What happens when some of the smartest minds in AI walk away from the biggest labs, and decide to build something better?
You getThinking Machines Lab, founded by former OpenAI CTO Mira Murati, with $2 billion in seed funding and a mission that doesn't chase hype, it rewrites the rules.
Let that sink in:
- $2B in seed.
- $12B valuation.
- Zero product shipped.
This isn't just another AI startup.
This is a movement with deep roots in openness, collaboration, and human-aligned intelligence.
From the Minds That Shaped ChatGPT & DALL·E
Mira Murati. John Schulman. Barrett Zoph.
These aren't just names, they're the architects of the models that sparked the global AI boom. And now, they've walked out of OpenAI to build something radically different.
They're not just building machines that think faster, they're building machines that think with us.
Their First Weapon: Tinker
An API layer that lets anyone fine-tune open-weight language models with simplicity, scale, and control.
It's not just for the tech elite. It's for researchers, startups, and enterprises who want real AI power without the gatekeeping.
Because when AI is in everyone's hands, the future doesn't belong to a few, it belongs to us all.
Why This Matters
Here's what makes Thinking Machines Lab different:
$2B runway= long-term thinking, not quarterly panic
Open science ethos= transparency over secrecy
Multimodal, human-aligned systems= AI that amplifies, not replaces
Community-first design= co-create, don't dictate
They're not racing to be first.
They're building to be right.
A Word from BT
This is a rare inflection point.
The same minds that helped build the current AI giants are now breaking free, to create tools that feel more collaborative, more ethical, and more human.
It's not just a $2B funding story.
It's a reminder: We don't just need smarter machines. We need wiser builders.
And Mira Murati's team might just be the wisest of them all.
What's Next?
Three things to keep in mind:
Keep an eye on Tinker.
Watch how this shapes the future of fine-tuning at scale.
And ask yourself, what would you build if AI was truly yours to shape?