VRUnderground's banner

VRUnderground

@VRUnderground • 2,985 subscribers

Promoting VR culture from the Underground. https://t.co/LyFhN8ZaHD https://t.co/OKMNpK2XrB #VRChat / #VRUnderground / #ChillOutVR

Shorts

I was working through Steam Early Access requirements for VRUnderground and I had a really spooky realization. The whole conversation around training images and videos is not the important part. The media itself is just a bridge. What is actually being extracted is cognitive patterning. Models learn how people think, not just how they draw. The images are only the residue of the process. The real map is of the minds that created them. It reminds me of poker. A great poker player is not "stealing your cards". They are reading the patterns in your behavior and mapping your decision space. In that sense, training AI on human-created data is not about automating tasks. It is about mapping cognition through metadata and behavior trails. If that is true, then the internet starts looking a lot more like a poker table. And nobody wins a hand against someone who can read their mind.

I was working through Steam Early Access requirements for VRUnderground and I had a really spooky realization. The whole conversation around training images and videos is not the important part. The media itself is just a bridge. What is actually being extracted is cognitive patterning. Models learn how people think, not just how they draw. The images are only the residue of the process. The real map is of the minds that created them. It reminds me of poker. A great poker player is not "stealing your cards". They are reading the patterns in your behavior and mapping your decision space. In that sense, training AI on human-created data is not about automating tasks. It is about mapping cognition through metadata and behavior trails. If that is true, then the internet starts looking a lot more like a poker table. And nobody wins a hand against someone who can read their mind.

26,437 views