AI is cutting edge but it still ain’t perfect

you are a million

Listen to Brenda tell the story

“Those durned ve-hicles are a curse to humanity!”

That’s probably what my great-grandfather said as he stood at the roadside watching a Model T Ford rumble past. New technology ruffles people’s feathers, particularly those of older birds. Artificial Intelligence or AI is no exception. Not a day passes without dire warnings about how it will not just transform our lives, but possibly destroy them. As a lifelong writer and translator, I fully embrace how it has made my work infinitely easier.

But AI still makes mistakes. Sometimes some pretty funny mistakes. For once, I have zero qualms about making fun of something, because you can’t offend a computer, right?

Let’s take a stroll through my life of translation to appreciate how much time and effort AI saves – and where it still messes up, which is far more entertaining.

In the beginning, there was the notebook

I did my first translation from Spanish to English decades ago with a pen, paper and a small stack of well-thumbed dictionaries. One was for common words like convert or twist, others were business and medical dictionaries for capitalization, return on investment, blood transfusion or neural aneurysm.

Sometimes I was obliged to sneak into the local bookstore to surreptitiously check on very specific terms. Translation earnings do not permit luxuries like specialized dictionaries. 

The electric Selectric

Once my handwritten copy was finished, I typed it up on an IBM electric typewriter with one of those balls covered with letters that jumped around into position when you typed:  the IBM Selectric typewriter. What a privilege to write on such a modern machine! I was an excellent typist, having honed my skills in a class at the YWCA. I practiced by typing up newspaper articles after hours at my travel agency job. Soon, I worked my way up to 75 words a minute (after subtracting mistakes) to get a decent-paying clerical job.


Then came the memory typewriter. Now what could possibly top this? It remembered what you typed. No longer did you have to put the paper back in the machine juuuust right to line up the mistake to correct it.

Unless, of course, you had left out a word or letter. Then you had to retype the whole page.


Along came the personal computer, finally. The first one I worked on was made by the pioneering but long since defunct firm, Wang. It was a word processor belonging to a secretary named Connie at the company where I worked as a student. She let me sneak in during the evening to type my term papers.

It was so easy! I felt like I was cheating.


At my next company, it was an early, DOS-based IBM PC – and that’s DOS the program, not dos, the number two in Spanish. A man named Kevin was the only one who knew how to operate it. I worked diligently on this clunky machine for two hours, typing a proposal. Then I hit a key, thinking I was saving the document.

Instead, the entire thing disappeared. My proposal had vanished with a single keystroke. The screen was a dark green, empty but for a blinking, accursed cursor.


I raced off to find Kevin, a kindly guy with a scraggly beard which he stroked thoughtfully while I pleaded with him to retrieve my file. I didn’t know it at the time, but Kevin was an early geek, sort of an IT Neanderthal. He already had the right outfit, at least, but the technology had a long way to go.

“Hmm, I’m not sure that’s possible,” he mused. “If you want, though, you can flip through the manual,” he said, hefting it in his hand in my direction.

It was approximately the size of a small barn.

Taking it gingerly with both hands (so as not to damage my back), I lowered it onto a sturdy table and flipped it open to the table of contents. But even the headings were unintelligible. What was all this gobbledegook? I couldn’t even figure out what to look up – and was afraid to keep turning the pages. Who knew what sort of bombastic buzzwords lurked inside! 

I retyped the whole thing. On the typewriter. At least the paper didn’t suddenly vanish if you happened to sneeze and accidentally hit the wrong key.

Fast forward to AI

Zooming ahead through the breathtaking galaxy of technology, we pass laptops, tablets, smartphones…

…and land in the world of AI. Here, you don’t have to work at all. The machine does it all for you.

Not quite.

I use AI tools for translations from German and Dutch into English. Since Dutch has a much smaller number of speakers than German, the Dutch algorithm is slower at learning. It’s like an intern who only works part-time. They simply take longer to learn the job.

Let me share some doozies produced by Mr. Dutch Algorithm Intern:


The Dutch are known to be very direct, a trait manifested perfectly by this word, which means “venture capitalists.” This profession was probably initially referred to in English as “people so rich they can afford to lose all their money.”

Surprise! When the banks looked around for such folks, they failed to raise enough capital, so they do what everybody does: they hired some PR guys. They rebranded this to “venture capitalists,” smoothly combining the idea of earning money with just a smattering of adventure. Very appealing.

She gave them big thinking

She advised them to think big. This is cute. It’s somehow better than what we would say in English: think big.

Just saying, “Let’s do some big thinking!” makes me feel like my head is expanding with additional creative capacity.

Shared grief is half grief

I love this expression! It works perfectly in German but not in English, even though it’s understandable.

Of course sharing your sorrows with someone makes you feel better. It isn’t “half grief,” but it sure feels that way. It sounds much more upbeat than the English expression: misery loves company. This conjures up images of a grumpy guy on a park bench commiserating with another grumpy guy slouched on the same park bench. Sure doesn’t seem like the kind of guy who would take away half my misery – more like someone who could use half my money.  

To try to knit a sleeve to it

This means to try to find a solution and it makes perfect sense to anyone who has ever tried knitting. Click-clacking the needles to produce a row of reasonable stitches for a sweater is hard enough. But knitting a sleeve that fits? A perfect metaphor for “finding a solution.”

It’s not an egg, but I’m used to it

This means it’s not easy. This expression was doubtless inspired by a fox, raccoon or other woods-dwelling animal that eats eggs whenever they can find them. Basically, fun forest fast food. Come to think of it, this applies just as well to a Sunday morning omelette.

Even people without a statistical lump should read their Covid-19 introduction.

Apparently those with a statistical lump – I assume this is on the head – are particularly good at statistics. Unfortunately, I must report that I don’t have one of these, which caused me significant anguish in college. Nor do I have a math, physics or chemistry lump.

I’m thinking a lump on my head might not be particularly attractive. On the other hand, if you manage to get enough of these lumps, taken as a unit they’d create a smooth surface. A worthwhile tradeoff for the expertise, especially when I recall my final exam in statistics.

Mass testing is of little use if only one man and one horse’s head are carrying the virus

I get it. One man, one horse. But a horse’s head? Why would the head get the virus and not the rest of the horse? And why does this expression evoke a picture of a man with a horse’s head tucked neatly under his arm, marching down the street?

And let’s face it: once the head has been severed from the horse, it’s going to catch a lot more than the coronavirus.

The only strange ducks in the bite are Swedes

This is in reference to early strategies for handling corona, which consisted pretty much in letting the virus take its course to achieve herd immunity. The Swedes were the only ones in Europe pursuing this strategy. You know, the only strange ducks in the bite.

My family used to have ducks and they would chase our cats and bite their tails. This would perhaps be an example of a “strange duck in the bite.”

Maybe, but imagining this would require some big thinking.

Some countries are above the threshold of three percent, but in terms of testing capacity, it is generally mustache.

Meaning not relevant. Because even a big handlebar mustache can be quickly shaved off. Instantly irrelevant, so to speak.

I hope it takes a while for the Dutch algorithm to catch up to the others. How boring it will be when it stops producing such abominable aberrations.

Perfection may be perfect, but it’s not much fun.

Brenda Arnold

Leave a Reply

Your email address will not be published. Required fields are marked *