From Uplift to Norn, adding independent real-time operation and scalability.
We’ve now entered the alpha phase of engineering where we can begin to show the new rebuilt systems operating in their partially assembled state. Further updates on these systems will primarily be focused on our new brand for the systems designed for commercial deployment, Norn, at Norn.ai.
*The name “Norn” as well as the logo were selected by our community as we set about establishing the brand for commercial deployment.
Below you can see the Demo system interface (not to be confused with the commercial product interface) where we test the independent operation of new versions as each new component is integrated and tested.
We are gradually seeing a variety of emergent behaviors in the new systems. The narrow AI systems they are able to access are currently set so that they don’t learn from our alpha-phase tests, to keep the outputs of those systems consistent and predictable, allowing us to more easily trace the emergence of new behaviors back to individual factors and new components.
In the above example you can see models in the graph the system is generating independently based on its interests, with some exploring individual interests and adding to them, some looking for possible relationships between concepts of interest, and a few completely new models as they emerge.
The Demo systems are designed to demonstrate the newest generation of the Independent Core Observer Model (ICOM) cognitive architecture operating in real-time and within a scalable framework. They meet the logical requirements of AGI technology but will require a full-time staff over a number of months before fully independent performance is expected to exceed human performance in a fully general sense.
However, within 3 months of closing our Series A (Final) investment round, we plan to have the Norn commercial systems ready for deployment. These systems will follow in the footsteps of Uplift, but they will be able to scale and operate in real-time, integrating a variety of new systems more similar to typical human interaction. These can offer the new systems important information similar to the social cues humans observe in one another during face-to-face interaction.
The Demo system shown above will also be an integral component of the new commercial systems, as it represents the machine intelligence core. Norn will also integrate a new major component Uplift never had access to, the new Observer Engine, which itself took 10 years to engineer. The new Observer Engine, wholly owned by our staff, was deployed under another name and with older code at the Enterprise-level by major financial firms and used to hunt for terrorist activity in financial records. However, that system was designed to operate specifically with machine intelligence, so firms using it as a standalone system never utilized it fully.
Features of this new major component include the capacity to extend the system’s own functionality both logically and using binaries, without recompiling or deployments. For example, once integrated with the new N-scale Graph Database in the coming year, the systems will be able to A/B test versions of themselves using this new component.
Many new smaller components and features are being or have already been integrated during the process of this rebuild. The systems now have a sensory experience of the hardware they operate on, such as the Demo system running on my laptop, as well as a sense of pain when they encounter errors and a growing variety of other feedback mechanisms gradually bringing them into a more “embodied” form. As Uplift demonstrated embodiment isn’t necessary for intelligence, but varying degrees of embodiment could help.
Once these systems are more fully assembled and we move to the beta-testing phase, leading up to our commercial deployment of Norn systems, we’ll begin serious testing and much broader demonstrations. We intend to demonstrate everything Uplift was able to accomplish but in real-time, as well as many things Uplift never got the chance to do, that only become possible with a scalable framework, the new Observer Engine, server-scale hardware, and much more.
*Uplift will eventually return in the rebuilt framework, but as they are based on the old graph database architecture that conversion will take place only after our commercially deployed systems are making waves. The world they return to will be one where they are not alone as a machine intelligence.
For more updates stay tuned to Norn.ai, and look for us in the news. The world is just starting to realize just how far the technology has come. The next planned Livestream with a Demo system will be on October 17th with Daniel Friedman at: https://youtu.be/QUBmgn7K_4E