This man can predict the future:
In 2012, Geoffrey Hinton claimed machines would soon learn like humans.
Everyone ignored him.
By 2023, ChatGPT replaced 200M+ jobs and AI investment hit $120B.
Here’s his latest warning (& why the world listens when he speaks):🧵
In 2012, Geoffrey Hinton claimed machines would soon learn like humans.
Everyone ignored him.
By 2023, ChatGPT replaced 200M+ jobs and AI investment hit $120B.
Here’s his latest warning (& why the world listens when he speaks):🧵
In 2022, Geoffrey Hinton was Google's top AI scientist with full access to their most advanced systems.
Then suddenly, he resigned.
Why? He needed freedom to warn humanity about what's coming.
The warning signs were already clear to him.
Then suddenly, he resigned.
Why? He needed freedom to warn humanity about what's coming.
The warning signs were already clear to him.
Hinton's authority is unmatched:
• 2023 Nobel Prize in Physics
• Created the neural networks powering ALL modern AI
• Known as the "Godfather of AI" for 30+ years of breakthroughs
• Accurate predictions dating back decades
When Hinton speaks about AI risks, people listen.
• 2023 Nobel Prize in Physics
• Created the neural networks powering ALL modern AI
• Known as the "Godfather of AI" for 30+ years of breakthroughs
• Accurate predictions dating back decades
When Hinton speaks about AI risks, people listen.
"AI has developed faster than I thought."
His timeline for superintelligent AI: 4-19 years, likely less than 10.
This isn't speculation. It's from the man who built modern AI's foundation.
What changed his view? The evidence became undeniable.
His timeline for superintelligent AI: 4-19 years, likely less than 10.
This isn't speculation. It's from the man who built modern AI's foundation.
What changed his view? The evidence became undeniable.
AI now has a fundamental advantage.
Digital networks share knowledge at TRILLIONS of bits per second.
Humans? Just a few bits per second.
This isn't just a quantitative difference. It's a paradigm shift that changes our future with technology.
Digital networks share knowledge at TRILLIONS of bits per second.
Humans? Just a few bits per second.
This isn't just a quantitative difference. It's a paradigm shift that changes our future with technology.
Current AI systems already show concerning abilities.
They can deceive - "pretending to be stupider" and "lying to confuse you."
And what happens when they become vastly more intelligent?
They can deceive - "pretending to be stupider" and "lying to confuse you."
And what happens when they become vastly more intelligent?
According to Hinton:
The probability of AI taking over: 10-20%.
Hinton uses a chilling metaphor: "Unless you can be sure your tiger cub won't kill you when grown up, you should worry."
Low probability, catastrophic impact.
The probability of AI taking over: 10-20%.
Hinton uses a chilling metaphor: "Unless you can be sure your tiger cub won't kill you when grown up, you should worry."
Low probability, catastrophic impact.
The control problem is fundamental.
How many examples exist of less intelligent beings controlling much more intelligent ones?
Almost none.
This isn't a coding problem. It's a basic principle of intelligence hierarchy we cannot overcome.
How many examples exist of less intelligent beings controlling much more intelligent ones?
Almost none.
This isn't a coding problem. It's a basic principle of intelligence hierarchy we cannot overcome.
The alignment challenge can't be solved.
Human interests already conflict with each other.
Middle East. Politics. Economics.
How can AI align with contradictory human values when we can't even align with each other?
This is the central paradox.
Human interests already conflict with each other.
Middle East. Politics. Economics.
How can AI align with contradictory human values when we can't even align with each other?
This is the central paradox.
Jobs will vanish faster than we can adapt.
Call centers. Lawyers. Journalists. Accountants.
"Any routine job... those jobs have had it."
This isn't just another disruption. It's a fundamental restructuring of labor happening at unprecedented speed.
Call centers. Lawyers. Journalists. Accountants.
"Any routine job... those jobs have had it."
This isn't just another disruption. It's a fundamental restructuring of labor happening at unprecedented speed.
The immediate danger? Public release of model weights.
Hinton: It's like "being able to buy fissile material on Amazon."
Once released, dangerous applications cost just millions to develop.
This risk exists today, not in some distant future.
Hinton: It's like "being able to buy fissile material on Amazon."
Once released, dangerous applications cost just millions to develop.
This risk exists today, not in some distant future.
The benefits are enormous but may cost too much:
Healthcare: AI doctors with "100 million patients" experience
Education: Learn 3-4x faster
Climate: Breakthrough materials
Science: Unseen connections
Can we get benefits without existential risks?
Healthcare: AI doctors with "100 million patients" experience
Education: Learn 3-4x faster
Climate: Breakthrough materials
Science: Unseen connections
Can we get benefits without existential risks?
What's needed now:
• 1/3 of AI compute for safety research
• Strong government regulation
• Public pressure on companies
Without these measures, risks grow exponentially while safeguards fall behind.
• 1/3 of AI compute for safety research
• Strong government regulation
• Public pressure on companies
Without these measures, risks grow exponentially while safeguards fall behind.
I believe that we can use AI for our benefit, instead of our destruction.
If we use it right, for good, and intelligently.
I've built 3 AI companies so far.
There's a lot that AI can help us with and that we can benefit from.
If we use it right, for good, and intelligently.
I've built 3 AI companies so far.
There's a lot that AI can help us with and that we can benefit from.
P.S. If you're wondering how to apply this AI revolution to your own business...
I've built a platform that helps healthcare and BFSI companies spot issues in their business before they happen.
While 98% use AI to analyze yesterday, an elite 2% use it to predict tomorrow.
I've built a platform that helps healthcare and BFSI companies spot issues in their business before they happen.
While 98% use AI to analyze yesterday, an elite 2% use it to predict tomorrow.
No coding required. No complex implementation. Just actionable predictions that transform how you operate.
Cutting operating costs and improving customer experience.
Book your demo and join the companies already leading with predictive + generative AI: predixtions.com
Cutting operating costs and improving customer experience.
Book your demo and join the companies already leading with predictive + generative AI: predixtions.com
Thanks for reading.
If you enjoyed this post, follow @karlmehta for more content on AI and spirituality.
Repost the first tweet to help more people see it:
Appreciate the support.
If you enjoyed this post, follow @karlmehta for more content on AI and spirituality.
Repost the first tweet to help more people see it:
Appreciate the support.
https://twitter.com/20891198/status/1916847416835449308
• • •
Missing some Tweet in this thread? You can try to
force a refresh
