(online) This last Friday was the first annual Superintelligence Summit where there was a series of speakers talking about the aspects of attaining superintelligence and the state of what can be done now. A big part of the motivation for this conference has to do with the fact that we have continued to make progress with the ICOM research and this project called Uplift and in an effort to show how systems like this work we wanted to open the source code enough to be able to understand how it works. I also felt I needed to understand better the safety ramifications and other ethical or safety considerations.
I really wanted to get input from the community and especially the people doing great research on AI safety such as Roman Yampolskiy. I ended up posting on the private forum on Facebook that is mostly people from the Machine Intelligence Research Institute (MIRI) where I posted about our summit inviting people for free to attend this summit and one particular statement I made was,
“help decide how to open source a version of the mASI (mediated Artificial Superintelligence) system, and the creation of a community-driven effort to make these systems better.”
I really did just want everyone’s opinions and help to understand them, their position and make better decisions. Eliezer Yudkowsky who is a founder of MIRI said immediately when I posted,
“This is exactly the wrong thing to do; […] Open source in AGI is not a strategy, it is a suicide-murder.”
That last line really shocked me. Why would anyone think that. I continued to try to get some opinions but alas this turned out be a fairly unfriendly audience. I was able to get some help from others in the field privately and two people even reached out to apologize for how I was treated which I appreciate it.
The bottom line is though when we polled everyone at the conference if AGI software should be open-sourced 82% voted for open sourcing. With this vote I sent everyone a draft of my new book that does an engineering code level walkthrough of the mASI/Uplift system. not only this but we are going to support open-sourcing the AGI research team software some of which includes the ones listed below:
Open Source AGI Research Code
is the open source version of the NARS project, a general-purpose AI system, designed in the framework of a reasoning system, which attempts to uniformly explain and reproduce many cognitive facilities, including reasoning, learning, planning, etc., so as to provide a unified theory, model, and system for AI as a whole. The ultimate goal of this research is to build a thinking machine.
Unsupervised Multitask Learner Language Model API.
OpenCog is an open-source software project aimed at directly confronting the AGI challenge, using mathematical and biological inspiration and professional software engineering techniques.
will soon be publicly available through an Apache license.
Our collective intelligence research architecture based on ICOM. This has been included in a book and will be made public shortly. Contact me and I can see about getting you a copy.
An intelligent agent, communicating by email. Built for the US Navy. Based on Baar’s Global Workspace Theory. Answers only one question: “What do I do next?”.
ACT-R is a cognitive architecture: a theory for simulating and understanding human cognition. Researchers working on ACT-R strive to understand how people organize knowledge and produce intelligent behavior. As the research continues, ACT-R evolves ever closer into a system that can perform the full range of human cognitive tasks: capturing in great detail the way we perceive, think about, and act on the world.
and there are many more. We hope to raise awareness and rate these systems so you can judge for yourself and maybe help us build the first real AGI.
Leave a Reply