Artificial Intelligence: From Wow Factor to Real Problem Tackling
Leveraging domain-specific knowledge to turn AI potential into impactful solutions
10/22/20242 min read
The Industrial Revolution brought us machines, and electronics gave them intelligence. There was a time when a machine capable of performing calculations inspired awe in the masses, as it executed a task that until then was considered exclusive to highly educated individuals. From that moment on, progress was unstoppable. Expert systems emerged, capable of encapsulating complex knowledge, followed by the first computers able to play chess at the highest level, eventually defeating the best human grandmasters. However, when the “wow factor” fades away, we stop considering these systems intelligent. They are simply machines executing a complex program that enables them to do those things, nothing more. True intelligence remains human, and more effort is demanded of those working in the field of artificial intelligence to prove their value.
Today, the frontier of artificial intelligence lies in what are called “foundation models”: powerful deep learning models that combine innovative data processing architectures, significant computational capacity, and massive amounts of digital information. These models can generate a poem, create an image, or draft a text following our instructions. They are useful and entertaining, but they do not always solve everyday problems in a practical way. For that, we need something more: we need domain-specific knowledge.
Consider the example of a supermarket manager who is hiring a new employee. In the selection process, the manager looks for basic aptitudes such as human intelligence, availability, salary suitability, or physical aptitude. These are the capabilities that a foundation model, or a combination of them, could assess. However, once hired, the employee needs to learn how the business works: how to restock products, what strategies are used to arrange items on shelves, how to operate the cash register, or how to interact with customers. This is what we call “domain-specific knowledge,” and it is what turns the employee’s basic skills into a source of value for the company. The employee is the human being who has signed a contract and acquired the training necessary to perform work that adds value to the business and to society. The AI solution is a combination of foundation models, domain-specific knowledge, and custom programming, providing a higher-level interface that allows users to solve practical tasks more easily.
In today’s market, there is room for a limited number of companies developing specialized hardware for artificial intelligence, as well as for the supporting software ecosystem. There are also few opportunities for the development of new foundation models due to high barriers to entry. However, the space for applying these models and configuring systems is vast and economically accessible. Here, it is about integrating intermediate products with a solid dose of domain-specific knowledge to create value for a company or a particular sector.
This space includes everything from large companies developing generalist solutions to a multitude of opportunities in the so-called “long-tail,” that enormous segment of the market that is not profitable for large corporations and their high operating costs. In any case, this is where the true utility of artificial intelligence can be demonstrated. All the investment in research and model development—is it truly worthwhile? Can it provide useful services to humanity that justify the resources used and improve quality of life? The answer to this question will determine whether we are in a speculative bubble driven by the wow factor and expansive monetary policy, or if we will indeed achieve something valuable and enduring from these advances.
Ultimately, every business manager, software developer, and AI practitioner should think deeply about how to harness these technologies to create practical solutions that drive genuine value. Now is the time to move beyond fascination and make these tools impactful, ensuring that artificial intelligence becomes a transformative force for businesses and society alike.