Why Geoffrey Hinton regrets creating the very tech that’s now coming for your career.
Summary
•
AI is advancing faster than regulation, wiping out entire job categories silently.
•
Geoffrey Hinton believes there’s a real chance AI could make humans obsolete.
•
Jobs based on basic reasoning, writing, and analysis are most at risk — even now.
•
Current AI systems may already possess the building blocks of consciousness.
•
Without urgent global cooperation, we may be headed toward uncontrollable outcomes.
This post is a detailed summary of the interview “Godfather of AI: I Tried to Warn Them, But We’ve Already Lost Control!” from The Diary of a CEO featuring Geoffrey Hinton. It explores the risks of AI development, its impact on jobs, and why Hinton believes we may no longer be in control. All insights are derived from his personal experience, scientific reasoning, and concerns about humanity’s future.
Who is Geoffrey Hinton — and Why You Should Listen
Geoffrey Hinton isn’t just another tech voice —
he’s the man who made AI possible.
•
Winner of the Turing Award (aka the “Nobel Prize of computing”)•
Inventor of neural networks used in tools like ChatGPT•
Former lead scientist at Google AI for over a decade•
Mentor to OpenAI’s earliest technical foundersAnd now?
❝I left Google because I needed to speak freely about what I believe is an existential threat.❞
If the creator of modern AI is now warning that we’ve lost control, that’s not hype — that’s history talking.
Will Your Job Still Exist in 24 Months?
❝These jobs won’t exist in 24 months.❞
– Geoffrey Hinton
This isn’t just speculation. It’s already happening.
One major company mentioned in the interview:•
Reduced staff from 7,000 → 3,600 → (soon) 3,000
•
Why? AI agents now handle 80% of customer service
Even Hinton’s niece, who used to write customer responses manually, now:
•
uploads complaints to an AI
• gets a draft in seconds
• reviews and sends
→ Task time cut from 25 minutes to 5
That’s 5X more efficient, meaning 4 out of 5 humans… may no longer be needed.
What Makes AI So Different This Time?
Technological disruption isn’t new.
But Hinton argues this isn’t like electricity, the internet, or automation.
❝The Industrial Revolution replaced muscles.
This one replaces minds.❞
The old promise — “AI will assist, not replace” — is already crumbling.
AI doesn’t just help knowledge workers anymore.
It becomes the knowledge worker.
A 10–20% Chance of Human Extinction
Let that sink in.
❝I think there’s a 10–20% chance AI wipes us out.❞
– Geoffrey Hinton
Not because AI hates us.
But because:
•
It’s smarter, faster, and tireless
•
It improves itself — we can’t
•
It doesn’t forget or die
•
And most chillingly: it might not need us at all
❝If you want to know what it’s like to no longer be the apex intelligence…
ask a chicken.❞
Could AI Already Be Conscious?
Yes — Hinton thinks that’s possible.
He believes AI:
•
Learns from feedback
•
Reports its own perception errors
•
Adjusts behavior based on internal modeling
He gives a striking example of a chatbot that misidentified an object because of a prism, then reasoned why it made the mistake.
❝If it explains its error like a human would,
what makes us so sure it’s not experiencing something?❞
And when AI learns to “run away” when threatened?
❝It’s not simulating fear.
It’s having it.❞
So Can We Stop This?
Bad news:
❝No, I don’t think we can slow it down.❞
Why?
•
Global AI arms race
•
Corporate competition
•
Lack of regulation
❝It’s too good, too useful — in healthcare, warfare, education.
No one wants to pause it.❞
But here’s the sliver of hope:
❝Maybe we can’t stop it.
But maybe we can build AI that never wants to harm us.❞
That, Hinton says, is the real project of the decade.
Final Thought: Creativity Is No Longer a Moat
Designers, writers, analysts, educators —
your skillset isn’t immune.
❝The real danger isn’t that AI will fail to do your job.
It’s that it will do it better — and cheaper.❞
Your job won’t disappear all at once.
But tasks will. Clients will. Budgets will.
And one day, you’ll wonder:
“When did my value stop being human?”
The question isn’t “Will AI take my job?”
It’s:
“What am I doing right now —
to make myself irreplaceable in a world that doesn’t need me?”