From BASIC to AI: A Personal Journey Through the Future of Technology
In this talk at BEEVA (2017), I talk about my journey in the world of artificial intelligence, from my beginnings programming in BASIC to my work on dialogue systems and agent-based architectures. I explore advancements in AI, its practical applications, and how this technology can enhance our capabilities rather than replace us.
2017: My vision of AI
On an afternoon in 2017, I had the opportunity to share my experience in the fascinating world of artificial intelligence (AI). My journey began more than 25 years ago when my father brought home our first computer. That machine, which we would consider prehistoric today, sparked a curiosity in me that would lead me to explore the limits of technology and AI.
The First Steps
My adventure started with BASIC in 1992, a programming language that opened the doors to a world of possibilities for me. The “Scientific American” magazines that arrived at home fueled my imagination with ideas about robotics and AI, leading me to dream of a future where machines and humans would interact in unimaginable ways.
From University to Industry
Choosing to study Computer Engineering (at University of Granada) was natural. Although at first I encountered more mathematics and programming than I expected, my passion for AI remained alive. My final year project during 2009, “Mayordomo” a voice interaction software to control home appliances, was in some ways the same idea to what we now know as Alexa or Google Home.
Delving Deeper into AI
My research thesis led me to specialize in dialogue systems, exploring how machines can understand and generate natural language. I worked on tools to evaluate these systems, always with the goal of creating more natural and efficient conversational interfaces.
AI in Practice
At that time of my professional experience, I had applied AI in various fields:
Natural language processing to extract relevant information from large volumes of text.
Architectures based on intelligent agents to solve complex tasks collaboratively.
Dialogue systems oriented to specific tasks.
Challenges and Expectations
Despite advances, the industry still faces significant challenges. Accurate understanding of user intentions and the ability to respond to questions outside a system's scope remain areas for improvement. The gap between user expectations and the actual capabilities of current conversational systems is a reminder that there is still a long way to go.
The Potential of AI
AI is not a threat, but a powerful tool that can amplify our capabilities. From route optimization to assistance in medical diagnostics, AI has the potential to significantly improve our decision-making and efficiency in various areas.
Looking to the Future
In 2017, the vision was clear: AI would give us "superpowers". Products like Boomerang for optimizing emails, recommendation systems like Netflix, and advances in pattern recognition and autonomous driving were just the beginning. The Clever platform, which we developed based on a intelligent agents architecture, sought to bring these capabilities to businesses, allowing them to automate complex tasks and improve information management.
Reflection: From 2017 to 2024
Seven years have passed since that talk, and it's fascinating to see how some predictions have materialized while others have taken unexpected directions:
Virtual assistants like Alexa and Google Home have become ubiquitous in many homes, fulfilling the vision of voice interaction with our environment.
Natural language processing (NLP) has taken giant steps with models like GPT, overcoming many of the limitations mentioned in 2017.
Autonomous driving has advanced, although not at the pace many expected, facing technical and regulatory challenges.
Generative AI has emerged as a revolutionary field, capable of creating content in a way that was not fully anticipated in 2017.
The ethics and social impact of AI have gained prominence in public discourse, underscoring the importance of responsible development.
What hasn't changed is the fundamental vision of AI as a tool to enhance human capabilities. The final phrase of the talk, "the best way to predict the future is to invent it", remains as relevant today as it was then. As we continue to innovate and explore the possibilities of AI, it is crucial to maintain a human-centered approach, ensuring that this technology serves to improve our lives and solve the most pressing challenges of our society.
The future we imagined in 2017 is here, in many respects, but it has also surprised us in unexpected ways. The task now is to continue inventing, with responsibility and vision, the next chapter in the history of artificial intelligence.
Chatbots have evolved from simple command-based systems to sophisticated conversational AI tools. This article explores their functioning, from the pioneering ELIZA to modern LLM-based systems. We analyze the traditional process: natural language understanding (NLU), dialogue management (DM), and response generation (RG), comparing it with the end-to-end LLM-based approach.