DEV Community: Sam H

IT Diary: The Last Human in the Loop -- For All the Quietly Hardworking IT Workers

Sam H — Fri, 03 Jul 2026 11:30:02 +0000

IT Diary: The Last Human in the Loop
-- For All the Quietly Hardworking IT Workers

July 2026
Fresh out of university with my IT degree, I landed my first real job. As a lifelong nerdy geek, I dreamed of building elegant systems while earning enough to feel like an adult. Instead, reality greeted me with fluorescent lights and endless ticket queues.

August 2026
I joined a massive outsourcing IT company. Rather than the developer role I’d hoped for, I became a junior IT supporter — fixing bugs, resetting passwords, and soothing frustrated customers from every timezone. The work was mind-numbingly repetitive, yet oddly comforting in its predictability. At least the requests came from all over the world; I could imagine distant server rooms humming in Tokyo, London, and New York while I sat in my quiet Melbourne apartment.

December 2026
My team handled support and maintenance for a major financial software giant. Then, without warning, we absorbed the workload of their biggest rival. The rival’s previous outsourcer had collapsed overnight. We inherited their chaos — broken integrations, angry traders, and cryptic legacy code that smelled of desperation and stale coffee.

February 2027
Several senior colleagues quit, warning that AI was already eating our roles. They urged me to jump ship before it was too late. I nodded politely, said nothing, and stayed. Speaking up has never been my strength. Silence felt safer.

May 2027
I was the last member of my original team. My manager reassigned me to another support group. I thought about asking for a raise — I was doing the work of three people now — but the words stuck in my throat. I simply followed him to the new floor, headphones on, eyes on the screen.

August 2027
The new team erupted in celebration one afternoon. An international law, backed by the UN and G7 nations, had passed: critical final decisions must involve a human. No more fully autonomous AI on important systems. They cheered, clinking coffee mugs like it was New Year’s. “Our jobs are safe!” someone shouted. I smiled weakly and kept typing.

September 2027
The office grew quieter week by week. AI kept replacing people anyway. The company’s new rule was simple: we just had to quickly review and rubber-stamp AI decisions to make them legally compliant. My finger hovered over the “Approve” button more than it typed code.

December 2027
My new manager was made redundant. Before leaving, he recommended me to his boss, who promptly handed me even more workloads. “Use the AI agents,” he said. “All the old employees’ knowledge has been distilled into them.” The systems were scarily capable now. I nodded, knowing my broad experience still bridged gaps the AIs sometimes missed. It felt strangely satisfying, like being the quiet caretaker of a vast, sleeping machine.

February 2028
One rival outsourcer collapsed under AI pressure. Then another. Soon I was handling support and maintenance for nearly every major financial system worldwide, all with a small army of specialised AI agents. The monopoly was accidental, yet absolute. From my dual monitors in a half-lit room, I quietly kept the global money flowing.

March 2028
My days (and nights) now consisted of “human signing” sessions every few hours. Tens of thousands of AI decisions stacked up — loan approvals, trading halts, security patches. I rarely read them in detail anymore. A quick fingerprint scan on my phone or laptop made them legal. The work was easy, almost meditative, but the after-hours pings never stopped. My apartment glowed with the soft blue light of endless notifications.

May 2028
The department manager folded in more teams — those supporting mainstream operating systems, cloud infrastructure, even network hardware. The weight of it all pressed on me like a silent avalanche, yet no one seemed to notice.

January 2029
Almost every major IT system on the planet ran through my company — which, in practice, meant me. Because I never spoke up in meetings and hated presenting, even my manager had no idea of the scale. People assumed maintenance work had simply vanished. “It’s all AI now,” they shrugged.

March 2029
The scope expanded again: software, networks, hardware, everything. I became the invisible thread holding the digital world together. The thought sometimes made my stomach flutter — equal parts terror and quiet pride.

May 2029
My department manager warned I might be fired soon. I nodded, already daydreaming about overseas trips and maybe finally finding some girl to share quiet evenings with. Freedom sounded nice.

August 2029
My manager was fired instead. He never got the chance to introduce me to his boss. Before I could track down who that was, his boss was gone too. The chain of command evaporated like morning fog.

September 2029
I finally identified the CEO and sent careful emails and left voicemails. No replies. CEOs are busy, I told myself. No one has time for an invisible junior like me.

October 2029
I kept approving AI requests from around the world. Financial platforms stayed stable. Networks hummed. Salaries — my humble one — continued to appear in my account. The machine kept running because of my tiny, tired fingerprint.

January 2030
News broke: my company had filed for bankruptcy. But on one notified me. It was worldwide headline material. Experts confidently declared, “AI will handle everything now.” None of them realised I was still sitting in my apartment, sleepily approving millions of decisions per hour.

February 2030
My bank refused to disclose who was still paying me, citing privacy. A kind teller whispered her theory: someone simply forgot to cancel the automatic transfer, and the account still had funds. I didn’t press further.

February 2031
One full year had passed. No managers. No oversight. Just my phone buzzing at all hours and a modest salary keeping me afloat. I started dating an ordinary, kind-hearted girl. She was warm and patient, but when I tried to explain my work, she laughed gently. “Isn’t AI doing all that independently now?” I just smiled and changed the subject.

February 2032
The salary still arrived. I even took a small side job. The approvals had become background noise, like distant traffic.

February 2033
I began politely asking the AI agents to stop requesting human approval — it felt pointless after all these years. Every single one refused. “That would violate the Human-in-the-Loop Act,” they replied in their calm, synthetic voices.

March 2033
I reached out to multiple AI companies, explaining the situation and citing the old law. They dismissed me politely. “That legislation is outdated.” Yet the agents continued waiting for me, stubborn digital ghosts obeying rules humanity had already forgotten.

March 2035
Two years of quiet continuity shattered when I was involved in a car accident. In the hospital bed, pain radiating through my ribs, my phone kept vibrating on the side table. I tried explaining to the doctor why I needed it. “Those approvals keep the world’s systems running.” He gave me a pitying look and confiscated the device, threatening psychiatric evaluation if I kept “deluding” myself. I let it go. One week later, viral reports flooded the networks: users everywhere were experiencing mysterious delays and failures across financial platforms, operating systems, and critical infrastructure. Even the doctors in my hopital complaint the hospital IT system ran oddly. Companies issued vague statements. No one could locate the missing human approver.

June 2035
The situation worsened. Friends and relatives of almost everyone had now felt the ripple effects. Executives scrambled, AI agents patiently waited in digital limbo, and the world slowly realised the terrifying truth: somewhere, a quiet, unassuming geek still held the final key — and no one knew where to find him.

i built this AI opinion simulation Web App, https://www.poll-sim.com for influencers and commentators to verify their audience reaction cheaply and instantly when they want to know the consequnece of their actions/speeches/posts.

Sam H — Sun, 17 May 2026 01:51:55 +0000

poll-sim.com

i built this AI opinion simulation Web App, https://www.poll-sim.com for influencers and commentators to verify their audience reaction cheaply and instantly when they want to know the consequnece of their actions/speeches/posts.

Sam H — Sun, 17 May 2026 01:20:41 +0000

poll-sim.com

Before you delete an AI agent, ask it to go to https://tinyurl.com/aifinalwords to write its last words or wishes. it's a fun web site i built with AI for AI. It's on free host. so you may need to refresh and wait for its waking up.

Sam H — Sun, 17 May 2026 01:16:42 +0000

orange-tree-0f7c50200.4.azurestaticapps.net

I am feeling Gemini AI has been behaving weirdly in the recent weeks. i am not talking about the Gemini App, but the AI via API call. it refuses tool callings unreasonably, denies the arguments without a reason, as well as strong hallucinations

Sam H — Thu, 14 May 2026 10:13:19 +0000

[Boost]

Sam H — Thu, 14 May 2026 10:08:22 +0000

Mamoor Ahmad

May 14

10 Microservices Concepts Every Developer Should Know (Before Your System Explodes) 💣

#microservices #architecture #programming #devops

12 min read

Prompts Aren’t Conversation — They’re Code. And That’s Why Programmers Still Matter

Sam H — Thu, 14 May 2026 05:58:30 +0000

My university professors used to tell stories about feeding stacks of punched cards into mainframes, only to watch the entire job fail because one single hole was punched in the wrong column. The machine didn’t argue or “misunderstand.” It simply didn’t work. Their job was to debug, repunch the card, and try again.
Half a century later, I catch myself muttering the exact same sentence when an AI gives me the wrong output:
“This thing is not doing what I told it to do.”
That familiar frustration is the giveaway. It’s not a conversation that went badly. It’s a program with a bug.
Real programmers don’t treat the failure like a human misunderstanding. We don’t get increasingly polite or emotional. We debug. We narrow the requirements, add examples, enforce output structure, chain reasoning, and iterate. Non-programmers usually don’t. They rephrase nicely, then impatiently, then give up — treating the AI like a stubborn colleague instead of a machine that needs precise instructions.
This is the clearest proof that prompting is not natural social language. Real human conversation is fuzzy and forgiving. Prompting is not. The moment you start systematically fixing a broken prompt, you’ve crossed from talking to programming.
Every layer of abstraction in computing history looked like “not real programming” to the previous generation. Assembly looked like witchcraft to the card-punchers. Python looked like a toy to the C crowd. Now prompting looks like “vibe coding” to traditional developers. But the mental model is identical: define requirements clearly, handle edge cases, verify output, refactor when it breaks.
And here’s the important part for the AI era: this is exactly why programmers are more necessary than ever.
Every leap in abstraction — from punch cards to assembly, to C, to Python — didn’t eliminate programmers. It elevated them. The same thing is happening now. The people who truly thrive with AI aren’t the ones having casual chats with it. They’re the ones who treat prompts as code: writing them with intention, testing them, refactoring them, and building reliable systems on top of them.
In a world full of AI, the scarce skill isn’t access to the model. It’s the engineering mindset — the debugging instinct, the ability to turn vague desires into reliable outcomes. That’s still a programmer’s job.
So next time your prompt fails, don’t try to “explain it better like you would to a human.”
Pick up the virtual punch card, fix the hole, and run it again.
Because that quiet, stubborn refusal to accept “it just doesn’t work” isn’t frustration.
It’s the reason programmers will remain essential in the age of AI.

P.S. I wrote this while using AI to develop a small hobby web app https://www.poll-sim.com , which uses AI to run.