Okay, so cards on the table: I wasn’t even gonna write about this.

But then a friend spammed me with YouTube videos by leading experts.

“AGI and the end of the world (lol maybe)” and now my brain won’t shut up about 2027.

Yeah, that year. Apparently it’s the new “Y2K” except instead of clocks breaking, it’s society. Fun.

Why is everyone obsessed with 2027?

So here’s the gist:

• Daniel Kokotajlo (used to be at OpenAI) threw out a forecast that AI could go from “neat blog post generator” to “actually running science itself” around 2027. He’s since nudged it closer to 2029, but once a number sticks, it sticks.
• Dr. Roman Yampolskiy… well, he just jumps straight to “99% unemployment by 2030.” Cool, thanks Roman. Really appreciate that bedtime story.
• And then Ross Coulthart, who is usually on UFO beats, keeps circling back to 2027 like he accidentally saw something in a classified file and can’t stop dropping hints.

Different people, different angles, same date. Honestly feels like one of those conspiracy theory cork boards with red string, except this one ends with me pacing around my living room muttering about Docker.

Right now vs. the scary version

Current AI: clever intern energy.

• Copilot finishes code when your brain flatlines.
• ChatGPT writes emails that sound professional but kinda soulless.
• MidJourney spits out fantasy art where everyone has extra fingers.

That’s fine. Annoying sometimes, but fine.

AGI (general intelligence) is the upgrade nobody knows how to handle. It doesn’t just copy patterns — it learns, adapts, reasons. And then crank that dial → superintelligence. Which is basically: smarter than us at everything, doesn’t need sleep, and could make a thousand clones of itself before lunch.

Nick Bostrom wrote about it like it’s stepping into an elevator that shoots straight up. Which sounds dramatic, Nick, but some of us are still waiting for the elevator in our crappy apartment buildings, so maybe chill.

What about dev jobs though?

This is the bit that actually makes my stomach twist.

Juniors:

Oof. The usual “cut your teeth on CRUD apps” path? AI already does that faster. By 2027, junior dev might literally mean clicking approve on AI pull requests. Which is… not a great way to learn.

Mid-levels:

Less about cranking out code, more about designing systems so the AI doesn’t quietly build a house of cards that collapses the second traffic spikes.

Seniors:

Half-architect, half-therapist, half-AI babysitter. Your role becomes making sure the AI doesn’t confidently do the wrong thing really well. And you’re still expected to mentor juniors who never got the classic “suffer through your own bad code” education. Good luck with that.

Teams:

Picture standup where one “teammate” is an AI that already closed 50 tickets while you were asleep. Do you clap? Do you quit? Do you quietly open Jira and assign yourself a “fix typo in README” just to feel useful?

Open source:

Brace yourself for GitHub spam. Thousands of AI-generated frameworks with README files that hallucinate features. Half the repos abandoned within a week. Open source might stop being about building new things and turn into janitorial duty.

Hiring? Fewer humans, more weird roles. Why hire ten juniors when two seniors + AI can do it? Interviews might pivot from “reverse this binary tree” to “explain why this AI microservice is about to catch fire.”

Outside the dev bubble

I’ll be quick here because honestly I’m not an economist and my brain leaks when I try to read econ papers. But yeah:

• Clerks, data entry, junior-anything = toast.
• Doctors, teachers, devs = they’ll morph. Still humans in the loop, but AI quietly running diagnostics or lesson plans in the background.
• Caregivers, nurses, performers = safer. People actually want people here.

The scary part isn’t job loss. It’s speed. Industries that usually take decades to change might flip in months.

Who owns the magic?

The real nightmare isn’t “AI gets smart.” It’s who controls it.

• Big corps? Hello, Cyberpunk Monopoly.
• Open-source utopia? Maybe… but also probably chaos.
• Reality? Somewhere in the messy middle. Some countries sprint ahead, some resist, inequality balloons, billionaires become trillionaires.

(And no, I don’t trust the Zuckerbergs of the world with this. Do you?)

Okay but… who are we if we don’t work?

This is where it gets uncomfortably existential.

We’ve wired identity to jobs. “So, what do you do?” is basically small talk 101. Kill jobs, and you don’t just nuke paychecks, you nuke identity.

Some optimists go: “Universal Basic Income will free us to make art and fall in love and stuff.” Others go: “Nah, people will just rot in VR headsets eating instant noodles.”

My gut? Humans are too restless. We’ll still build weird things, chase meaning, annoy each other. But the scoreboard shifts. Less productivity, more creativity, connection, “who made the weirdest project this week?” vibes.

So how do we prepare (without losing our minds)?

I mean, none of us have the full answer. But here’s what feels sane-ish:

• Stop memorizing syntax. Seriously. If AI can write it, why are you cramming it? Learn systems. Tradeoffs. Stuff that doesn’t autocomplete.
• Get AI-fluent. Doesn’t matter if you love or hate it. Prompting, auditing, fine-tuning — this is just another tool you need to wield.
• Lean into human skills. Communication, leadership, vibes. (Yes, vibes are a skill.)
• Don’t tie your identity 100% to your job title. Be a builder, a tinkerer, whatever.
• Support open ecosystems. Otherwise it’s “Congrats, five companies now run reality.”
• And yeah — don’t do it alone. Community matters. Share knowledge, ride this wave together.

Should we be worried?

…yeah. Of course. But not in the “killer robots march down Main Street” way. More like:

• Jobs evaporating before retraining catches up.
• Wealth condensing into fewer hands than ever.
• Governments using AGI as the ultimate surveillance toy.

Yuval Noah Harari once said AI could be “the end of human history — or the beginning of something much bigger.” Which is dramatic, sure, but when you slap “2027” on it, it suddenly feels way too close.

Wrapping this ramble

So yeah. 2027 probably won’t be the apocalypse. But it might be a hinge moment — the year the door swings open and we realize we can’t go back.

For devs: your job won’t vanish, but it’ll mutate. (And yes, LeetCode will still somehow survive. Evil never dies.)

For everyone else: economies will wobble, identity will need a reboot, inequality will spike before anything stabilizes.

Should we worry? A little.

Should we also be hopeful? Yeah, because panicking alone in a dark room isn’t a strategy.

The real story of 2027 won’t just be about AI.

It’ll be about what we become alongside it.

And no — AI still won’t fix Jira. Some problems are beyond superintelligence.