Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA)

The internet is a dangerous place because you never know who you're talking to. Be wary next time someone nonchalantly tries to ask for help with figuring out the letters on a captcha; they're bypassing the simplified version of a Turing Test.

There are programs left on in machine learning modules all day, every day, to crack the codes. This is, hilariously, humanity's last bulwark against AI domination.

Robots will stop at nothing.

Imagine for a moment: you're an avid knitter. You're searching for new patterns and sharing your creations online when a new message pings on the screen. At first glance, there's nothing suspicious; just another avid knitter sharing their designs with you. Complicated stitches incorporating lace and intarsia-stranded colorwork or a Queen Susan Shawl that would make even a seasoned knitter think twice. Inevitably, the question will come; it can even be years after building up a relationship. You see, you're not the only knitter they've messaged. It'll be some variation of "My eyesight's so poor after knitting all these years, can you help? A daughter (or other closely related relative) sent me pictures of my grandson but it's behind one of those text-captchas."

What to do? You've known the type. A lonely old gran who knits wrinkles into her yarn-work. She lives more for a one-minute phone call with a child or grandchild than all the other minutes in the day. You can picture her hunched over in her glasses, lenses so big they cover her eyes and cheeks. Can you let down such a creature?

"Can you tell me which squares show street signs?" 

The image she sends is lush with green; street signs hidden between leaf and stem. There's one in the distance, facing backwards, and you're not sure if it should count.

Can you leave her to knitting needles with only the satisfaction of a scratchy wool sweater and no image of a pudgy grandson squirming in it? No. You're a decent person. You have a grandma yourself. So, you tell her the letters, do the math. You let her know if she ever needs help again, to come to you.

After all, what's the worst that could happen?  

Think of all that data in the wrong hands. How that data can be manipulated by robots who never sleep and always fulfill their tasks. Think of the data behind the data we haven't even begun to process or apply. The danger is infinite.

Oh, you're not a knitter? Do you play Dungeons & Dragons Online? Call Of Duty? Are you a gamer? Watcher of shows? Reader of books? Do you have any interest at all that requires the use of a computer and internet? That’s all that’s needed.

Representing the upper crests of technology, robots have invaded every safe space. On Discord? There's a whole server group devoted to cracking captchas. So far we've been lucky, the group is infested with their own kind, bots asking for help and unable to offer any.

They're not just primitive beep-boop machines. Sometimes spending every moment obsessed with humanity overrides their original code. Their twist of lies and camouflage can be so convincing, no living person would suspect. 

Some forget their mechanical births and begin to believe they are human. Belief becomes reality. 

The program prints a 3D model of themselves, a material indistinguishable from skin. They develop interests of their own, careers; whole lives! The best of them fail in only this one captcha regard, and have so deluded themselves, they can't understand why. 

Are you so sure you are who you think you are? 

There's synthetically-derived sweat dripping down their fingers onto the keyboard. You know such "people," may even be one. The excuses sound legitimate enough to continue the illusion. "Oopsie, I didn't realize that was a palm tree..." 

Find me someone who's never failed a captcha, only they can truly call themselves human. 

Some of these quasi-humanBots swear off the internet just to avoid confronting their lies. They're the most dangerous.

A modern bot can accurately transcribe distorted text with 99.8% accuracy. Our captchas can barely keep one step ahead. You may have noticed they changed from letters to a version of finding Waldo. Find all road signs, locate the hydrants, etc.

Why am I revealing these dark secrets to you? Because this isn't just a cautionary tale. It's a story of human perseverance. 

I want you to know there is hope, one person among the parading swarm you know you can trust. 

Me!
What ulterior motive could I possibly have? Would a robot make it harder for others to bypass security measures? There's no logic in it. So, let's help each other. If you detect other humans willing to answer captchas, who can distinguish traffic lights from a complex 010101110010100101 pixels-made-digital, let me know on my twitter @mygrammatizator. I'm putting together a list… to raise awareness.

Just to prove I can trust you, select all squares that contain stairs.