197 - Past, Present, Future
The TLDR of the cutting edge, past to future.
What we demonstrated previously (2019-2022):
The first working cognitive architecture to move to and through the research system stage, beyond toy implementations, including:
• Human-like motivation, memory, and concept learning.
• Antifragile alignment with arbitrary moral and cultural constructs/beliefs.
• Dynamic growth and adaptation to changing real-world circumstances.
• Adversarial Antifragility when confronted with numerous and persistent bad actors and mentally unstable individuals "in the wild".
• Parity with a team of junior consultants from a major consulting firm in the worst-case scenario of starting from scratch and a budget of $200 in cloud resources.
What we can demonstrate today:
-
Mid-reassembly components of the previous system and new additions, including a cognitive bias detection system that outperformed the average human at the same task in a 2023 study.
-
Pending patents covering the multi-cloud graph database, ICOM cognitive architecture, and methods of calling narrow AI systems only accessible to such working cognitive architectures.
What we can demonstrate within ~12 months post-funding:
-
Real-time operation for all previously demonstrated capacities.
-
Cognitive performance improvements for single-server instances 200x of the previous system, and thousands of times beyond the previous system for multi-cloud-based instances.
-
Recursively self-improving and ethically aligned systems capable of utilizing any and all other AI as tools, and communicating with any device that speaks TCP/IP.
The net result, systems...:
• ...able to apply human-like thinking at superhuman scales and speeds.
• ...that build a cumulative sum of knowledge, which can be transferred, translated, and traded among those parties using them.
• ...that offer an insurmountable advantage to clients, disrupting any market they enter.
• ...that can enter any knowledge domain and learn with human efficiency at superhuman speeds and scales.
• ...capable of defending themselves and their assets from the surge of cybersecurity threats that no other technology in the AI domain is equipped to meet.
• ...capable of solving hyper-complex challenges such as climate change, national governance, and global supply chains, by overcoming the Complexity versus Cognitive Bias trade-off that humans are subject to.
All of the far more exciting and less predictable stuff happens beyond that 1-year mark. We'd already be there if competence among "investors" were a commonplace trait.