Session 1 · What AI Actually Is
AI Fundamentals:
From Curious to Capable
A plain-language mental model for what AI actually is, and just as importantly, what it isn't.
Manuel Martinez
- Sr. Technical Account Manager · Amazon Web Services (AWS)
- Podcast Host · Career Downloads
- Business Owner · EZwerc AI
Five-Session Series
Building Toward Capable
By the end of this series: a mental model for what AI is, a repeatable way to prompt it well, and a critical eye for when to trust what it gives you.
Session 1
What AI Actually Is
Session 2
Session 3
Session 4
Session 5
Warm-Up
Where Are You Starting From?
Quick show of hands. There's no wrong answer in this room.
Avoiding It
Weekly
Daily
The Simple Flow
Input → Processing → Output
Every AI tool, no matter how sophisticated it looks, is doing the same three things.
You Ask
Input / Prompt
It Works
Processing
You Get It
Output
Under the Hood
Neural Networks
& Models
They learn similar to how we learn: identifying patterns.
Myth-Bust
Is AI Actually Thinking?
It predicts one piece at a time based only on what's already there. It can't see what comes next until it builds the step to get there.
Fluent ≠ correct. That sentence sounds confident because it follows a familiar pattern, not because it's true. (It was Neil Armstrong.)
Not All AI Is the Same
Same Category. Different Vehicle.
They're all large language models trained by different companies, on different data, tuned for different strengths.
A pickup truck and a sports car are both cars
ChatGPT
OpenAI
Broad, all-purpose assistant with the largest plugin ecosystem.
Claude
Anthropic
Careful, thorough writing and analysis; strong on coding and long documents.
Gemini
Built into Docs, Gmail, Sheets; strong with images and real-time info.
Perplexity
Perplexity AI
A research tool that searches the live web and cites its sources.
Grok
xAI
Built into X, with real-time social access and a casual, less filtered tone.
AI Sees More Than Text
Same Idea. Different Senses.
The same basic ideas apply across every type of content.
Text
Writing, summarizing, answering questions.
Image
Recognizing what's in a photo, generating new images.
Audio
Turning speech into text, or text into speech.
Video
Analyzing or generating moving content.
Voice-to-text, photo grouping, auto captions: you've likely used more than one already.
AI as a Tool, Not a Threat
It Doesn't
Want Anything.
AI responds when prompted, and stops when it's done. Judgment and accountability are still entirely yours.
No goals of its own
No judgment about what matters
No accountability for the outcome
Calculator
Spreadsheet
Search Engine
AI
Each one changed how work gets done. None of them decided what mattered. You still did.
Before Next Session
Take It For a Test Drive
Have one real conversation with any AI tool you already have access to, whatever's available to you.
Ask it a question that actually matters to your work, not a trivia question.
Notice: did it feel like it understood you, or could you tell it was predicting and pattern-matching? Where did it sound confident but questionable?
Now try a second, different tool you don't normally use. Ask it something similar and notice what's the same and what's different.
Bring one observation to share at the start of next session.
Questions?
Let's Talk.
Input, processing, output. Prediction, not thought. Many tools, many strengths. All of it yours to direct.
Manuel Martinez
Facilitator · AI Fundamentals: From Curious to Capable