Riccardo Di Molfetta is a young Italian entrepreneur who founded Symbiotica startup based in Boston and London, which has already received seed funding and is focused on aligning artificial intelligence “with the values and intentions of humanity.”
That is: working to ensure that future “superintelligent” AIs are on humanity’s side, avoiding existential risks. He recently met Sam Altmanthe CEO of OpenAI (CROSSLINK HERE AND IT MUST BE DONE THE REVERSE ON THE OTHER PAGE on this page you will find the interview with the father of ChatGpt, which we co-publish exclusively). Di Molfetta, 24 years old, has a humanistic background and his studies – different from other founders in the world of AI – were a strong point to leverage: «Yes, I studied at the Brera Art School in Milan , and later History and Philosophy at King’s College London. My interest has always been in the human condition, trying to answer existential questions about what we do in the world and what the human being fundamentally is. I believe this is a multidisciplinary issue that involves humanistic fields such as history, philosophy and art, but also science and technology such as artificial intelligence.” Then he adds: «AI is a tool to better understand who we are and during my studies I realized that the important thing is not just to acquire knowledge, but to develop a proactive and multidisciplinary approach to understanding things. Given that AI can compress huge amounts of human information, we should reconsider what analytical and critical skills we want future generations to learn, rather than relying on passive knowledge. The art school gave me prospects for creative thinking, interest in humanism and the Renaissance. At Harvard and MIT I translated my humanistic background into a more scientific one, studying AI, cognitive and computational sciences.”
What was the perception of your humanistic approach when you started promoting your startup in the United States?
«My humanistic and multicultural background was a precious added value in approaching the American technological context. Coming from a continent with a deeper history and in close contact with different cultures has allowed me to bring creativity and new perspectives to the conception of innovative products. There is a very standardized understanding of the go-to-market and creating products in America. My European cultural background has given me more tools than many American technicians, allowing me to imagine new product categories thanks to a different vision. It is an added value that is difficult to find in America and which must be fully exploited.”
At what point is your startup’s journey and how many people work with you?
«After a formative period raising small funds in Silicon Valley, we successfully closed a seed funding round. We are currently a team of five, but we are not hiring new staff at the moment. It’s true that AI talent is difficult to find, both in Europe and America. This is why we have chosen to have two offices, one in London and one in Boston. The idea is to attract European talent, who often end up moving to California, while maintaining an American base. It allows us to unite the two worlds. Furthermore, for the development of new generations of models, we are not only looking for skills in data science and computer science, but also in cognitive science, an approach that we believe is crucial.”
Tell me more about your meeting-interview with Sam Altman and the questions you asked him.
«We knew Sam was coming to Harvard and MIT in early May, so we tried to arrange a meeting with him. It was an ideal experience, I had the chance to ask him some questions about our field of application, i.e. how to create a more intimate and human interaction between artificial intelligence and humans. I asked him what is missing from a technical-scientific point of view in current language models to achieve this connection. He replied to me that OpenAI is interested in customizing models, but that we can and must go further, knowing the user’s personal data is not enough. After the interview we also had a short, very productive private exchange.”
Can you help us better understand what AI model alignment is, which you deal with? And what is the “superalignment” that your start-up wants to dedicate itself to?
«Alignment is about the ability to align advanced AI models – with cognitive capabilities superior to those of humans – with humanity’s values, preferences and intentions, so that they can be truly beneficial to society. We currently have no established techniques for achieving this alignment with super-intelligent models. As human beings, we align ourselves on a social level through laws and institutions, but also through the interpersonal dynamics developed in family or between friends, sharing emotions and life experiences. The Symbiotic is exploring precisely this “social” approach to create a more intimate and empathetic bond between AI and human beings, going beyond the simple customization of models through users’ personal data.”
Do you believe that true Artificial General Intelligence (AGI) will be achieved in a reasonable time?
«Personally, and as a company, we are quite skeptical about the term AGI. We believe it will be more plausible to see the emergence of “superintelligences” that specialize in certain domains, rather than general human-like intelligence. Current models are already capable of generalizing across various cognitive abilities, but we do not know whether this “generality” paradigm is accurate. It is possible that some aspects will far surpass human capabilities, while others will lag behind. No one can predict future developments with certainty. The current Large-langue-models (Llm) have already amazed everyone with their unexpected progress. There will probably still be significant improvements, but it is also possible that we will reach a wall that will require us to completely rethink the approach, as has happened several times in the history of AI.”
