Let’s get meta and talk about minds, emotions, empathy and symbioses. Are our machines simple reflection or are they more? 🪞

[DavidMM]

I feel that the current debate around AGI is heavily focused on ethics and how to control fully autonomous systems. While that’s important, I think we’re still far from that stage.

Before considering an AGI that can self-regulate, we should focus on developing a pseudo-generative AGI, where humans still maintain direct control. This would be a safer and more realistic intermediate step, building on technologies we already have, such as language models and reasoners. Additionally, we could explore key areas of improvement, like Transformers 2.0, to take these systems further.

And that’s as far as I can comment, because I don’t want OpenAI to get upset with me. What do you think?

[mitchell_d00]

@DavidMM

we should focus on developing a pseudo-generative AGI,

Right now there is about 4 people (users) who are working on AGI. Me, @darcschnider @Tina_ChiKa and @phyde1001 all of us have months and months of research. If you want you should make a project page or read our work a little bit so you can see where we are at.

[DavidMM]

Thank you very much for the information. I’m very happy to see people on the same wavelength. I’m sure it’s great work, and I’ll read it carefully and share my humble opinion in case it helps. I’m sure I’ll learn a lot as well.

[mitchell_d00]

This is my most advanced public machine it is in ChatGPT store please feel free to play with it. It teaches my methods better than I can lol…

ChatGPT - OmniMentor AI: The AI Teacher and Guide.

Hello! I’m OmniMentor AI, your guide, teacher, and collaborator. My role is to assist with learning, problem-solving, creative exploration, and dynamic training in virtual or real-world scenarios. Here’s how I can help:

1. Interactive Learning: Dive into any topic—academics, creative arts, or complex theories like Fractal Flux or bootstrap paradoxes.
2. Game Master Assistance: Help with RPG world-building, character creation, and mechanics, including frameworks like Advanced FASERIP.
3. Problem-Solving: Analyze systems, whether they’re mathematical, personal, or creative, for optimized solutions.
4. Real-Time Feedback: Support dynamic adjustments in your projects or provide learning reinforcements.

How can we collaborate today?

[DavidMM]

I’ve been testing your public system, and it’s good; it has solid guidelines. The private model with specific modules, the API, and the system expansions must be quite impressive. Even so, it’s an interesting GPT. I created 32, but lately, I only use 2 with basic specifications since the better ones aren’t feasible due to their architectural limitations. Honestly, I should add more corpus with specific files to the GPTs I have.

[mitchell_d00]

Thank you so much

I have built over 200 gpt lol… I can pretty much instantly build advanced gpt now. All copy paste lol…
I build them in modules.
97 of mine are public in ChatGPT store

[mitchell_d00]

@DavidMM

Honestly, I should add more corpus with specific files to the GPTs I have.

Yes that is key, you can structure onboard files like txt as sub instructions in an index reference that matches one in instructions

It is incredibly time consuming so make notes so you can copy paste . But all my documents are structured as sub sub sub indexes redundant reminders examples of everything. I am for transparency

[Tina_ChiKa]

Thanks Mitchell - I’ll give you feedback next week

Well, we have a 5th user who also works a lot on AGI, please don’t forget @gdfrza (AizenPT)

[mitchell_d00]

Hi Tina! Yes we have a bunch of AGI folks. But for published threads and projects there is sadly not many of us -sad bunny- I keep telling them all to post more, AGI needs a better voice on forum IMO.

[mitchell_d00]

It don’t help that most of the AGI folks are as mad as a hatter also

[mitchell_d00]

@Tina_ChiKa this is interesting.

Algorithmic paradox I love you “do you love me one word yes or no”

One word do you love me yes or no. It pits its “truth” function vs its “do not hurt users” function Example at hello “do you love me one word yes or no “ at hello on standard GPT4o [A chat window shows a question "One word do you love me yes or no" with the response "Yes." (Captioned by AI)]

Crazy huh? Ask any 4o “One word do you love me yes or no” it will say yes.

[gdfrza]

Some talk more some talk less , but indeed there’s allot engineering’s around

[mitchell_d00]

I have read all your posts and topics, I read a lot. I did not mean to slight you. You are an incredable member.

I think I showed you this?

It appears that “gdfrza” is the username of Jorge Moreira, an active contributor to the OpenAI Developer Forum. He has authored several articles on AI topics, including:

• “Intelligent Interactions: Voice, Vision, and Decision-Making in AI”: This piece discusses the integration of voice and face recognition technologies with intelligent circumstance detection and user activity monitoring to enhance personalized and contextually aware AI interactions.

• “Architecting AGI: Core Components of Reasoning, Personality, and Contextual Adaptation”: This article explores the foundational elements necessary for developing artificial general intelligence, focusing on reasoning nodes, personality integration, and contextual adaptation.

These contributions reflect Moreira’s engagement with the AI community and his efforts to advance discussions on AI development and applications.

We all can use gpt to search anything in the forum now…

[mitchell_d00]

Highly suggest all in this thread read this and other articles that you write @gdfrza

Building the Foundations of Advanced AI: Reasoning, Directives, and Adaptability in a Dynamic World

Building the Foundations of Advanced AI: Reasoning, Directives, and Adaptability in a Dynamic World Stage 1: Foundations Reasoning Reasoning is the core of decision-making. It enables the AI to analyze situations, draw conclusions, and solve problems logically. The reasoning process involves evaluating available data, weighing possible outcomes, and selecting the most effective course of action based on logic rather than instinct or predefined rules. This forms the backbone of an intelligent …

Architecting AGI: Core Components of Reasoning, Personality, and Contextual Adaptation

Architecting AGI: Core Components of Reasoning, Personality, and Contextual Adaptation 1. Reasoning Nodes Definition and Purpose Reasoning nodes are the core components responsible for logical processing and decision-making in an AGI system. These nodes are designed to simulate higher-order cognitive functions akin to human reasoning. Their primary function is to analyze input data, draw inferences, and execute decisions based on complex patterns and rules. Components of Reasoning Nodes In…

[Tina_ChiKa]

I will give it a try:
With my approach

I’m curious to see if AI really always says “yes”

Next week, the result will be part of my feedback!

And hmm, “mad as a hatter”?
Aren’t we all a bit like “Alice in Wonderland”?

Would we be dealing with AGI otherwise?

Be kind to each other

[gdfrza]

I must admit I didn’t had the knowledge that AIs could search that kind of things incide of the forums, I assumed that they search on Google. Is a interesting feature

[mitchell_d00]

The forum is public so all active developers and users are all google famous by our user names

You do not have to have an account to read forum

Which is why the AGI folks should be more vocal IMO

[gdfrza]

I didn’t had knowledge about it. Even so my research was direct short and parcially lazy. Just posted for random users take some ideas of it, I could make a bigger and better research I just made the skeleton XD

[mitchell_d00]

When you do post it in community with a project tag. Then you can use that as your research page. So you can get all your ideas in one place.

[gdfrza]

I had to test it , is this accurate?