Dear Dr. Boli: Every so often I hear people talk about “the mysterious East,” as though everyone knew what that meant. What do they mean when they talk about “the mysterious East”? —Sincerely, A Professional Sinologist.

Dear Sir or Madam: They mean that, in spite of having some of the most delicious cuisines on the face of our planetary conglobation, ordinary people in China and Japan and nearby places have enthusiastically adopted American fast food. This is a mystery no ordinary theory can account for, and experts believe it can only be explained by quantum psychology.

How much does water weigh?

Water weighs as much as it likes. There is no use trying to put water on a diet.

If I gathered together seventy-five discarded phones with working batteries, would I have enough power to build a Bluetooth death ray?

Probably. But it is likely that it would kill only people who are wearing headphones or earbuds, and thus it would be of limited utility.

Is there any way to demagnetize an owl?

Owls are not generally magnetic. If your owl has become magnetized, it may be a symptom of a deeper problem. Consult a veterinarian with a specialty in avian electromagnetism.

What are the health effects of sarcasm?

Sarcasm leads to robust health and well-being, because obviously people who approach every human contact with an inflexible sneer are both living happy lives themselves and spreading happiness to the people around them, so that no one could possibly ever want to beat them to death.

If an anaconda swallows itself, does it get any fatter?

No. It remains exactly the same size. Anacondas are in fact entirely self-sustaining, and predation is nothing but a hobby for them.

It is always good fun to make predictions about our friends who are artificially intelligent and then watch the predictions not come true, and that is all the excuse we need for offering another one.

An email correspondent mentioned that she had been a bit put out when AI invaded her email. She had been ready to write a reply to a message from Dr. Boli when she saw that Google’s pet bot had already drafted a complete reply for her, plausibly human-sounding, and the bot offered to send the message off with one click and zero actual input from the putative author of the response. Is this the future? she asked. Will it soon be impossible to tell whether our correspondents have actually read and replied to our messages, or whether they have just let the bot do all their thinking and pressed the button without even reading their own replies?

Dr. Boli thought for a little while, and then gave opposite answers to the two questions. Yes, he wrote back, this is the future. More and more people will simply push the button and allow the bot to do all their thinking for them, because most people find thinking an unrewarding slog and will be glad to be relieved of that responsibility. But for that very reason it will be easy to tell which messages have been written by bots. They will all sound alike.

Right now, the purveyors of botty services emphasize how well the bots can imitate you. They learn your style, we are told; they write in your voice.

And that is true now to a certain extent. Large language models work by predicting what word is likely to come next. Give one access to all your accumulated emails and other writings, and by simple statistics it will start to sound like you.

But that depends on your having an archive of writings that come out of your own brain. Suppose you are one of the large majority who will be happy to accept whatever the bot writes and declare it good enough. From this point on, almost all your communications will be written by the bot. As time goes on, these botty effusions will form the majority of the communications that have gone out under your name. Older messages will fall victim to the various forces that cause old data to crumble away, and eventually your own unbotted writing will have vanished from the universe. The bot’s only models for “your voice” will be the things the bot itself has written.

This is a feedback loop, and a feedback loop always ends in a high-pitched squeal.

More and more of the world’s writing will consist of bots replying to bots. They will learn from themselves and each other, and they will all sound alike.

In the field of software documentation, this has already happened. Look at the descriptions of new projects on GitHub; look at the changelogs; look at the user manuals. They are all written by bots, with few exceptions, and they all speak in the same voice. And it is a distinctive voice. You quickly learn to recognize it.

That is the future of communication. It will separate the human writers from the bots very neatly, but—as we have already seen in the world of technical writing—the human writers will be hobbyists. Bots will be considered necessary for professional communication.

That is not necessarily a dismal future. Think how much you treasure a handmade sweater or quilt that no machine could ever duplicate. Messages you receive from people who care enough to write their own words will be messages you treasure the same way. Likewise, if you are one of the small number of people who still prefer to communicate in their own words, you will be regarded as a kind of prodigy. It will not increase you chances of remunerative employment any more than the ability to crochet an exquisite shawl will increase them. But it will earn you a kind of social awe, and that is worth more than employment in the long run.

The Eiffel Tower at Kings Dominion, Doswell, Virginia, photographed by Julia Traub.

---

Eiffel Tower.—If all the Eiffel Towers in the world were laid end to end, they would form an enormous horizontal metal pagoda.

Briggs’ Patent Mote-Magnifying Glasses. See the mote in your neighbor’s eye clearly and offer him the help he obviously needs.

---

50-Caliber 6-Inch Mosquito Eliminator. Solves mosquito problems expeditiously and permanently. Accurate aim in the hands of a skilled operator. Reloads in minutes.

---

Excelsior Bicycle Trap. Harmlessly traps bicycles and their riders for safe release in the wild.

---

Talking Bust of President McKinley. Repeats “The mission of the United States is one of benevolent assimilation” until you switch it off (off switch located on back of head, activated by sound thwack).

---

Dual-Action Snow-Removal System. Removes snow by scraping it from the ground and by accelerating global warming.

Dear Dr. Boli: My wife sent me to the state store to pick up some connecticut, but when I got there all they had was booze and stuff. So I decided I’d get some vodka, since my boss is driving me to drink, and I might as well get started. When I got to the vodka aisle, I was amazed and a little intimidated by the variety of brands, and even more by the variety of prices. There were big bottles priced at six bucks, and bottles the same size priced at eighty bucks. At first I thought the expensive ones must be the good ones, and the cheap ones must be the bad ones. But then I looked at the cheapest of the lot—$4.99 on sale till the end of the month—and I noticed the label said “Premium” and “Finest Made.”

So I guess I just don’t understand vodka. If the finest vodka made costs five dollars, why do other vodkas cost eighty dollars? I can’t make any sense out of it. —Sincerely, a husband who decided to get drunk on a good Côtes-du-Rhône instead.

Dear Sir: Vodka is basically distilled compost, so there is very little difference in quality between one vodka and another. In that sense any brand of vodka may claim to be the finest made. The price, therefore, depends on the quality of the graphic design on the label. The higher the price, the more the label conforms to the latest trends in graphic design; whereas you probably noticed that the five-dollar vodka has a label that looks like a thirteen-year-old’s idea of high class in 1958. If you are planning on impressing your friends with your taste in vodka, of course you will pay for the better label; and in addition you will eschew flavors like strawberry or lemon in favor of up-to-date flavors like pomegranate or açai.

Prairie Dogs.—From a municipal point of view, a so-called prairie-dog town is technically a borough.

Why is this car not going as fast as the driver would like? It is missing one essential thing.

---

Dear Dr. Boli: I understand the basic principles of internal-combustion engines: they operate by exploding, and as long as we pray hard enough, the explosions don’t kill us. I even have some acquaintance with electric motors, which operate by magic. But the thing I don’t understand is racing stripes. Whenever people want a car to go fast, they put racing stripes on it. They put racing stripes on racing cars. They put racing stripes on expensive sportscars, so that they can make it to the red light at the end of the Liberty Tubes before anybody else. Just this afternoon I saw racing stripes on a Honda Odyssey, and the woman who was driving it certainly acted as though she was on her way to a fire, or from a fire. So what is it about racing stripes, anyway? Do they really make cars go faster? —Sincerely, Bicycles Are a Mystery Too, but No One Wants to Talk About It.

Dear Sir or Madam: Racing stripes do indeed make cars go faster, but there is no mystery about it. It is simple physics. At high speeds, the single thing that retards the progress of a vehicle most is air resistance. Left to themselves, air molecules will just stand around in front of the car waiting to see what will happen. This creates resistance. It is desirable, therefore, to make the air less resistant, which is best done by making it move easily along the length of the car. Now, air molecules are simple souls, with a limited vocabulary, and it is not easy to make them understand us if we try to express what we want from them in complex sentences dripping with relative clauses. Yet they are not unwilling to please us—quite the contrary, in fact. What they need is for us to be simple and direct with them, and therefore visual communication is preferable to verbal. The racing stripes point the direction of desired air flow along the vehicle in a way that even the simplest air molecule can understand. Many mysteries have similarly simple explanations once you understand the physics of them. To balance a bicycle, however, requires divine intervention, which is a very complex subject indeed.

Dear Dr. Boli: I was talking to some guy on the streetcar, and he said that artificial intelligence isn’t really intelligent because it doesn’t meet the definition of intelligence. But he said I’d have to pay him twenty bucks to hear what the definition of intelligence is, and I didn’t have twenty bucks with me. So I figured I’d ask you, since you seem to answer questions for free: What is the definition of “intelligence”? —Sincerely, Ardis Wallslair McFurtle, Stowe Township.

Dear Madam: Intelligence is the quality that separates human beings from machines or animals. That is the only definition of “intelligence” that is of any permanent value. Specific criteria may be mentioned in some academic definitions of “intelligence,” but when we come across a machine or animal that meets those criteria, we change the criteria. The ability to make and use tools has been rejected as a criterion, because crows and apes can do that, and obviously they are not really intelligent. The ability to make reasoned arguments has been rejected, because AI bots can do that, and obviously they are not really intelligent. Eventually intelligence may be found to reside in the small number of genes we do not share with the chimpanzees. But, meanwhile, “the quality that separates human beings from machines or animals” is a reliable permanent definition, because it will not change no matter how much the facts have to be revised to fit it.

We had been astonished that the spammers who flood our site with thousands of comments had not adopted artificial intelligence to make their comments more plausibly human. Instead, they had been leaning on the old reliable techniques that go back to the very earliest days of comment spam twenty years or so ago.

1. Throw a specific-sounding comment at hundreds of sites and hope it may be relevant to one or two of them:

I’ve tried many horoscope sites over the years, but this one stands out for its accuracy and clarity. The daily forecasts are comprehensive yet easy to understand. I love how they break down the influences for each zodiac sign. This is now my go-to site for daily astrological guidance!

This comment was not left on one of the articles here that mentioned astrology.

2. Write something generically complimentary and hope that a few vain site owners will not notice that it could apply to any writing anywhere:

Your writing has a way of resonating with me and making me feel understood Thank you for being a relatable and authentic voice

3. Ask a question that might provoke the site owner to interact with you, thus bringing your comment more attention. Again, the question will have to be generic, since you are going to throw it at thousands of sites.

Thank you for any other informative site. Where else may just I get that kind of info written in such an ideal manner? I’ve a project that I’m just now working on, and I have been at the look out for such information.

4. Use article-spinning software to create endless retreads of the same comment. Article-spinners are based on the assumption that “synonym” means “exact drop-in replacement for a word.”

Pretty portion of content. I just stumbled upon your website and in accession capital to assert that I acquire in fact enjoyed account your blog posts. Any way I will be subscribing in your augment or even I fulfillment you get admission to consistently rapidly.

5. Do all of the above, but in Russian. There are bursts of a few days at a time when the overwhelming majority of comments sent to Dr. Boli are in Russian.

Almost all these comments are intercepted by our simpleminded spam-blocking software and sent to the spam holding cell, where they accumulate (right now there are 18,425 of them) until Dr. Boli decides to rummage through them to find a few amusing things and then deletes the rest.

But things may be changing at last. Yesterday we got our first AI-powered spam comment.

Dr. Boli’s satirical voice has always been a masterclass in gentle humor, and the carefully curated magazine format under his editorship continues that brilliant tradition. The way he blends scholarly pretense with absurdity creates a unique reading experience that rewards repeated visits.

It is immediately obvious that the bot has “read” at least some of the site, in that it “understands” what kind of site it is commenting on and adds enough specific detail to make that understanding obvious. It is using the same psychological manipulation as No. 2 above, but it is much more effective. If you did not notice the slick AI style of it, or the link in the commenter’s name that leads to a chatbot site, you might believe it was from a genuinely appreciative reader.

We also notice that it made it past the simple bot fence we have set up, meaning that the AI-powered commenter was able to outwit our little mousetrap, and we may have to build a better one. (Or it is possible that the comment was posted by a human who used AI to write it; we have no automatic defenses against human commenters except that their comments are held for moderation if they have not commented before.)

Well, now, what are we to do about this new AI comment spam? Obviously the same thing we would do if a real human left a spam comment. That does happen occasionally, and for the moment we see no reason to make any distinction between bots with bad intent and humans with bad intent. Or if there is any distinction to be made, it will be in favor of the humans. Dr. Boli has usually been happy to consider it news rather than spam if a human reader mentions and links to her own book, for example; he would not apply the same generous assumption to a bot.

Meanwhile, we can take this opportunity to admire how fast the bots have developed. They are turning into competent criminals very quickly.

And, by the way, if you wonder why your electric bills keep going up, it is because huge server farms are providing artificial-intelligence services to scammers and cheaters of all descriptions. In an ideal world, perhaps, the costs of all that extra power consumption would fall entirely on the criminals who do the consuming, who would thus be forced to learn valuable lessons about balancing current expense with anticipated income. In the world we have created, however, criminals are a privileged class, and it is the duty of the rest of us to support them by spreading the cost of their crimes through the general population. This is how we make sure that virtue is indeed its own reward: by not sullying it with any other material considerations.