We are available from Monday to Friday from 9:00 a.m. to 6:00 p.m. (Central Mexico Time)
MSP Mobility Corp. MéxicoFill out this short form and well get back to you in less than one business day.
Fill in the contact formTalk to a specialist, open a case or ask a question in the forum, or manage your support account preferences.
Go to MSP SupportThank you for your interest in our products and solutions, we invite you to complete the following contact form and a specialist from the Mobility in a Box suite will be in contact with you shortly.
Thank you for your interest in our products and solutions, we invite you to complete the following contact form and a specialist will be contacting you shortly.
Thank you for your interest in our offer, we invite you to complete the following form and we will be in touch with you shortly.
Thank you for completing the following form, so that a business specialist can contact you and clear up your doubts.
(*Required fields)
Only applies to Mexico.
As developers who began our careers in the 1990s, we were both witnesses to and protagonists of a golden era of software development centered on relational databases. Likewise, our approach to problem-solving was fundamentally different from what we see today with modern artificial intelligence.
This reflection explores the profound paradigm shift we have experienced: from structured queries on stored data to deep learning based on neural patterns.
In that decade, our technology stack revolved around relational databases such as Oracle, SQL Server, and MySQL. We learned to “think in tables” and to model reality through entities, relationships, and normalization. Therefore, every problem was broken down into:
Our mindset was deterministic: for every question, there was a specific query which, when executed on structured and pre-stored data, returned an accurate and reproducible answer. Likewise, the knowledge resided in the structure of the tables and the relationships between them.
The typical pattern of a 1990s application followed this sequence:
This model worked exceptionally well for transactional and business management systems, where information had a predictable structure and queries followed defined patterns.
Contemporary artificial intelligence, exemplified by Large Language Models (LLMs) such as GPT, Claude, or LLaMA, operates under fundamentally different principles. There is no “database” in the traditional sense, but rather a neural architecture that has “learned” patterns during an intensive training process.
Today’s models use the Transformer architecture, introduced in the paper “Attention is All You Need” (Vaswani et al., 2017). This architecture:
The “query” process in modern AI is radically different from an SQL query:
Emerging Skills:
Architectural Considerations:
At MSP, our work specifically involves applying these skills to solve business problems through the use of AI.
This evolution does not imply total replacement. Future systems will likely combine:
As IT professionals who have journeyed from JOINs to attention networks, we have witnessed one of the deepest paradigm shifts in computing history. The transition from querying explicitly stored knowledge to inferring answers from learned patterns represents not only a technological evolution but a transformation in how we conceptualize information processing.
The future will belong to those who can integrate both worlds: the precision and reliability of relational databases with the flexibility and generative capacity of modern artificial intelligence.
Discover some applied AI use cases for business by visiting us at MSP IA.
Error: Formulario de contacto no encontrado.