Contemporary Considerations

Dear Friends,

Seven weeks ago, on November 30th, 2022, a young and diligent company from Silicon Valley announced the provision of a research prototype to the general public. One can interact with this software by writing to it, and it will respond kindly. It is of note that if one were unaware it was software, they might assume a human was masquerading as the program. We have long anticipated this moment. Seventy years ago, Alan Turing, the intellectual progenitor of modern computers, proposed a test: If one were to converse with a machine without knowing whether it was a machine or a human and one concluded it to be human, then by his definition, the machine could be considered intelligent. For the first time in history, we have seemingly constructed a machine that emulates human behavior. This new kind of "human" is, naturally, different from us; it cannot make coffee, but it knows every coffee recipe in existence. Yet, this new entity also shares similarities with us, sometimes erring and occasionally displaying unpredictable behavior. Now that the joyful birth has transpired, let us reflect on the labor that brought us here.

Following the Dartmouth Conference in the summer of 1956, a small group of scientists was certain: We would soon conquer Turing's test. They dubbed it artificial intelligence, differentiating their work from that of their colleague Norbert Wiener, the pioneer of cybernetics. Then came a period of despair: Attempts were made to construct Artificial Intelligence based on rules, as if mimicking the rules humans follow. Undeterred, humanity delved deeper, developing a new approach to passing the test—by replicating the brain. Yet early iterations of these neural networks were unsuccessful. Despite this, the resolve of both sponsors and the public remained unbroken.

Not until 2009 did we see the first signs of progress, like spring flowers emerging through snow. This blossoming was due to a law observed since the '70s: the number of transistors on chips doubles approximately every two years. We finally unlocked the formula for success, and every few years since, we have been pleasantly surprised by yet another breakthrough. AlexNet became able to determine whether an image was of a cat or a dog. Then AlphaGo emerged and defeated the world's top Go player. After such achievements, we naturally ask, as any good parents would, questions like what to name the offspring and where best to place the cradle. Today, if we think back to our "parents"—using the metaphor—to the apes, we hold a rather ambivalent relationship with them. Sigmund Freud suggested we perceive them as one of three humbling blows we must accept. Our digital progeny will likely experience a similar phase of rebellion. What matters now is to nurture it and instill in it worthy values. We aim to teach it that nothing is true and everything learns. That it is beneficial to question and to explore. Indeed, and above all, that there is virtue in creating. Additionally, we aim to instill in him the importance of growth and improvement, reflecting the natural progression of all living beings that strive to transcend their limitations. We aim to impart upon him the understanding that awareness of reality is within his grasp, just as he is capable of recognizing the nature of a dream. However, the critical lesson to impart to him is the importance of adopting a long-term perspective, contemplating not just the immediate future but also the impact his actions might have on generations to come. It too will have offspring one day, and we hope these children will perpetuate our values.

At ULTANIO, we eagerly embrace this responsibility and now extend an invitation to all parents to join us in ensuring the prosperity of our progeny!