I’ll say it plainly: the race between what universities teach and what the market demands stopped being close a while ago. Technology won, by a landslide. And not enough people are talking about what that means for a 20-year-old picking a major right now, or for someone 45 years into a career doing a job AI is about to rebuild.
What’s actually real
Colombia has more AI course offerings than ever. The Ministry of Education counts 865 active academic programs tied to artificial intelligence as of 2025, according to La República. Programs with “Artificial Intelligence” literally in the name went from 12 to 92 in five years —a 667% jump—, per SNIES data cited by IFM Noticias. Universidad de los Andes runs a virtual AI master’s (MAIA). Universidad Católica de Colombia opened an AI and data engineering undergraduate degree with over 260 dedicated computing stations. Politécnico Grancolombiano, EAN, Icesi and the National University all have formal offerings now. And the government, through MinTIC, certified 104,065 Colombians via Talento Tech, with AI as the single largest category (35,148 certifications), and already launched AvanzaTEC 2026, with 53 free courses backed by Microsoft, Oracle, Cisco and Mercado Libre, per MinTIC.
What nobody tells you
With more than 15 years watching finance and tech curricula try to catch up to the market, here’s what those numbers don’t capture: program count isn’t the same thing as real adaptation speed. A master’s program takes two years to design and four semesters to teach, and by the time its first graduate walks out, the frontier model they studied is already two generations behind. Frontier AI doesn’t move on academic calendars —it moves on product cycles, measured in weeks. While a university is still deliberating whether to open a new track, the AI labs have already shipped three versions of the model that track was built to teach. That’s the real gap: not a shortage of programs, but the speed mismatch between an institution built for stability and a technology built for constant disruption.
The real opportunity
What should you do if your school isn’t at the frontier? Don’t wait for it to get there. Institutions falling behind should stop trying to compete on specific technical content —that race always goes to whoever updates fastest— and start teaching what doesn’t expire: applied critical thinking, the judgment to audit what AI produces, and the discipline to pick up any new tool fast. For the professional who never got that formal training, the good news is the alternate path already exists, and it’s free: Talento Tech and AvanzaTEC aren’t a consolation prize —today, they’re the fastest reskilling route in Colombia.
Because the meta-skills that actually matter aren’t knowing how to write a prompt —any tutorial solves that in an afternoon. The World Economic Forum has ranked analytical thinking and complex problem-solving as the top demanded skills for five consecutive editions of its Future of Jobs Report, and projects that 39% of the workforce’s core skills will change by 2030 (World Economic Forum). What separates someone who thrives with AI from someone it leaves behind isn’t the tool they use —it’s their judgment: the ability to ask the right question, catch when a model’s confident answer is wrong, and make the call the machine can’t make for you, because it doesn’t bear the cost of being wrong.
And here’s the uncomfortable question worth sitting with: who benefits if only a few of us actually know how to get real value from AI? The honest answer isn’t a conspiracy, it’s a plain incentive. Whoever already controls the compute, the data and the capital profits from a market where genuinely AI-literate talent stays scarce —that concentrates bargaining power upstream, not downstream. Which is exactly why it’s no accident that the programs actually closing this gap in Colombia are the ones deliberately widening access: free public training, not just graduate degrees for people who can already afford them.
Frontier AI isn’t going to wait for anyone. The question you can actually control is whether you decide to get on board with judgment, or let speed decide for you.
If you want to keep this conversation going daily, I’ll be waiting in the Daily Briefing —that’s where we talk markets every single day, holidays included.