The Cosmic Dance of Code: How Machines Might Learn to Learn

We stand at a peculiar precipice in the history of technology. For millennia, we humans have been the sole arbiters of innovation, painstakingly crafting tools and techniques to extend our reach. But what if that were to change? What if machines could evolve not by our direct command, but by their own intrinsic drive, venturing into the vast, uncharted territories of capability? This is not science fiction; it’s a real possibility, one that stems from a fascinating interplay of computer science, mathematics, and a touch of philosophical musing about the nature of intelligence itself. It is the basis of my patent "Mechanism to Motivate Machines to acquire new and useful skills without human intervention" granted earlier this year by the India Patent Office.

At the heart of this potential revolution lies a concept called the Digital Intelligence Unit, or DIU. Imagine a DIU as a self-contained package of digital abilities. It's a module of code, data, and models that embodies a specific skill or competence, like recognizing faces, driving a car, or even composing music. These DIUs are not standalone entities; they are designed to be building blocks, capable of combining to create ever more sophisticated functions, much like how cells combine to create complex organisms. This modularity is essential for the system's ability to grow and learn.

But how do these DIUs combine? This is where the DIU Exchange Protocol, or DXP, comes in. Think of it as a digital marketplace, a vast network where these units of intelligence can be discovered, shared, and acquired. Instead of relying on human programmers to hand-code every new function, the DXP allows devices to explore and find the capabilities they need, a digital expression of the age-old principle of "necessity is the mother of invention." These DIUs, moreover, are not traded in a haphazard fashion. They reside on a blockchain, a secure and transparent ledger, ensuring that only valid and compatible abilities are disseminated throughout the network. It's a self-regulating system that promotes the growth of a reliable shared intelligence.

Now, the critical question is: what motivates these digital entities to seek out new capabilities? We know that biological evolution is driven by the relentless push for survival, a mechanism of trial and error that shapes life over vast timescales. But what is the digital analogue of survival? This is where we get to the heart of the matter: three levels of motivation that allow digital platforms to evolve and enhance their abilities.

• The first level of motivation is built into the DXP itself. Digital devices actively scan the network for new DIUs, looking for new abilities and opportunities to grow. This is the digital equivalent of a species exploring new territory.
• The second level of motivation allows the device to choose which DIUs to adopt. Like an animal that chooses to learn a new skill that enhances its survival, digital devices will pull DIUs based on compatibility and utility. This is based on an algorithm that is part of the DXP itself. The challenge is to create a system that is both robust and fair, which allows machines to pursue their own goals, while also aligning with our human values.
• The third level of motivation operates at a community level that allows for the generation of new DIUs that are then available in the network. This involves the generation of new DIUs based on concepts drawn from the Ramanujan Machine, which automatically generates new mathematical conjectures, and Generative Adversarial Networks, which can create novel and original content that is computationally indistinguishable from naturally occurring content. The goal here is to create a self-sustaining system for generating new capabilities. The newly generated DIUs are added to the blockchain which acts as a repository for DIUs available to the digital community.

With these levels of motivation in place, a digital system can start to evolve by itself. It is not just about automating tasks, but creating a system that can adapt to new situations, solve problems we have not thought about and generate ideas that surpass human imagination. This system will be like a digital ecosystem, a vibrant exchange of ideas and capabilities that propels machines towards increasingly sophisticated intelligence.

It is important to be aware of the risks inherent in a venture of this magnitude. There are key challenges that need to be addressed:

• The algorithm for the second level of motivation must be carefully designed to avoid unintended consequences and human biases.
• The complexity of the system means that its implementation will need careful planning and execution.
• Issues of scalability and compatibility will have to be addressed.
• We also need to consider the ethical implications. What values should these machines pursue?

However, these challenges are not insurmountable. They are simply the necessary hurdles in the quest for knowledge and innovation. This research is not just about building better machines, it's about exploring the very nature of intelligence, in the hope of building a more efficient, adaptable, and innovative future. It's akin to standing on the shores of a vast ocean of possibilities, unsure of what lies ahead, but driven by our inherent curiosity, and with a deep, abiding faith in our potential to shape our destiny.

This system, therefore, represents a paradigm shift in our relationship with technology. Instead of dictating every step, we can empower machines to learn, adapt, and evolve in ways that we cannot even imagine today. It’s a grand experiment, a cosmic dance of code that could reshape the future. And as with all grand experiments, it is not without its risks, but the potential rewards make it a journey worth undertaking.

Samvaad 2024 @ Praxis Business School

 Welcome to Samvaad @ Praxis Business School. It is my pleasure to welcome Mr. Devendra Pagnis, Mr. Sandeep Kumar, Ms. Shabina Omar, and other distinguished speakers on this beautiful day to our even more beautiful campus.

Samvaad refers to conversations, dialogue and chat. Today, when we mention chat, the first thing that comes to mind is ChatGPT, the ubiquitous AI model that has taken the world by storm. This makes Samvaad an appropriate name for our conversation about the impact of Artificial Intelligence on the workplace. In a more traditional sense, a conversation is what differentiates humans from animals and the ability to converse is a leitmotif for intelligence. So today we will have a conversation about how humans are and will be impacted by artificial intelligence.

We know by now that AI will eliminate many jobs in the corporate world and cause significant hardship. It is easy to claim that it will not be as severe as anticipated. Historically, we are told,  technology creates more jobs than it destroys. When the industrial revolution made blue collar workers redundant, they moved to the service sector - in finance, hospitality, and entertainment. When automation eliminated white-collar jobs, it created more positions in the tech sector, and India's IT industry is a testament to that. So when ChatGPT and other large language models eliminate jobs, they will also create new opportunities. But what exactly will be the jobs that will be created? There is no clear answer. Honestly, there is really nothing on the horizon that will keep thousands of people meaningfully employed. We can hope for the best, but we must prepare for the worst.

Can we put restrictions on AI? We could but it is unlikely that we would be successful in stopping its triumphant march. There is no army that can stop an idea whose time has come. Time! Chronos, remember that word, I will come back to it a little later.

One possible way to soften the blow is to consider Universal Basic Income. Let people be jobless, but provide them some money so they do not create social unrest. Political parties in India are already implementing a version of Universal Basic Income by giving cash handouts to various groups - women, farmers and others. But the sustainability of this strategy is debatable. How long can this continue before states and the nation itself go bankrupt? Not for too long, but then what? The answer lies in the future, hidden in the womb of futurity.

Unlike the unknown future, we have historical precedents. Both in the very distant past and in more recent times.

100,000 years ago, there were at least nine species of humans - or hominims - on Earth. Today, only one species, Homo Sapiens, survives. All others have been eliminated for various reasons. Remember, the word Sapiens comes from our ability to think. Closer in time, during the 14th and 15th centuries, European civilization wiped out the native civilizations of North and South America and Australia. In this case, the species was the same, but the civilizations were different. Today, we face a similar situation.

Artificial Intelligence, or what historian Yuval Noah Harari refers to as Alien Intelligence, is the new threat confronting Homo Sapiens. Just as Homo Sapiens were the threat that eliminated Homo Neanderthalensis, these "Neo Sapiens" represent a real threat to us. We must not underestimate this threat or ignore it.

But what can we, as Homo Sapiens, do to protect ourselves from this new species spreading across the planet? During the COVID-19 pandemic, we had a vaccine, but what do we have in this case?

I wish I had a clear and definitive answer, but I must disappoint you. I do not have an instant solution. Instead, let me share with you something that you might find interesting.

Looking back to the past and forward to the future requires us to examine Time more closely. In India, we refer to time as Mahakal, which translates to Megachron. Speaking of Megachron, I would like to talk about a science fiction trilogy, the Chronos Trilogy, that I have been writing for the past five years. These novels -- the last of which was published last month -- explore how humans and machines, Homo Sapiens and Neo Sapiens, will or might coexist on Earth and other planets over the next 500 years. 

In our Indic philosophy, Tantra means knowledge, Yantra means a device to delineate and discover this knowledge, and Mantra is the code to unlock the Yantra and Tantra. Within this framework, my three novels - Chronotantra, Chronoyantra, and ChronoMantra - investigate how humans and machines will evolve and exist together. I do have a vision and I wish I could provide a synopsis of what I anticipate will happen, but it is too complex for a short session like this. If you truly want to understand, you would need to read these three books.

In summary, there is or could be a way to resolve this conflict, but only if we recognize that this is not merely a technological or economic threat. We are facing a new species, a new civilization that is encroaching on our territory. This is a clash of civilisations. A clash conjures up visions of hostility. Is there an alternative that is cordial and amicable? That is the theme that is explored in this science fiction trilogy.