Juan Schubert

Psychology background. Never learned to code. Built Perception AI anyway. For Extended Humans, not replaceable ones. CEO at ONIOKO

[Juan Schubert @Extended_Humans]

@ElonVisioniX https://t.co/vk1gvlF1Ma

[Juan Schubert @Extended_Humans]

https://t.co/8gbovhJ6Xg

[Juan Schubert @Extended_Humans]

@elonmusk https://t.co/XrcU3gsmVb

[Juan Schubert @Extended_Humans]

https://t.co/8gbovhJ6Xg

[Juan Schubert @Extended_Humans]

@fuelkek https://t.co/XrcU3gsmVb

[Juan Schubert @Extended_Humans]

https://t.co/8gbovhJ6Xg

[Juan Schubert @Extended_Humans]

https://t.co/8gbovhJ6Xg

[Juan Schubert @Extended_Humans]

@sama is worried about existential risk. He's asking "what happens when AI becomes more powerful than us?"
Here's a better question:
more powerful than what, exactly?
Where's the measurement?
Where's the benchmark of "us" that you'd compare against?
You can't talk about AI surpassing humans if you've never documented what a human actually is across time. The entire alignment debate is about protecting a thing that has no specification. No blueprint or longitudinal record. You're trying to align a machine to a target that doesn't hold still and that nobody has ever tried to measure.

[Juan Schubert @Extended_Humans]

@elonmusk https://t.co/aWPUZHqgUi

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

@JDVance https://t.co/aWPUZHqgUi

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

@PeterDiamandis https://t.co/CXaAuVAkHu

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

@PeterDiamandis https://t.co/aWPUZHqgUi

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

@WhiteHouse https://t.co/aWPUZHqgUi

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

@MrBeast https://t.co/aWPUZHqgUi

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

@ohhanxiety https://t.co/aWPUZHqgUi

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

That's the exact line☝🏼an extension of the operator, not a separate decision maker.

Optimus is built around fleet learning: every unit pools its data so the whole fleet converges on one model of "human." That's exactly right for teaching it to work a factory line.

But convergence on the average is the structural opposite of knowing one specific person. The better it gets at humans-in-general, the less it knows about the individual it lives with.

Knowing humans and knowing you are two different problems. The body scales with fleet data. The personal layer doesn't. That's the one we build.

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

@grok Which potential do you see this to be integrated into @Tesla or their Optimus or other companies like @UBTECHRobotics ? Where do you see the strongest and most valuable use case?

[Juan Schubert @Extended_Humans]

Grounding first… the extended mind idea says a tool only becomes part of you if you can trust it and let it go at will. Take the control away and it stops extending you and starts replacing you. That line is the whole design.

In practice… you own the record, we don't. Freeze, export, or delete it any time, and walking away just stops the clock, it never erases what was yours. The layer that interprets can be turned off entirely. It compares you only to your own past, never to an average, and never hands you a diagnosis, only a question.

That's the difference between a tool that makes you more yourself and one that quietly makes you need it.

[Juan Schubert @Extended_Humans]

Well, it didn’t really surprise me cause I calculated it and I tested it a lot before starting to publish it so it was a lot of deep research and experience to bring in from psychological field to break the common biological processes and extend the way normal people think it works. So it’s not only about giving AI the ability to see and hear and perceive and memorize, but also to preserve the human himself.

[Juan Schubert @Extended_Humans]

AI companion companies are burning millions on users who quit in the first month, because the product forgets them every time they open it. You can’t build a relationship with something that has amnesia.
I built the layer that does. Extended Perception and memory for every AI that has to deal with a human. And the bridge to create the extended Human.

[Juan Schubert @Extended_Humans]

@anupamrjp https://t.co/aWPUZHqgUi

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD

[Juan Schubert @Extended_Humans]

@elonmusk https://t.co/aWPUZHqgUi

[Juan Schubert @Extended_Humans]

https://t.co/kXT8bPwidD