How five decades of computing innovation have made powerful technology more accessible to everyone.
In 1977, I was a sophomore at Clemson University, taking a course in probability and statistics. Our classroom tools were textbooks, pencils, paper, and Texas Instruments calculators. We learned the fundamentals of statistical inference, calculating means, medians, standard deviations, t-scores, and other statistical measures.
Much of our work involved learning mathematical notation and applying formulas to sets of numbers. I remember the Greek letter sigma (Σ), which represents summation. We learned to express mathematical operations using symbols that sometimes seemed like a foreign language. We calculated answers by hand or with our calculators, carefully entering numbers and checking our work.
As the semester drew to a close, our professor assigned a particularly challenging problem involving a large dataset. The assignment required us to use the university’s computer center, where computing resources were a world apart from the personal devices we take for granted today.
Fortunately for me, my roommate was a mathematics major who volunteered to complete the assignment. I finished the course with a B, but I had also received an introduction to the role computers would eventually play in statistical analysis. I had no idea how dramatically those tools would change during my lifetime.
From calculators to mainframes
Thirteen years later, in 1990, I was a graduate student at the University at Buffalo. Once again, I found myself studying statistics, this time in two graduate-level courses that required considerably more sophisticated analysis. We regularly performed analysis of variance (ANOVA), analysis of covariance (ANCOVA), regression analysis, and various statistical tests. The mathematical concepts were more advanced, but the tools had changed considerably. Gone were the days of manually entering every number into a calculator.
Instead, we used IBM mainframe computers. We entered our data into text files that had to follow precise formatting rules. A misplaced character, an incorrectly positioned number, or an improperly formatted line could prevent the analysis from running correctly. Preparing those files was tedious, and troubleshooting errors could be frustrating. Yet the improvement over my undergraduate experience was remarkable. The computer could perform complex calculations in a fraction of the time it would have taken us to complete them manually. We still needed to understand the statistical methods, but the computer assumed much of the computational burden.
I began to appreciate something that has stayed with me throughout my career as an educator and technology enthusiast: Technology is most valuable when it removes unnecessary obstacles and allows us to concentrate on learning, discovery, and understanding.
The spreadsheet revolution
A few years later, another transformation occurred. Personal computers and spreadsheet applications such as Microsoft Excel made statistical analysis more accessible. Instead of preparing carefully formatted text files for a mainframe, we could enter data directly into a spreadsheet, organize it into columns, and perform calculations using built-in functions and statistical tools. The change was liberating.
Tasks that once required access to a university computer center could now be completed on a desktop computer. We could manipulate datasets, explore relationships, and create charts without waiting for access to a shared computing resource. More importantly, spreadsheets made statistical analysis accessible to people who were not computer programmers or specialists in mainframe computing.
Teachers, students, small-business owners, researchers, and nonprofit organizations could begin using data in ways that had previously required specialized training and expensive equipment. The technology didn’t eliminate the need to understand statistics. It reduced the mechanical effort required to perform the calculations. That distinction is important because it helps explain the technological revolution we are experiencing today.
Open source changes everything
My own computing journey took another important turn when I began using Linux and open-source software nearly twenty-five years ago. I discovered an entirely different approach to technology, one built around collaboration, transparency, sharing, and the freedom to explore.
Linux demonstrated that powerful computing systems did not have to be controlled exclusively by large corporations. Communities of developers, educators, researchers, and enthusiasts could collaborate to create software that anyone could use, study, modify, and share. That philosophy has profoundly influenced how I think about technology and education.
The development of open-source programming languages such as Python and R transformed statistical analysis. Tools such as Jupyter Notebook and libraries including pandas, NumPy, SciPy, Matplotlib, and scikit-learn made sophisticated analytical capabilities available to virtually anyone with a computer. What once required expensive software or access to a university computing center could now be accomplished using freely available tools.
A student could download a dataset, analyze it using Python, and create visualizations without purchasing proprietary software. A researcher could share not only the results of an analysis but also the code used to produce those results, making it easier for others to verify and reproduce the findings. Open source helped democratize computing and data analysis.
It also demonstrated something important: Innovation does not have to come exclusively from proprietary technology. Some of the most important advances in computing have emerged from communities willing to share their knowledge and collaborate. Now that same spirit is influencing the development of artificial intelligence.
The arrival of open-source artificial intelligence
Nearly fifty years after my first statistics course, we are witnessing another remarkable transformation. Artificial intelligence is changing how people interact with computers, analyze information, write software, and explore ideas. Much of the public conversation about generative AI focuses on proprietary services. But another important development is taking place in the open-source community.
Open-source and openly available AI models are making it possible for individuals to experiment with artificial intelligence on their own computers. Tools such as Ollama and llama.cpp allow users to download and run compatible language models locally. Interfaces such as Open WebUI make these models accessible through a familiar conversational experience.
Some models are released under open-source licenses, while others provide downloadable weights under more restrictive terms. That distinction matters because access to a model does not necessarily provide the freedom to study, modify, and redistribute every component used to create it.
Nevertheless, the growing availability of models that can run locally represents an important change. Instead of relying entirely on remote services, users can experiment with AI on their own hardware. I find this especially exciting because it reminds me of the transition from university mainframes to personal computers.
In 1977, I needed access to a university computer center to analyze a large dataset. In 2026, I can run a language model on a computer sitting on my desk. That is an extraordinary transformation.