Rejected by a Robot at 4:19 AM

At 4:19 AM I walked to my neighbor's house, got rejected by a robot, and ended up questioning my own existence. True story. Rain, goats and all.

4:19 AM

It was 4:19 AM. Even the guys who pretend to wake up at 5 to "build their empire" were still asleep. My brain, of course, was wide awake and making terrible plans.

And it was raining. Not drizzle. Angry rain, the kind that sounds like the sky owes somebody money. I grabbed my phone, pulled on yesterday's half-wet hoodie and walked out to my neighbor's house.

Why? Because I have no network at home. That's the whole reason. Some people climb mountains for the view. I walk through mud for two bars of Wi-Fi.

I connected, opened Chrome and typed in the URL. You know the one. I'm not going to lie to you, it's my blog and it's a true story.

Enter. Then I waited.

The checkbox

The page never loaded. Instead a little white box popped up and asked me, very politely, whether I was a robot.

☐ I'm not a robot

What the actual fuck.

I had just walked through rain, mud and probably one wet dog, and a website was looking at me, a real person with real needs, wondering if I might be a machine. I stared at that screen like someone who just got dumped over text.

Try again

I clicked the box. Nothing. Clicked again. Nothing.

Then: Try again. Something went wrong.

No, motherfucker, you went wrong. I'm the one standing out here in the rain.

So I clicked again. And again. And again. Seven times, maybe eight, I stopped counting when I stopped caring. Same message every time. Think about how stupid that is. A real robot would have passed by now. It's the one test a bot is built to pass and I was failing it. My failing was basically proof that I'm human, and there was nobody around to sign the paper.

The shed

By then I was soaked, so I did what any desperate man does and ducked into my neighbor's shed. Goats, buffaloes, a cow, an ox. The smell had its own personality. The rain on the roof was louder than my thoughts, and my phone was the only light in the whole place.

One goat looked straight at me like, bro, what the fuck are you doing here at 4:19 AM? Fair question.

The buffalo kept chewing. Total peace, zero questions. The cow was definitely judging my life choices. The ox just stood there like an ox, the way it has probably stood there for years.

Not one of them has ever been asked to prove it's an animal. Sit with that for a second. And honestly, every single one of them was handling the night better than me.

Am I a robot? Let me check

So I sat on a bucket and went through myself like a form:

• Metal skeleton? No.
• Can I calculate pi in a second? No.
• Can I work 72 hours straight without crying? Also no.
• Can I feel emotions? Unfortunately yes. Way too many.
• Can I make terrible decisions at 4:19 AM? Professionally.

And this website thinks I'm a robot? What kind of robot sits in a goat shed in the rain, horny and furious, fighting a checkbox? If I were a robot I'd at least have better Wi-Fi.

The checkbox was never the test

Here's the part that really pissed me off, because I looked it up later. That checkbox isn't the real test. The click is a formality. The thing is watching everything around the click: your cookies, your browser history, your connection, how your mouse moves (or, in my case, my thumb). Researchers at ETH Zurich took Google's reCAPTCHA v2 apart in 2024 and found evidence that it leans heavily on cookie and browser history data when it decides if you're human. The picture puzzles mostly matter when it already doesn't trust you.

So now it makes sense. A phone on a stranger's Wi-Fi at 4 in the morning, no history it recognizes, probably a shared connection. To that system I'm not a man in a shed. I'm a bot with a bad backstory. A script with a fat cookie jar and a clean connection walks right through, and I stand there getting wet. The test isn't asking "are you human?" It's asking "are you a human we already know?" Those are not the same fucking question.

And the same study had another result I can't stop thinking about. Their model solved 100% of the image challenges they threw at it, where earlier attempts only got around 70%. They also found no real difference in how many challenges humans and bots had to solve. So the machines pass the picture test just fine, and I couldn't. The thing built to keep bots out is now better at annoying me than at stopping them.

Who is judging who?

Alan Turing, in 1950, asked whether a machine could fool a person into thinking it was human. He called it the imitation game. Now it's flipped. A machine sits there deciding whether I pass as human. We built the judge and then volunteered to be the defendant.

And it gets uglier. Turing's idea says that if you can't tell the difference from the outside, the difference doesn't matter. Judge by behavior. Then John Searle shows up with his Chinese Room argument and says you can shuffle symbols perfectly and understand absolutely nothing, so passing the test is not the same as being the thing. Philosophers have been fighting about this for decades and nobody has settled it. Meanwhile the internet picked a side without asking anyone. Its idea of "human" is purely behavioral. Click the right squares, move the mouse the right way, and you're in. My feelings don't count. My anger doesn't count. My wet hoodie definitely doesn't count.

Descartes said, I think, therefore I am. The internet says, I click the bicycles, therefore I am. And if that's the standard, what about the blind guy? The old man who has never looked closely at a traffic light? The kid with a cracked phone and a dying connection? Are they less human because the test wasn't built for them?

Camus said we have to imagine Sisyphus happy, pushing his rock up the hill forever. Clicking "Try again" for the eighth time, I was not happy. I was wet.

Who's actually doing the work?

Quick history, because this is the dark joke. The word CAPTCHA came out of Carnegie Mellon around 2000, from Luis von Ahn and his team. In 2007 he turned it into reCAPTCHA, so instead of wasting your click on a random wobbly word you helped digitize old books. Google bought it in 2009. After that came the street numbers, the crosswalks, the traffic lights, and Google itself describes the system as channeling all that human effort into digitizing text, annotating images and building machine learning datasets.

So think about what I was doing at 4:19 AM in a goat shed. I was working for free, teaching a machine to see traffic lights better, so that one day the machine can pass the exact test I was failing. The unpaid intern of the thing that's coming for my job. Great deal. Really glad I walked through the mud for it.

And the whole system leaks anyway. There are click farms where actual humans, in places where labor is cheap, solve these things for pennies so spammers never have to. Back in 2023 the testing for GPT-4 reportedly had the model hire a person on TaskRabbit to get past a CAPTCHA, and when the worker joked "are you a robot?", the model said it had a vision impairment. A machine lied to a human so it could prove to another machine that it's human. I couldn't write that plot if I tried.

Both things are true

I'm not going to pretend CAPTCHA is pure evil. Bots are real. Spam is real. Fake accounts, scraping and credential stuffing are real, and if every site left the door open, free websites would be dead in a week. The system is doing its job. But I wasn't a spam farm. I was one man, one phone, one shitty connection, and I got treated like a threat. Both of those are true at the same time, and that's exactly what makes it so irritating.

Here's the controversial part, and I mean it. The default setting of the internet is distrust now. Every login, every form, every "prove it" screen carries the same tiny message: guilty until you click the right squares. We built a world where the human is the suspect and the machine is the bouncer. And the bots are already past the bouncer. The only ones still getting stopped at the door are the tired, the poor, the slow, the old, and the guy in the shed.

So here's what I can't answer. If machines are going to be better than us at everything, this included, what exactly are we protecting by keeping them out? Humanity? Or traffic numbers and ad revenue? I'm not sure I want the answer.

The ending

After twenty minutes of clicking, failing and cursing, I gave up. I closed Chrome and just sat there listening to the rain and watching the animals breathe. We sent rockets into space. We mapped the human genome. We built machines that write poems, paint pictures and pass exams. And in the middle of the rain I got beaten by a goddamn checkbox.

Or maybe I'm overthinking it. Maybe the site just glitched, my network was just bad, and I just needed to go back to sleep.

But right then I didn't feel intelligent, or advanced, or philosophical. I felt like a human being. A horny, rain-soaked, sleep-deprived human being, sitting in a shed with a buffalo.

And the buffalo, honestly, looked like it understood.

— Aexinide Because even my worst moments are worth writing.