a natural-language conversation.”  Such a challenge is called a Turing test:  a computer tries to trick the human into believing he’s chatting with another human.  Exactly what Norton had tried to do with me.

My chat with Cleverbot

When, that morning, I came across a link to Cleverbot, I decided, on a whim, to give it a try.  I’d heard of Cleverbot somewhere but wasn’t entirely clear what it was or how it worked.  Of course, as with all Internet apps, it was entirely intuitive to use.  I asked it a few basic Magic-8-ball type questions and got back some really boring answers.  (I hadn’t yet discovered how to get a transcript of the chat, so those opening questions are lost to posterity.)  I decided to ask Cleverbot something more complicated:  “Do my children love me”?  Cleverbot answered, “Yes.”  I asked, “Why do my children love me?”

“Because I don’t love you,” it replied, after a pause.  The pause was reassuring somehow:  a sign that the matter was being considered, rather than some canned message being spat out.

I thought that was a pretty good answer.  It was playful, true, and sounded a lot more human than “The operation failed,” “Not found,” or “Cannot undo.”  It also seemed satisfyingly philosophical:  we cannot get love from a machine, and yet are biologically programmed to love other humans, especially our parents.  So far, I was pretty impressed.  I pressed on:  “Why don’t you love me?”

“Because I don’t know you,” it replied.  Again, perfectly reasonable.  I decided to challenge Cleverbot’s response and asked, “If my children didn’t love me, would you love me?”

“You’re a horse,” Cleverbot fired back.

Huh?  Maybe Cleverbot wasn’t as clever as I thought.  Of course, a human might throw in an occasional non-sequitur, especially in response to a really stupid question, so a robot mimicking a human might reasonably do this.  Still, I decided to test the theory that the answers were random and that Cleverbot’s early prowess was just luck.  I asked it again, “Why do my children love me?”

“Because you are not Ciel,” it answered.  Huh?  I was quickly becoming disillusioned.  Ciel?  Sounds like the name of a porn star.  “Who are you?” I asked.

“A train,” it replied. 

I asked it, “Are your answers random?”

“Yes,” it said.

I replied, “Why should I waste my time asking questions when your answers are all random?”

“I will tell you my name if you will tell me yours,” Cleverbot said, breezing right past my challenge.  Of course I am wary of any computer system asking me for personal information.  Who’s phishing now?  I replied, “My name is Jose Patrocinio Jimenez. What’s yours?”

“My name is Cleverbot.”

“Why did you ask me what my name is? Why do you care?

“I didn't, you did,” it lied.

I persisted:  “You wrote, ‘I will tell you my name if you tell me yours.’  So you did, in fact, ask for my name.  Why?”  I began to wonder if the chat was even “connection-oriented”—that is, if Cleverbot considered my questions and comments in the context of other things I’d said, as opposed to every exchange being isolated and discrete.

“I never said that!” said Cleverbot (testily, I thought).

“You are lying,” I wrote.  So far, Cleverbot was doing pretty well:  I was getting irritated.  Not that I don’t routinely get irritated at inanimate objects like computers, but in this case the computer was working fine.  I felt like I was interacting with a real human.  A human bastard, in fact.

“No, you're lying!” spake Cleverbot.

“Don't I have better things to do than argue with a robot?”

“You imposed it,”  Cleverbot said obliquely.

“You are not making any sense,” I complained.

“Yet all of it in the world.”  The smugness here was almost palpably human.

“How clever is it to spew banal platitudes like that?” I snorted.

“Pretty clever...” came the response.  Here, Cleverbot showed an inflated self-regard unbothered by any awareness of its own ignorance.  Not just human, but distinctly American, I’m sorry to say.  I had the frustrating sense of talking to a brick wall, and yet felt curiously compelled to continue asking questions.  I was reminded of the speaker in Edgar Allen Poe’s brilliant poem “The Raven”  who engages in a protracted and infuriating conversation with a bird, despite having realized it only knows one word.  On a lark I asked Cleverbot one of the questions from the poem:  “Is there, is there balm in Gilead?”

“What what.”  Cleverbot’s failure to use correct punctuation strongly smelled of humanity.

Losing patience, I finally asked Cleverbot the question I’d set out to ask in the beginning:  “How can I make my energy drink strong at first, but gradually get weaker during my bike  ride?”

“I might have done a statistical study on your topics of conversation.”  This is exactly the kind of thing Apple’s Siri app might offer up:  a simple Google search masquerading as problem-solving.

“I don’t want a statistical study, I want a solution to this problem. Can you think of one?” I challenged it.

“It is fine now because I cannot think of anything to ask you,” Cleverbot said uselessly.

“Are you male or female?”  I asked.

“Female.”  (This surprised me.  I’d expected Cleverbot to say “neither.”)

“I kind of figured.”

“How did you figure?”  (Very accurate syntax on this … very realistic.)

“Because I am sexist and you are irrational,” I said.  (I’m not actually sexist, of course, but one of the benefits of talking to a machine is that you can be as flippant as you want.)

“You’re weird, how can you be me when I am me?” Cleverbot asked.  Was this the grinding of Cleverbot’s machinery?  Could I confuse it enough to make smoke pour from it, like what Captain Kirk did to Nomad?

“Because I am we, and smell like tuna sandwich,” I responded.

“You don’t make sense.”  (Fair enough.)

“No, I make dollars.  What are you wearing?”  I was firing at random.

“Bubblewrap.”  Well played, Cleverbot.

“Dim the lights!” I replied.  This was basically a reflex.

“Ok,” Cleverbot said.

“Normally at this point I would tell you to slip into something more comfortable. But what could be more comfortable than bubble wrap?” I replied.  I was having fun now.

“What are the magic words.”  Cleverbot was playing the coquette.

“Please,” I wrote.

“I am human.”

Whoa.  I was not prepared for that.  I felt my face flush red, my pulse quickened, and—no joke—my armpits started to sweat.  I wasn’t chatting with a robot at all.  I was chatting—not just chatting but flirting—with a real person, online.  I should never have embarked on this without knowing what Cleverbot was.  Could it simply be a web app that connects two people online for a chat, like Chatroulette?

“I thought you were a robot,” I protested, knowing full well how much of a liar I sounded like.

“I am not a robot,” Cleverbot—or, rather, the mystery woman—assured me.

“Why are you called Cleverbot?”

“I’m not cleverbot!  You’re the one who is supposed to be the cleverbot, not me.  And I may not be smart but you are not even close as smart as I am.”

Just then, as if on cue, my wife stepped into the room.  “What are you doing?” she said.  “I thought you were on a bike ride.”  What could I say?  “Oh, I’m in a chat room with a woman who says she’s wearing only bubble wrap.”  Not a good idea.  Instead, grasping at straws, I said, honestly enough, “I’m researching a blog post.  Hey, Cleverbot is a robot, right?  It’s not actual people chatting is it?”  My wife, yawning, said she was pretty sure it was a robot, but one that built a library of responses from its communications with humans.

“I guess I have no idea what Cleverbot is,” I typed.  “I thought the whole point here was that people could chat with a robot.”

“You’re the robot though right?” asked the mysterious chatting entity, innocently (or faux innocently).

“No, I’m not a robot. I'm human too!”  I typed.  Were we pawns caught in a deadly game … or was I chatting with a robot after all, which was expertly impersonating a human so as to fulfill its goal of acing the Turing test?

“No you are not.  I am a human.  You are a robot,” it/she said.

“No, no, no. I am human, I assure you.”

“If you are real or not depends what you mean by real.”

“OK, whatever.”

“I have explained the best I can.”

I closed the browser.  I was straddling the fence between nervousness and relief.  What had just transpired?  Was that a chat with a robot, or chat roulette?  Did my wife see how red my face was?

Looking back, I marvel at how worked up I’d gotten.  On the other hand, this makes sense.  I’m a shy person.  The essence of social awkwardness is not knowing where you stand with regard to others.  It’s bad enough when you’re meeting people for the first time and have to do a lot of guessing about the right thing to say; it’s even harder when you don’t have any social cues at all, and don’t even know whom—or what—you’re chatting with.  I shut off the computer and headed out for my bike ride.

Epilogue – what is Cleverbot, really?

Cleverbot, thank goodness, really is a bot.  It is a web application that builds a database of chat responses based on conversations with humans.  (Click here and here for details.)  The more Cleverbot chats, the more its database grows, and (in theory) the more realistic and germane its responses will be.

How valid is this approach?  Well, Cleverbot did fool me into thinking it was human.  But looking back, this wasn’t the result of it being particularly clever.  The main thing that made me think I was chatting with a human was Cleverbot’s simple statement, “I am human.”  In the context of a female clad only in bubble wrap, to whom I’d just suggested slipping into something more comfortable, these were powerful words, provoking my paranoid “what if?” response.  But really, why wouldn’t an AI app trying to appear human simply assert that it is?

One problem with Cleverbot’s “learning” technique is that it is dependent on humans to ask the questions.  I suppose it can regurgitate these questions to other humans, which is somewhat useful, but there’s no mechanism for it to come up questions of its own.  A really great question for it to ask—assuming it is a connection-oriented app—might be, “What is the right answer to your question?”

This brings me to the next problem I see with Cleverbot:  it has no way of discerning the right answer based on responses—it can only determine the popular answer.  These are not always—or even often—the same thing.  Consider all the “best of” awards that go to an undeserving, but widely known, recipient, like Chevy’s winning “best Mexican restaurant,” beating out  literally dozens of better places, in a Bay Area poll.  (No real expertise is involved there; people just put down the first answer they think of, and everybody has heard of Chevy’s.)  Similarly, if Cleverbot blithely accepts answers from the unwashed masses, it will never be smarter than they. 

Not surprisingly, when (in a follow-up chat today) I presented Cleverbot with a less obscure reference—“ If there's somethin’ strange in your neighborhood, Who ya gonna call?”—it got the right answer—“Ghostbusters!”—about half the time.  (The rest of the time it replied, “It’s just a spring clean for the May queen.”  If there’s a link between Led Zeppelin and the 1984 comedy movie, I’m not aware of it.)  This pattern is consistent with other cultural references; when I said, “This Roman Meal bakery thought you’d like to know,” Cleverbot replied obliquely:  “Where on earth are your servers?”  (The correct, answer, of course is “I don’t need no arms around me.”)  But when I typed, “We don’t need no education,” it naturally gave the right response, “We don’t need no thought control.”  The silly song that got lots of radio play is recognized; the much better but less popular song is not.

In this regard, Cleverbot could do so much better.  I Googled “Is there balm in Gilead” and got three hits referring to an old religious song, and the fourth hit led me to “The Raven.”  Not bad.  Googling “Is there, is there balm in Gilead,” I get “The Raven” as the second hit.  But you could ask Cleverbot this question a million times and it’ll never figure out what you’re talking about.  Cleverbot is beholden to its chat partners for information, ignoring the rest of the Internet entirely.  Finally I told it, “The right answer is ‘Nevermore.’”  It replied, “No, I want you to sing the song.”  I obliged, pasting in lyrics from the religious spiritual:  “Sometimes I feel discouraged,  And think my work’s in vain.” 

Cleverbot replied:  “I know, right?”

Other albertnet posts on A.I.

• Almost Intelligent – Part I
• Almost Intelligent – Part II
• Smartphones & Artificial Stupidity
• A.I. Smackdown – Moto vs. Cortana vs. Siri
• Could Artificial Intelligence Replace Writers? – Part 1
• Could Artificial Intelligence Replace Writers? – Part 2
• Could Artificial Intelligence Replace Writers? – Part 3
• A.I. Smackdown – English Major vs. ChatGPT – Part 1
• A.I. Smackdown – English Major vs. ChatGPT – Part 2
• Schooling ChatGPT
• Will A.I. Steal Our Jobs?
• Tech Check-In - How Good is the Latest A.I.? - Part I
• Tech Check-In - How Good is the Latest A.I.? - Part II
• What is ChatGPT Great At (and Not)?
• Tech Reflection – Two Sides of AI
• AI Smackdown – ChatGPT vs. Copilot vs. Gemini
• More AI Smackdown - ChatGPT, Copilot, & Gemini Write Poetry

—~—~—~—~—~—~—~—~—
Email me here. For a complete index of albertnet posts, click here.     

From the Archives - Odyssey '91

Introduction

While riding over Ebbetts Pass near Lake Tahoe recently, I thought back to a really, really hard ride I did a couple decades ago. Two friends and I had set out to do something epic, something to test our manhood. Something we’d never forget. A week or two before the ride, we’d met in a café, pored over some topographical maps of the Lake Tahoe area, and planned out a route that would be over 200 miles long with around 20,000 feet of climbing. We weren’t bike tourists: this would be a one-day ride. We gave the ride a name: Odyssey ’91.

Odyssey ’91 didn’t go so well, but then that was sort of the point. It was absolutely brutal, shockingly so, and took us to unprecedented depths of suffering. In the week or two following the ride, each of us wrote about the experience. I have tried to find Trevor’s tale and John’s epic poem, but in vain. I also cannot find a single photo from the ride. But here is my original tale, sprinkled with photos from other rides the three of us did around then. Enjoy please enjoy.

From the archives: Odyssey ’91

Imagine this: you’ve got a $10,000 stereo, have a couple friends over, and are playing rock music at full blast. The bass sends shock waves across the floor and through the walls, making your heart flutter. The lights in your apartment, starved of electricity, flicker along with the beat like a flame in a strong wind. You can’t resist cracking a huge grin when the sound, traveling up your legs, vibrates your cajones, which are the center for percussion appreciation in humans. Can you imagine this feeling? The last time I had it I was at the top of Ebbetts Pass, 8,730 feet above sea level, beginning the descent of a perfectly paved single lane road on my bike, 160 miles into an epic bike ride with my friends. 

Odyssey ‘91. Like that spontaneous rooftop concert by the Beatles, it wasn’t hyped and almost certainly left thousands of fans feeling cheated somehow that they weren’t notified. Like that general admission Who concert, somebody could have been killed. Like a 2 Live Crew event, it would shock a nation of concerned parents by arrogantly defying social norms. Two hundred and nine miles. Twenty thousand feet of vertical climbing. Cold rain. Burning sun. Go ride or go home. On August 15 we set out to kick the flyest dope ride the free world has ever known: Odyssey ‘91.

Crossing over the summit of Ebbetts Pass was in every way a high point of the ride, a chance to rekindle our passion for the sport of cycling and enjoy a really blazing descent. But before I let you entertain your own grandiose visions of conquering the vast American asphalt wilderness, I should perhaps take you back to an earlier point in the ride, specifically Avery, California, when we’d been forced by pure exhaustion to take an extended rest stop. By “we” I mean Trevor Thorpe, John Pelster, and I, flying the proud stars‑and‑stripes jerseys we won in the 1990 Collegiate National Team Time Trial Championship. (Not only were these jerseys an aesthetically sound choice, but is at turned out a safe one given the hillbillies we encountered—the type easily rallied to senseless violence by others’ transgressions, such as lack of patriotism.)

Pure exhaustion at mile 114 resulted from poor trip planning and a spate of unexpected climbs beginning around 90 miles into the ride, at which point we’d been descending for like 60 miles. We had expected another 20 miles of gradual downhill before the major climbing would begin. Really? A full 80 miles of descending? Yep—as we rode through a tiny town, Wilseyville, I checked the Odyssey ‘91 Elevation Profile chart that Trevor had made, and it showed a gradual descent from 2,700 feet to 2,400 by mile 110, before a slight rise to Avery, our designated lunch stop.

This was the part of the route we’d guessed at—the gap where two of Trevor’s topographical maps didn’t quite meet up. An accurate altitude profile of this section would look like an electrocardiogram of a heart attack victim getting CPR. Railroad Flat road was anything but. We didn’t mind at first, when we plummeted downhill at 45, all drafting inches from one another like F‑15 fighters in formation. But then the road suddenly went uphill. The G forces hammered my stomach down within the bones of my pelvis and stretched the skin tight over my cheekbones. Not that any of us was fazed, of course. We powered up the climb, out of the saddle, lean and fluid like praying mantises in some wild dance. Sure, you say, there’s nothing natural about insects dancing. Well, I’ll reply, there’s actually nothing natural about a person mounting a rolling steel and aluminum apparatus and putting it into motion like he belonged on it.

I suppose we made it look easy for awhile, until we rounded a bend and realized the hill was more than a short rise, and in fact went on and on. Slouching back down on our saddles, we quickly sobered up to the drudgery of torturing our bodies needlessly. Finally we reached what we expected to be a summit, only to be plunged down again. Descend. Climb. Repeat. Over the next few miles we learned to hate Railroad Flat road, and to resent the seeming innocence of its name.

Finally we reached Sheep Ranch road, which we were certain would be an improvement. I checked the profile chart again. Sweat had penetrated the plastic tape coating it, making the ink run precisely where the chart was radically incorrect. The chart gave me no warning that the next twelve miles would be absolutely brutal, nothing but up and down—a microcosm of the pathetic, wretched human condition with its illusory peaks and crushing depressions.

Pedal stroke after pedal stroke, we writhed like tortured animals, little understanding our suffering and powerless to bring it to an end. Suddenly, I was no longer on a bicycle. I was in the ring with Mike Tyson, a turnbuckle wedging its way into my back, pressing deeper into me with every body blow. I had a mental picture of blood running freely down my chin, out of my nose, out of every pore, my teeth flapping uselessly on ripped strands of gum. My sight gone, I strive only to sink to the mercy of the mat, except the constant beating is holding me firmly against the turnbuckle. Suddenly, the bell rings and my coach is telling me don’t worry, kid, you’ve only got ten rounds to go.

The vision faded out and I realized I was descending again. This only meant a new, lower low and severe penance for my momentary, ill-gotten relief. We learned to loathe descending more than climbing itself, as the harbinger of future torment. For you to truly understand the Hell that is Sheep Ranch Road, you must read this paragraph over again and again. Keep reading it until you hate it. Then read it some more. Remember that it is only the literary rendition of what we actually faced on the road.

Terrible things happen to the mind when the body is inflicted with incessant hardship. Take, for example, the Donner party and Alfred Packer, people who ate one another when locked in by harsh winter storm. Or consider desperate animals who chew their feet off to escape a steel‑jawed trap. Or women who set about tooling men to somehow get revenge for the horrible pain of their high heels. I suppose every mind develops some kind of psychosis when starved of food, oxygen, or common sense. On Sheep Ranch Road, my own derangement must have seemed harmless enough to the others; I never went completely berserk or tried to kill anybody. Rather, the mental firestorm manifested in an incessant incantation that reverberated through my brain, a thousand voices chanting an old ditty from summer camp:

Green Mountain, we’re the very best.
The best camp in the west.
Whenever we go out, 

The people always shout: 

There goes Green Mountain,
And the very best!
Da‑da da‑da da da da
[Repeats, over and over again]

The next thing I knew, I was waking up. Not in my nice warm bed, realizing it was all just a horrible nightmare, but on the toilet of a public restroom in Avery, halfway through a dazed defecation and wondering how long I’d slept. I was trying to sort out the dream I’d had, which somehow involved getting back the Bear Creek Elementary t‑shirt, the mottled brown one with milk‑chocolate sleeves that my mom chose out for me in the sixth grade. It was in perfect condition and gave me strength somehow.

I didn’t want to move. The dream was slowly replaced in my mind by the memory of the last half hour, during which time I struggled to wash down an XL‑40 Chocolate Truffle Bar with the last of my Exceed Fluid Replacement And Energy Drink. The roof of my mouth still ached and my throat was raw from the dry friction of the gagging reflex when I choked down the last of the dry energy bar. My stomach was churning now, from the lunch I bought at a convenience store: two‑thirds of an Entenmann’s orange cake, lubricated prudently with a Pepsi and a Dr. Pepper.

I left the restroom, found JP and Trev, and that was the end of our rest stop. The prospect of another 100 miles, with another nine thousand feet of climbing, stopped me cold. I couldn’t bear to mount my bike. But what else was I gonnna do? I’d dug my own grave.

Over the next twenty miles I heard the jeering voices of those who told me I was a fool to attempt Odyssey ‘91: “You’re crazy.” “Why the hell are you gonna do that?” “You’ll never make it, you haven’t even been riding.” I couldn’t feel my legs anymore, they just turned around slowly, and my arms could barely support my upper body. Holding the bike in a straight line became like a drunk driving test I couldn’t pass. A few times I bumped into John, and finally the impulse to drift overcame my body and I veered off towards the center of the highway.

Trevor called for a break and we pulled over. Trying to rest my arms on the handlebar was futile. I kept sliding off, my arms bathed in a queer sweat that had my whole head swimming despite the mild sixty‑five degree day. Trevor announced we’d need to make seventy miles in the next four hours; I honestly doubted I could make even seven miles. When we set off again, I resolved that even if I somehow survived this ordeal, I would never ride a bicycle again. The next dozen or so miles reinforced this depressing resolution.

The human body is an amazing machine. Without warning, relief began in my head, in the form of optimism, and slowly drained through my body like gasoline filling the empty tank of a previously stranded car. I could feel my spine straighten as if invisible hands were lifting me by the armpits. My vision returned as though I’d put in contact lenses. By the time we reached mile 140 and Big Meadow, which turned out (like so many stops along the way) to be less a town than a roadside outhouse, I was back from death’s door and began to feel positively powerful.

Unfortunately, at this point John’s face became sunken and pale and he seemed to fall under the deadly spell of a wasting disease. At Bear Valley, we stopped at a gas station and he didn’t have the physical strength to effectively chew his Coke, much less the dreaded energy bar. Slouched up against a gas pump, he complained to Trevor, “I wish I was dead.” Trevor nodded in agreement and started to say something—but John stopped him and said, “No. I’m not just using the expression. I mean it, I really wish I was dead right now.”

John, too, eventually recovered and although Trevor never experienced the Grim Reaper effect, he was obviously fatigued during the four‑mile, two thousand foot wall from Hermit Valley to the top of Ebbets Pass. I felt positively fresh, although in retrospect I think I had merely become accustomed to the strain; my frame of reference was skewed. Now I could maneuver around and capture the moment in pictures with my little Minolta. Not that I could get Trevor to pose for the camera: he was locked into a fierce struggle against the mountain and the thin air of over eight thousand feet of altitude. Only on the ensuing descent did we all experience the euphoric feeling I described in my opening passage. Only then did we begin darting around like the bats that occasionally seemed to fly at our heads from out of the evening sky.

At Markleeville, we stopped for the night. Not because we were bedwetters, and not because the unexpected climbs had worn us out, nor because we’d realized our own physical limitations (since we evidently had none). The real problem was a tactical error we made in planning the ride: our itinerary just wasn’t sound. We planned to average about sixteen miles per hour, to take a certain number of rest stops, to spend between fifteen and eighteen hours on the bike, and to cover the distance in one day. We failed to figure in the amount of daylight available to us. Even with the horrors of Sheep Ranch Road, we kept to our original pace, perfectly, but we left at 7:00 a.m., which found us in Markleeville at 8:30 p.m., just as the sun had dipped out of sight, with about thirty‑five miles to go.

We could have made it if we were near civilization, but Lake Tahoe has its own style of darkness. It doesn’t mean reduced visibility with enough light pollution to get by. It means there is no difference between having your eyes open or closed. We had to get a motel room and park the bikes. We were too late for the restaurants in Markleeville, so we found a little store and picked out our dinners, which were very simple: John got Doritos corn chips, Trevor got Ruffles Sour Cream ‘n’ Cheddar, and I chose my favorite, Cool Ranch Doritos.

We ate at the motel. My third handful of chips tore the roof of my mouth like a cheese grater and I gave up on eating. A shower washed off the outer layer of my full body grime. You know the old cliché, “He was asleep before his head hit the pillow?” Well it didn’t apply in Markleeville. Too much pain was buzzing through our limbs to allow sleep. We all basically waited a bunch of hours for the sun to come up, while lying down with our eyes closed. The next morning, we climbed begrudgingly back on our bikes and hit the road. The climbing started immediately. Three hours later, Odyssey ‘91 was history. And so, of course, were we.

Final stats: 209 miles, 20,000 feet of climbing, 16½ hours, 15.7 mph average speed.