Week 2 — Readings & Resources · How AI Actually Works (Conceptually) & Its Limits

Course: Using Artificial Intelligence (AI 101) · Silver Oak University (fictional sample) · Prof. Quinn
Objective covered: Objective 1 — Explain what generative AI and large language models are, how they work conceptually, and their core capabilities and limits.

How to use this page

Everything here is a link to an external resource — open it in your browser. Nothing needs to be downloaded.

This week's load is deliberately manageable: 2 short videos + 3 short readings, grouped by the lecture's ideas, plus the token-visualization tool. Watch or read one item per group and you'll be well prepared for the quiz; do all of them and you'll have a thorough, multi-angle picture. Total time is roughly 40–55 minutes if you do everything, much less if you pick one per group.

Order that matches the lecture: ① tokens → ② the context window → ③ hallucination → ④ search vs. AI + the Turing test → ⑤ try the tokenizer.

Keep the habit going: before you trust any claim in these resources (or from the AI), ask: Is this generated or verified? How would I check?

① What Is a Token?

Maps to Lecture Segment 2. The LLM processes text as tokens — small chunks predicted one at a time. Seeing this once makes it click.

Video — "Tokens, Transformers, and Language Models" (Google DeepMind / 3Blue1Brown-style explainers)
🔗 https://www.youtube.com/watch?v=wjZofJX0v4M
Why it earns the click: a clear walkthrough of how tokens work and why the "next-token prediction" framing is the key to understanding every LLM behavior — including hallucination. No math required.
⏱ ~10 min

Reading — "What Is a Large Language Model?" (Google Cloud)
🔗 https://cloud.google.com/discover/what-are-large-language-models
Why it's assigned: a plain-language overview that covers training, tokens, and how the model generates text — exactly the "open the hood" picture from lecture, in readable prose.
⏱ ~8 min

② The Context Window

Maps to Lecture Segment 3. The context window is the model's "working memory" — everything it can see at once. A bigger window holds more text; it doesn't make the model more truthful.

Reading — "What is a context window in AI?" (IBM)
🔗 https://www.ibm.com/think/topics/context-window
Why it's assigned: a concise explanation of what a context window is, why it matters, and what happens when it's exceeded — with plain-language analogies. This is the clearest written version of the idea from lecture.
⏱ ~8 min

③ Hallucination — Why AI Is "Confidently Wrong"

Maps to Lecture Segment 4. Hallucination happens because the model predicts likely text — not verified truth. Understanding the mechanism makes the cure obvious: verify checkable claims.

Video — "Why does AI 'hallucinate'?" (BBC Ideas / accessible explainer)
🔗 https://www.youtube.com/watch?v=hfIUstzHs9s
Why it earns the click: a short, accessible video on why LLMs produce confident wrong answers — exactly the mechanism from lecture, explained for a general audience.
⏱ ~5 min

④ Search vs. AI & The Turing Test

Maps to Lecture Segments 5–6. Search finds existing pages; AI writes new text. The Turing test (Turing, 1950) is a behavioral test — it checks whether a human can tell the difference, not whether the machine "understands."

Reading — "Computing Machinery and Intelligence" — overview (Stanford Encyclopedia of Philosophy)
🔗 https://plato.stanford.edu/entries/turing-test/
Why it's assigned: a thorough, reliable academic overview of the Turing test, its history, and the debates it sparked (including the Chinese Room). Skim the first two sections — "The Original Imitation Game" and "Turing's Prediction" — for this week; the deeper sections are optional.
⏱ ~12 min (skim sections 1–2)

Note on the original paper: Alan Turing's 1950 paper "Computing Machinery and Intelligence" was published in Mind, vol. 59, no. 236, pp. 433–460. If you want to read a portion of the original, it is freely available at many library databases and archived sites — ask a librarian or search "Turing 1950 Computing Machinery and Intelligence Mind."

⑤ Try It Yourself — The Token Visualizer

Maps to the live demo in Lecture Segment 2. See exactly how a model breaks your text into tokens.

Tool — OpenAI Tokenizer
🔗 https://platform.openai.com/tokenizer
How to use it: paste a paragraph you've written, a tweet, or a short article excerpt, and watch the color-coded token breakdown. Notice how rare words, long words, and names get split differently than common words. This makes the abstract concept concrete in about two minutes.
⏱ 2–5 min

Pick-one quick path (≈18 min total)

In a hurry? Do exactly these two and you'll be ready for the quiz:
1. Watch "Why does AI 'hallucinate'?" (group ③) — the mechanism and the cure in 5 minutes.
2. Read "What Is a Large Language Model?" (group ①) — the tokens + generation story in 8 minutes.
3. Skim the Stanford Encyclopedia entry on the Turing test (group ④), sections 1–2 only.

Approved assistants (reminder)

You only need one for this week's tutorial, practice, discussion, assignment, and Studio. Free tiers are fine.

• ChatGPT (OpenAI) 🔗 https://chatgpt.com
• Claude (Anthropic) 🔗 https://claude.com
• Gemini (Google) 🔗 https://gemini.google.com
• Copilot (Microsoft) 🔗 https://copilot.microsoft.com

Heads-up (links rot): these point to outside sites that occasionally move or rename pages. If a link ever fails, let Prof. Quinn know and use a known-good search (e.g., "IBM context window AI") or the Wikipedia article on "Large language model" as a fallback. Nothing here is downloaded or redistributed — all resources stay as links to their original sources.