Bridging Realms: Navigating the Present and Future of AI and the Metaverse through Philosophy & Mindfulness
The Big Picture: How AI, Philosophy, Mindfulness & the Metaverse are Connected
What does AI, philosophy, mindfulness, and the metaverse have in common? Why am I speaking and focusing on these topics, and how are they interrelated?
Let’s jump in.
We will start with AI.
Artificial Intelligence & LLMs
I have been in the field of Conversational AI since 2016. The field has come a very long way since then.
The big breakthrough came when the Google research team wrote the infamous paper "Attention Is All You Need" by Vaswani et al. which led to the emergence of the Large Language Models (LLMs) we are seeing today.
In the old days, creating Conversational AI Agents, took a lot of meticulous work. Conversational AI agents were using NLP and NLU, in which a sentence would be broken down into intents (verbs) and entities (nouns).
The conversational designer would then have to map out intents (e.g., buy pizza) and consider all of the ways a user would ask to ‘purchase pizza’ and all of the other related intents and entities (e.g., buying soda) within the context of the use case.
Creating bots this way was time-consuming, expensive, and addressed a very limited set of questions.
LLM solved these problems.
What is an LLM and How does it Work?
A Large Language Model (LLM) is an advanced AI system designed to understand, generate, and interact with human language at a large scale.
It's trained on massive datasets of text and learns the statistical structure of the language, including how words tend to relate to each other and form meaningful sentences.
LLMs are not knowledge bases, and they don't explicitly store information; instead, they store and learn patterns of information.
When you ask an LLM a question, it reads the entire text and turns your query into tokens and vectors. After identifying the patterns, it forecasts the subsequent token in each pattern by taking into account the context that each of the prior words has provided. This process continues until a full response is generated.
And in a nutshell, this is how LLMs work.
This solved many of the problems and challenges of previous NLP, NLU, and rule-based Conversational Agents but created a new set of problems.
The New Challenges LLMs represent
When we start exploring AI and LLM, we quickly find ourselves in a very strange place. On the surface, these systems seem easy and straightforward to understand.
We project our confidence on them and assume that all is going well, but underneath the hood, there is a lot of complexity. As soon as you start working with LLMs, you’ll notice this right away.
The challenges of working with LLMs:
Answers can vary quite a bit based on how a users poses a question
It is hard to get the model to perform accurately and predictably
Models don’t actually know anything and they don’t have any information
There is Bias in the Models
The models can’t tell the difference between real and fake ontologies
To address these issues, we created a number of techniques, including:
Prompt Design, Engineering & Tuning
RAG and Knowledge Bases
Fine-Tuning The Model
Pre-Training
In essence, we are now in a place where we have to work with data, information, and knowledge. We have to consider how we organize data and information into ontologies, taxonomies, and semantic relationships that give rise to meaning.
And with this, we are quickly coming into the world of philosophy.
Philosophy
At its core, Philosophy deals with the nature of being, what existence is, what’s true, the nature of knowledge, our values, reason, mind, and language, seeking to understand and provide insights into the nature of reality and how one should live their life.
Ontology is a branch of philosophy that studies the nature of existence, reality, being, and becoming.
In data science, Ontology is used to structured frameworks that define the relationships among a set of concepts and categories in a domain. They are crucial for organizing information systematically.
However, LLMs do not know what is real and what is not, what are real ontologies and what are fake ones.
All LLMs know are the relationship patterns within their structure. LLMs then make predictions based on these patterns, which can give rise to new knowledge.
On the surface, what this means is that it is up to us to investigate what is real, identify these relationship patterns, and then train the AI accordingly.
And this is the beginning of our trouble.
We take it for granted by assuming that the world is basically a set of facts, that these facts represent themselves to us, and that we know the world.
This aligns with the common-sense intuition that the world is as we perceive it.
This philosophical point of view is Direct Realism and it is assumed to be true by many learned men.
In spite of its common-hood, over the past 100 years, this philosophical approach has been falling apart.
Here is how:
Problem of Too Much Information: There is too much information in our environment, and we can’t observe everything. So, instead, we focus our attention on what we believe is valuable and ignore the rest. As a result, it is really easy to miss essential information and to get fooled. Watch a simple magic trick, and you will know what I mean.
Sense Organs: Our sense organs are limited. You only see what is in front of you, and you only see about.003% of the visible light spectrum. That would be like walking into a room with 300 objects and only seeing one of them.
Perception: This is the process of organizing and interpreting sensory information and giving meaning to it. Our personalities, prior experiences, internal states, and preferences all influence perception. For example, one person may perceive a dog jumping on them as a threat, while another may perceive it as the dog being excited to see them.
Perspective: We all see things from a single point of view in time and space within a given context. We have a very small slice of information we are dealing with at any moment in time, but assume to actually know the things we are looking at.
Bias: Humans have over 180 well-known and documented cognitive biases, which are irrational and distort our perceptions.
Culture: Culture defines and interprets a lot of social information for us. For example, shaking your head left to right means ‘No’ in almost every country except Bulgaria, where it means ‘Yes’.
Language: Language allows us to think, talk about, and experience things that would not be possible without the right words. The right words help us perceive! For example, notice the difference in how a wine tastes before and after it is described by a sommelier.
Values: As we scan our environment, we look for and move towards the things that we value. Our values act as information filters.
Memory & Meaning: We don’t remember things as they happened even a few seconds ago. We forget most things very quickly and only really notice and remember information that is meaningful for us.
Beliefs & Assumptions: Our assumptions are like hidden beliefs about the nature of the world, who we are, and our relationship with the world, and this has a philosophical foundation. Our beliefs then act like guides that move our behavior forward towards what we believe is valuable and possible for us.
Now, we have come full circle. The world is a lot more complicated than simply looking at it and perceiving it as it is. It has three components: the subject, the objects, and the interpretation.
We don’t perceive the world as it is, but as we are.
And this sheds light on our differences, on our conflicts, and why it is so difficult for people with different beliefs, values, or cultures to agree on anything.
Often, it feels like we live in a different world because we do.
None of us live in ‘the world'; instead, we live in ‘our world’.
In some ways, this is the core of our problem, and one possible solution is mindfulness and self-inquiry.
Mindfulness
Mindfulness is the process of bringing your attention back to the present moment and connecting it to your body's experience of reality.
Mindfulness is like a superpower.
It allows you to begin to observe your experience as it actually is instead of how you believe it is.
It helps you become aware of your thoughts and feelings.
How your thoughts trigger your feelings, how your feelings trigger your imagination, and how this leads you to action.
Until you watch this process play out, you will believe that you are the thinker. That voice in your head that is reading these words and providing commentary is you.
Yet, when this thought goes, will you go with it?
When the feelings pass, do you pass with them?
No, you are still here.
When you start watching your thoughts and feelings, you turn them into objects that you are observing.
As objects, you will see them come and go.
Watching your thoughts as you would watch an object helps you untangle your identification with them.
After all, you can’t be a thought if you are also the one observing it.
Practicing Mindfulness will help you with:
Become more self-aware
Recognize your patterns
Notice your own Cognitive Biases
Identify your values and inspect them
Overcome limiting beliefs and fears
Become more aware of your hidden assumptions
Become more aware of your Identity and self-concept
Notice your perception
Mindfulness can untangle many of your assumptions and core beliefs and make you aware of your identity, which extends beyond your ideas about yourself.
Mindfulness can help bring us back down to earth, which will make it far easier to address some of the philosophical questions that AI poses, such as:
Understanding of Language and Meaning: Exploring the Nature of Understanding and the Relationship between Language and Meaning.
Consciousness and Artificial Intelligence: The sophisticated capabilities of LLMs to mimic human-like text generation spark discussions about whether these models can be truly conscious or possess understanding, or if they are merely advanced tools without awareness. These questions arise because the majority of people don’t know the nature of their own consciousness and assume that it is related to thinking, language, and the mind.
Ethics and Responsibility: AI raises ethical questions regarding bias, the attention economy, and other types of misuse. There is a profit motive to make AI assistants that are deeply personalized and have the ability to suck up our attention and sell us products.
Epistemology: What is knowledge? Even though LLMs don’t have any actual knowledge and are instead statistical representations of knowledge, they are able to make valuable predictions that add new knowledge.
AI and Human Identity: LLMs also prompt philosophical reflection on human identity and self-worth. If our values are tied up to our intelligence or results in the world, what happens when an AI can do our job better than us? AI invites philosophical inquiry into the essence of human uniqueness.
The Metaverse & Technological Convergence
Last but not least is the Metaverse. I am using these terms to represent Augmented Reality, Virtual Reality, Spacial Computing, Web3.0 (NFTs, Blockchain & Crypto) and any other terms that the industry invents.
I believe that over the next decade, all of these technologies will merge with AI and the internet.
We will be able to go to a 3D-embodied version of the internet by putting on a pair of glasses and one-day contacts and having a fully immersive experience.
We will be able to meet friends there, talk to AI agents, and have unique experiences.
Imagine being able to live and experience your favorite movie or video game from the inside out!
Very niche communities will spring forward and become 3D meta-representations of their members’ inner worlds. Many people might not even have friends that live in their city and engage with others mostly through the screen.
As core needs aren’t fully met, this can lead to deeper disconnect, more loneliness, and the proliferation of synthetic relationships to address loneliness.
All of this will make us question our reality even more. It will bring more philosophical questions to the forefront.
On the bright side, this technology can lead to a revolution in self-awareness. It can help us become more aware of the role our senses play in constructing our experience of reality. As we are able to experience life through another person’s eyes, it can also help us become more empathetic and compassionate and to one day see ourselves as one humanity.
Join the Journey
Join me as I journey and navigate through these waters.
Over the next few weeks and months, I will share my insights on AI, philosophy, and the results of my experiments in implementing the best versions of these technologies.
Indian philosophy has lot of answers to these realities
Hi Stefan, it's a good read. Thank you.
I am wondering if you may like to read this article I wrote on a similar topic.
https://medium.com/@soharab.hossain/wisdom-meets-tech-elevating-humanity-unleashing-the-power-of-ai-infused-wisdom-82a9f14ac5da