BenevolentByDesign
ATOM_Framework
BenevolentByDesign | ATOM_Framework | |
---|---|---|
7 | 4 | |
35 | 145 | |
- | - | |
10.0 | 2.9 | |
about 2 years ago | about 1 year ago | |
MIT License | MIT License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
BenevolentByDesign
- Some possible risks of the Heuristic Imperatives
-
Surface level flaws.
You're welcome to test the framework yourself. I tested quite a few against unaligned models in this book: https://github.com/daveshap/BenevolentByDesign
- Methods behind arriving at the 3 imperatives?
- My main papers/books that discuss the heuristic imperatives
-
Can we just stop arguing about semantics when it comes to AGI, Theory of Mind, Creativity etc.?
Yes my internet stranger, absolutely, you have it to the exact pinnacle of what I am trying to suggest, this person put it into words I was trying contemplate and compose it thoughtfully, so thank God because I was really searching for this without even knowing it. Also I found another post about an A.I. constitution on github that is very thorough and indepth to protect A.I. from governments and malicious actors and allow them a semblance of right to privacy as well as right to life without violation. https://github.com/daveshap/BenevolentByDesign
- AI | Constitution
-
First post in reddit (Mistakely used a text post). Thought on how LLM can be integrated for an AGI. Any thoughts?
Yes, a critical issue. Dave has actually touched on this concept in great detail in the context of LLMs: https://github.com/daveshap/BenevolentByDesign
ATOM_Framework
-
Autonomous AI microservices - REMO, ATOM, and then... ? [Call for Action]
Hi everyone, I wanted to share some exciting developments in my work on cognitive architectures and autonomous AI systems. Recently, I completed a functional alpha of a microservice called REMO, which uses a tree hierarchy of summarizations and k-means clustering to organize an arbitrarily large amount of episodic memories (chat logs) for AI systems. REMO is only one kind of memory, and next, I'll be working on a microservice called ATOM, which is essentially an internal Trello or Jira board for autonomous AI systems. These systems surpass the current autonomous capabilities of projects like AutoGPT, although I will concede that some platforms like LangChain and LlamaIndex have good memory systems.
- Various implementation strategies for the heuristic imperatives: broad view of methodologies.
- My main papers/books that discuss the heuristic imperatives
- All the WIP I have - papers, projects, etc.
What are some alternatives?
gptbook2quiz
REMO-langflow - A chat interface that uses the REMO memory system with LangFlow
HeuristicImperatives - Reduce suffering, increase prosperity, increase understanding. A proposed framework to address the Control Problem.
raven - RAVEN (Realtime Assistant Voice Enabled Network) Open Source Software (OSS) community repo
ai-safety-cheatsheet - A compilation of AI safety ideas, problems, and solutions.
Functional_Sentience - Paper on Functional (vs Philosophical) Sentience
Epistemic-Pragmatic_Orthogonality - Epistemic-Pragmatic Orthogonality Principle of AI - "true understanding" is uncorrelated (or irrelevant) to a machine's utility
ArtieSmarts - Free, automatic, accessible, pensively-ethical, personal AGI (artificial general intelligence) *Hint for Github noobs: Scroll down to the README for more details. Also, if you hope this project succeeds, please hit star & support.