I met Ju Lin 7 years ago while on a project in Hong Kong. He was a programmer. We connected again during Covid when he told me of this AI project he’d been hired to do. It was through an Internet recruiter who had hundreds of code developers around the world working on this one project. It was funded by several VCs who remained secret.
The project did not have a name but was called Mantis AI by some of the developers because it was stealthy and captured data. The purpose was to glean information bits from everyday functions humans do.
The terabytes of data collected over time could be used to predict on a granular level everything from life span to job promotions or if you are likely to blow up a building or take an AR15 to school one day. More broadly, political groups, organizations, and Governments can be monitored for early exposure of nefarious outcomes. The data would be monetized to insurance companies, other Government agencies, and corporate HR throughout the world.
Software is no longer written from scratch but is built on Open Source libraries written by others available on Git repositories for free. From this, new software can be developed on powerful computer instances and automatically scaled to any size. The virtual disk image can be copied and placed in availability zones throughout the world by a 17 year old in his parents basement that would create a Borg like collective with an enormous reach.
Two or more developers were assigned to different disciplines such as facial or voice recognition depending on the number of sub-variants. For example there would be a variant for voice recognition that had been changed through voice distortion software and how to reverse it. The facial recognition did not just look at the face, but also the shape of the head, the body, arm movement and foot/leg movement patterns.