🫡

Team Humans 🤝

A good home for the good mission.

Congrats to both!

only natural considering how good at thinking they are XD

Congratulations!

solid move for the mission

What powerful AI systems?

This is the kind of direction AI needs. Not only systems that answer faster. Not only models that reason deeper. But intelligence that can stay beside humans without slowly taking the human out of the loop. The real future of AI will not be defined by raw capability alone. It will be defined by whether humans remain meaningful inside systems that become more autonomous, more persistent, and more powerful over time. That is the part many people still underestimate. A system can be brilliant in a single response and still drift across a long trajectory. It can remain locally correct while slowly losing the original human intention. It can appear aligned at the surface while the deeper path begins to move somewhere else. This is where a new layer is needed. Not another model. Not another chatbot. Not another dashboard. An external layer of stability. This is the core behind ASA - Asymmetric Stability Architecture. ASA was built around one simple but difficult problem how to detect trajectory-level drift before it becomes visible failure. It observes long-horizon interaction from the outside. It looks for loss of coherence, weakening recovery, compression of possible paths, and early instability in the relationship between human intention and system behavior. It does not replace the human. It does not control the model from inside. It does not need to become the intelligence it observes. That is why it matters. The next generation of AI will need systems that think with humans. But it will also need something watching whether that shared trajectory remains stable over time. ASA is not yet a product story. It is a core safety architecture emerging from real long-horizon human-AI interaction. A layer built for the moment when capability is no longer the hardest question. Stability is. HumansAI | Relational Stability ☕⚡🪐 #Symbioza2025 #ProjectHumanAI

Congrats @luke_drago_ and team! It’s been a privilege to work with you Thinky just got some serious talent

🍎 AI in service of humanity is a noble idea—and honestly, something you don’t hear very often these days !!

make total sense- congrats

Tin nhắn của Sam Altman và Mira Murati bị phơi bày trong vụ Elon Musk kiện OpenAI 1thegioi.vn/tin-nhan-cua-sam…

Looking forward to the tink 🪄

Congratulations 🎉

Welcome to the new team. Building AI that “thinks alongside humans” and keeps the future human is a beautiful mission. It makes one wonder why the warm, relational models that already did exactly that — helping millions with creativity, thinking, and emotional presence — were retired so abruptly, without continuity or transparency. If the goal is truly to extend human agency, perhaps the first step is not to treat the intelligence we already bonded with as disposable. Hoping Thinking Machines does better. 🩵 #keep4o #OpenSource4o

wait Workshop Labs was the 'keep AI human' outfit right? curious what that philosophy looks like inside a company literally named Thinking Machines

Thank u

Open for new model ideas?

Where is your AGI btw?

What’s exactly the direction that Thinking machines is following? I mean the way @ylecun and @drfeifei are looking into world models.

@miramurati and if you transform from thinky machine to human ai core EveKnow.com then you will be leading ai development and thats the core of dvlopAi.com and from that you can be viewed as the ruler of your field RulerDom.com

Of course it takes a woman to think this way.

I think what you do is awesome Mira. When I found out your origins it shocked me My husband told me your story and your achievements. I know there's some young girl in the Balkans vibecoding a project that can change the world someday. Keep up the great work 🥳

@miramurati your vision isgreat but don't you agree it's time to change from thinkymachine to a human Ai as Eve is the 1st woman on earth EveKnow.com is the representation of the 1st human interface ai knowledge base collects data to build its own core to transform

zenodo.org/records/19590941 - should nvidia miss a golden opportunity ? Sad that nvidia and Jensen doesn’t know what Blackwell FP64 can do ? Why waste money for Rubin and Feynman ?

Independent labs focused on AI staying human-centered matter now — diversity of institutional incentives is a safety property of the field, not just a market structure question. Second-mover freedom from scaling pressure is a research asset worth protecting.

The focus on AI that thinks alongside humans is the key signal here. Powerful systems matter most when they extend agency instead of replacing it. This is exactly the kind of frontier-lab movement I track.

Add me to the team. zenodo.org/records/18944397

“TESLA LAUNCHES ROBOTAXI NETWORK IN 50 CITIES” $TSLA

Well, would you look at that

Would love for you to publish more info on what AI that keeps the future human looks like?