The world is moving fast. But some things don’t change.
Maths. Physics. Chemistry. Philosophy.
Everything we call innovation is built on top of these. The iPhone runs on physics and logic. A vaccine is chemistry and biology working together. A courtroom argument is structured philosophical reasoning. The fundamentals always show up somewhere.
The person who knows why something works will always be able to use AI tools effectively.
AI Skills worth learning right now:
Communication and Output
Prompt Engineering — how to talk to AI models clearly and get consistent, useful results. Most people are still winging this.
AI Workflow Automation — tools like Make, Zapier, and n8n that connect AI into real business processes. People building these are saving hundreds of hours a month.
AI Video, Audio and Visual Generation — content production is changing fast. Sora, Runway, and similar tools mean you can produce what used to require an entire creative team.
Building and Technical
RAG (Retrieval Augmented Generation) — connecting AI to your own documents and knowledge bases so it works with your specific context, not just general knowledge.
AI Agents — systems that take actions autonomously, not just answer questions. This is where things get genuinely powerful.
API Integrations — connecting AI models into existing tools and products. The people who can do this are in serious demand.
Fine-tuning — training models on your own data so they perform specifically for your use case.
Data and Judgment
Data Literacy — being able to read what numbers are saying and ask the right questions. AI generates enormous output. Knowing what matters is a skill on its own.
Critical Evaluation of AI Output — knowing when the model is wrong or confidently mediocre. This comes from domain knowledge. Which brings us back to the fundamentals.
Cost and ROI of AI Tools — a lot of businesses are spending money badly here. Knowing what's worth it is already a competitive advantage.
Financial Engineering and related fields
Quantitative Analysis — using AI to model risk, price assets, and find patterns in market data that humans would miss or take weeks to find.
Algorithmic Trading — building and refining trading strategies with AI assistance. The speed and pattern recognition here is unlike anything before.
Financial Modelling — AI dramatically accelerates scenario planning and forecasting. What used to take days now takes hours.
Fraud Detection — understanding how AI systems flag anomalies in transactions. Every financial institution is building this out.
Credit Risk Assessment — how models evaluate borrowers now versus traditional methods. The criteria have expanded significantly.
Regulatory Compliance Automation — one of the biggest pain points in financial services and AI is starting to solve it well.
Physical Skills
Welding, electrical work, plumbing — robots still struggle enormously with unstructured physical environments. A burst pipe at 11pm still needs a human.
Surgery and procedural medicine — AI assists, imaging improves, but hands and judgment still matter enormously at the table.
Farming and agriculture — understanding land, seasons, soil, and ecosystems is knowledge that took generations to accumulate.
Construction and structural work — reading a site, solving problems on the ground, adapting in real time. This doesn't happen on a screen.
Physical therapy and rehabilitation — deeply relational, deeply sensory. The human presence is part of the treatment.
Woodworking, glassblowing, ceramics — making real things with your hands. Craft at a high level involves instinct and embodied judgment that is genuinely hard to replicate.
