"There’s a 10% chance AI wipes out half of humanity"
“There’s a 10% chance AI wipes out half of humanity.”
That’s what 𝘢𝘭𝘭 𝘧𝘰𝘶𝘳 leading AI modelers told Paul Tudor Jones at a private summit.
And no one in the room – packed with the world’s top investors, policymakers, and technologists – pushed back.
What Paul Tudor Jones shared on CNBC this week is one of the most fascinating (and alarming) takes I’ve heard on the current state of AI.
And it’s something everyone building in this space should listen to.
Here’s what Paul shared live on CNBC this week:
AI is improving at 25% 𝘵𝘰 500% efficiency every few quarters.
So what’s 𝘵𝘩𝘦 𝘳𝘦𝘢𝘭 𝘳𝘪𝘴𝘬?
It’s not your job getting automated. It’s not AI writing emails or building apps.
It’s this: 𝗔𝗜 𝗶𝘀 𝗰𝗼𝗺𝗺𝗼𝗱𝗶𝘁𝗶𝘇𝗶𝗻𝗴 𝗸𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 𝗳𝗮𝘀𝘁𝗲𝗿 𝘁𝗵𝗮𝗻 𝗼𝘂𝗿 𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝘁𝗼 𝘀𝗲𝗰𝘂𝗿𝗲 𝗶𝘁.
And that means:
- The tools for building bioweapons, cyberweapons, and large-scale disruption are getting easier to access
- Geopolitical competition is accelerating deployment, 𝘯𝘰𝘵 safety
- No global governance exists — and companies are too deep in the race to stop
One of the panelists literally said:
“It will take 50–100M deaths for the world to take this risk seriously.”
That is chilling.
And yet, it tracks with what 𝘦𝘷𝘦𝘳𝘺 leading AI expert is quietly saying behind closed doors.
And no one’s hitting pause. Not Big Tech, not governments because this is now a 𝗴𝗹𝗼𝗯𝗮𝗹 𝗮𝗿𝗺𝘀 𝗿𝗮𝗰𝗲 between companies 𝘢𝘯𝘥 countries.
We need more than just adoption. We need:
- Governance conversations
- Real security frameworks
- Standards around agent autonomy
- Checks on where and how AI is deployed
AI is the most powerful technology we’ve ever built.
Let’s treat it that way.
– Marc
You know the pain: finding the right use AI cases and vendors is slow, messy, and full of risk.
We are fixing that: faster discovery, smarter evaluation, real trust. Powered by AI.
And we want you to help shape it.