Architecture for Artificial General Intelligence
“The only way to win this game, is not to play at all”
It breaks my heart to join those who argue against open-source architecture for Artificial General Intelligence (AGI).
I have come to this position after much thought, internal debate and seeking the opinion of others.
The creator inside of me wants to build AGI-like AI systems, and I have and do with the work I do with Perceptible.AI.
The academic inside of me wants to share knowledge.
The child inside of me wants to be and share excitement with new technology.
The competitor inside of me wants to demonstrate prowess.
The businessperson inside of me wants to have marketing leverage of demonstrable aptitude.
The cautious and careful side of me thinks heavily of the community.
The safety side of me thinks of self-preservation.
The pragmatist inside of me say that a machine with free will is useless to me.
The realist inside of me says that Large Language Models, even today (2024), are very capable systems as they are for a vast array of autonomous agents and applications not yet thought of or successfully marketed.
