Recommended First Steps for New Learners

Learning Objectives

• Identify your personal next steps for building genuine AI fluency
• Reflect on the scope of what you have learned across all 11 modules
• Develop a mindset for continuous learning in a field that changes faster than any curriculum can keep up with

You Have Made It to the End — Now the Real Work Begins

This is the final section of the AI Fundamentals curriculum. You have covered a tremendous amount of ground:

• The foundations of how AI, machine learning, and large language models actually work
• AI's deployment across ten major industries
• How AI is reshaping careers and what the resilient professional looks like
• The full landscape of US foundation models — from OpenAI to Amazon — and international models from France, China, and Canada
• A comprehensive directory of AI tools across thirteen categories
• The enterprise application platforms and SaaS ecosystems built on top of AI
• How AI agents work — their architecture, frameworks, and failure modes
• The full developer and infrastructure landscape: IDEs, CLIs, hosting, hardware, databases
• Near-term, medium-term, and long-term AI trajectories — with honest uncertainty
• AI's societal impact, responsible AI principles, and how to stay current

That is not a small thing. Many professionals who work alongside AI every day have not developed the integrated understanding you now have.

But understanding is not the same as fluency. Fluency comes from doing — from actually using these tools, making mistakes, iterating, and building the instincts that only come from hands-on experience.

These six first steps are designed to move you from understanding into fluency. They are concrete, low-cost, and high-impact.

Step 1: Create Free Accounts on ChatGPT, Claude, and Gemini — Use Them Daily

This is the most important step, and the most frequently skipped.

All three of these products have genuinely capable free tiers. There is no reason to wait until you understand AI better to start using them — you will understand AI better by using them.

The instruction: Use AI for actual tasks in your real life, starting today. Not experiments. Not playing around. Real tasks:

• Summarize a long document you need to read
• Draft an email you've been putting off
• Research a topic you've been meaning to understand
• Brainstorm approaches to a problem you're working on
• Ask it to critique something you've written

The goal for the first 30 days: find five tasks in your life where AI assistance saves you meaningful time or produces meaningfully better output. Then make those five tasks AI-assisted by default.

Using all three tools — ChatGPT, Claude, and Gemini — helps you develop intuition for where each performs best. You will notice differences in reasoning quality, communication style, and capability that no benchmark can fully convey.

Step 2: Watch Andrej Karpathy's "Deep Dive into LLMs"

Andrej Karpathy — former Director of AI at Tesla and OpenAI founding researcher — published "Deep Dive into LLMs like ChatGPT" in 2025. At roughly 3.5 hours, it is the most comprehensive, accessible explanation of how large language models work — from training data all the way through to the chat interface you use every day. No coding background required.

This is worth your time because after watching it, "tokenization," "attention mechanism," "context window," "temperature," and "next-token prediction" will stop being mysterious terms and start being things you understand structurally.

When you understand — even at a high level — how these systems work, you become a much better user of them. You know why they hallucinate. You know why longer prompts sometimes help. You know why they are sensitive to framing. The mechanisms make sense.

Search YouTube for: "Andrej Karpathy Deep Dive into LLMs"

For technical learners: Karpathy's earlier series "Neural Networks: Zero to Hero" builds a GPT-like model from scratch in Python. It is more code-heavy but gives you a hands-on understanding of the architecture. Both are free.

Step 3: Try a Free AI Image Generation Tool

Image generation is one of the most immediately visceral demonstrations of what modern AI can do — and it takes about five minutes to try.

Free options:

• Gemini (web or mobile): Nano Banana Pro image generation is integrated directly. Just describe what you want.
• ChatGPT: GPT Image 1.5 is available in the free tier with limited generations per day.
• Adobe Firefly (firefly.adobe.com): Available free on the web.
• Image Creator from Microsoft Designer (formerly Bing Image Creator): Free, integrated into Microsoft's design tools.

The goal is not to produce professional work. It is to develop direct intuition for what generative AI means — you describe something, and a capable AI creates a visual representation of it. That intuition is valuable for understanding the implications of AI across creative industries.

Step 4: Subscribe to One AI Newsletter for 30 Days Straight

Choose one newsletter and read it every day for 30 days without skipping.

Recommendation: Start with The Rundown AI (therundown.ai) for broad daily coverage, or Ben's Bites (bensbites.beehiiv.com) for a product-focused daily digest.

The 30-day commitment matters: the first week, most of the names and products will be unfamiliar. By the fourth week, you will begin to recognize patterns, understand context, and distinguish signal from noise. That pattern recognition — knowing what matters vs. what is hype — is enormously valuable and only comes from sustained exposure.

After 30 days, assess: is this newsletter worth continuing? Is there a better one for your specific interests? By then, you will have enough context to answer that question.

Step 5: Attempt a Small AI Coding Project

This step is for everyone — including those who have never written code before.

Why: Building something with AI — even if AI writes most of the code — teaches you how AI-assisted development actually works. You learn what you need to specify precisely, where AI makes confident mistakes, and what the experience of directing an AI system feels like from the inside.

Tools that require minimal or no prior coding experience:

• Bolt.new — describe an app in plain language; AI builds it; you can deploy it instantly
• Lovable.dev — conversational full-stack web app builder; describe what you want; iterate in natural language
• Replit — cloud-based coding environment with AI assistance; no local setup required
• v0 by Vercel (v0.dev) — describe a UI component or page; AI generates production-ready code
• Claude.ai — describe a specific small program or script; ask Claude to write it and explain what it does

Suggested first project: Build something small and personally useful. A personal expense tracker. A simple tool that reformats information in a way you find useful. A landing page for a project you're working on. The output is not the point — the experience of going from description to working thing is the point.

If you hit errors or things that don't work: that is part of the experience. Debugging with AI — describing the error and asking for help fixing it — is itself one of the most valuable skills in AI-assisted development.

Step 6: Re-Read the Platform Goals After 30 Days of Daily AI Use

At the beginning of this curriculum, we stated seven platform goals. After completing the curriculum and practicing daily AI use for 30 days, read them again.

1. Build practical AI understanding
2. Develop critical thinking about AI
3. Understand AI's impact on careers
4. Understand AI's broader impact on society
5. Explore honest AI futures
6. Make AI accessible to everyone
7. Help you keep learning as AI evolves

After 30 days of use and 11 modules of study, you will understand these goals differently than you did when you started. Concepts that were abstract will be concrete. Questions that seemed simple will have revealed their complexity. And — hopefully — you will have developed some of the genuine fluency that turns understanding into capability.

What You Have Accomplished

Take a moment to recognize what completing this curriculum means.

You can now:

• Explain how transformer-based LLMs work to a colleague who has never thought about it
• Evaluate a new AI model release and understand what the benchmark numbers actually mean
• Map the competitive landscape of foundation models — US, European, and Chinese
• Assess how AI is affecting your industry and your specific role
• Use AI tools fluently for research, writing, analysis, and coding assistance
• Engage critically with AI claims, hype, and risks
• Understand the governance frameworks — EU AI Act, NIST AI RMF — that are shaping the industry
• Think clearly about near-term, medium-term, and long-term AI trajectories without false certainty in either direction

That is a significant, genuinely useful body of knowledge. Most people who work alongside AI every day have not developed this level of integrated understanding.

AI Fluency Is a Practice, Not a Destination

Here is the most important thing to understand as you leave this curriculum: there is no point at which you are "done" learning AI.

The field will look meaningfully different in 12 months than it does today. Some of the models described in this curriculum will have been superseded. New tools will have emerged. The deployment patterns and use cases will have evolved. A section of this curriculum that was current in early 2026 may need significant updating by late 2026.

This is not a flaw. It is the nature of the field.

The learners who will be most capable and most valuable in three years are not the ones who studied AI intensively in 2026 and then stopped. They are the ones who built the habit — daily use, regular reading, periodic skill deepening, constant curiosity — and maintained it. They are the ones who stayed genuinely interested in what AI can do, what it cannot do, and what it means for the world.

AI fluency is a practice. Like physical fitness, it is not built once and maintained effortlessly — it is maintained by continued engagement. Unlike physical fitness, the engagement is genuinely interesting, because the field keeps changing in ways that reward attention.

A Closing Thought

The AI era is genuinely significant. It is not the end of human expertise or human meaning — but it is a substantial restructuring of what professional expertise looks like, what kinds of problems are tractable, and what tools are available.

Informed, thoughtful humans are not made redundant by this transition. They are made more important. The judgment to use powerful tools well, to evaluate their outputs critically, to understand their limitations, to direct them toward genuinely valuable ends — these are distinctly human capabilities that become more valuable, not less, as AI becomes more capable.

You have built the foundation for being that kind of thoughtful, informed person in the AI era. The rest is practice.

Key Takeaways

• The six concrete first steps — daily AI use, Karpathy's Deep Dive video, image generation, one newsletter, a small coding project, and revisiting the platform goals — move you from understanding to fluency
• Daily use is the non-negotiable foundation: occasional AI use produces marginal fluency; daily use produces compounding fluency
• AI fluency is a practice maintained over time, not a destination reached once
• The most valuable thing you can bring to the AI era is informed, critical, thoughtful engagement — the kind of judgment that AI cannot supply for itself
• You have built a meaningful foundation. Now use it.