01
How LLMs generate text
Token prediction, temperature, sampling, what each one means for the output you’ll see.
Pillar 01
AI Foundations & Limits
Understand how AI actually works, and where it reliably fails.
You are here · pillar 01 of 07
Foundations
Interaction →
A working definition
What this pillar is, what it isn’t, and why it matters.
What it is
+
A working mental model of what’s under the hood.
Knowing in plain language how language models predict text, what training data does and doesn’t include, and why hallucinations happen at all.
What it isn’t
−
A computer-science degree.
You don’t need to derive transformer math. You need enough intuition to predict where a tool will be reliable and where it won’t.
Why it matters
→
Every other pillar depends on this one.
Without a working model of how AI works, every output looks equally trustworthy, and that’s how new graduates get caught.
76%
of recent hires can’t explain how a model knows what it knows. (NACE, 2025)
What’s inside this pillar
The sub-skills the assessment is actually measuring.
Each pillar breaks into 4 sub-competencies. The assessment scores each one independently so you can see what to work on, not just an overall band.
02
Training data, in plain terms
What’s in the corpus, what isn’t, and how cutoff dates create blind spots.
03
Where hallucinations come from
The structural reasons a model invents citations, and the patterns that predict when it’ll do it.
04
The capability frontier
Knowing what current models can and can’t do, and how fast that changes.
Try a sample
A real question from this pillar.
The full assessment has 44 of these for Pillar 01. There’s no time limit and you can pause anytime.
Sample · 1 of 44
You rephrase a prompt slightly and get a different answer. What best explains why that happens?
One correct answer
Proficiency levels
Where the assessment will place you, and where to aim.
“Proficient” is the workplace-ready bar. Most graduates leave school at Developing; recruiters look for Proficient or above.
Level 01
Novice
Limited working knowledge. Outcomes are unreliable and depend on luck.
Score · 0–25%
Level 02
Developing
Foundational fluency. Can complete simple tasks; struggles in novel contexts.
Score · 26–55%
Level 03
Proficient
Workplace-ready. Handles real tasks at expected quality with light oversight.
Score · 56–80%
Level 04
Advanced
Expert practice. Sets standards for peers, anticipates failure modes, adapts quickly.
Score · 81–100%
Reading list
All articles →
To go deeper on this pillar
You Learned AI as a Chatbot. Work Will Ask You to Manage One.
You can use a chatbot. That is one layer of four. From analytical models to agents to the open frontier, your job moves up toward judgmen…
Confident and Wrong: Why Reasoning Models Still Make Things Up
The newest AI models reason better than ever, which makes them more dangerous, not less. Capability has outpaced reliability. Here is the…
Your AI Doesn’t Know Anything: How LLMs Actually Work
Your AI does not know anything, it predicts text. Understanding how large language models actually work, through tokens and next-word pre…
Start with the assessment
Find out where you stand across all 7 pillars.
The free AI Literacy Assessment scores you independently on each pillar and shows you exactly where to focus next.
✓
10 minutes · No email required to see results