SPARK Speeches 2026 • Tech Novice Tools

Your Code Can Talk.
Can You?

You spent time collaborating with an AI model, wrangling prompts, debugging outputs, and assimilating code you didn't write line-by-line. That experience is genuinely interesting. Your audience — classmates who went through the same thing and a teacher who has seen it from every angle — wants to hear your story. Not a recitation of features. Your story.

The 15–30 minutes ahead of you are a conversation, not a report. The tips below are drawn from the world's foremost researchers and practitioners in communication, psychology, and technology — applied specifically to the challenge of presenting an AI-assisted project to a technical peer audience. Read them. Use them. Return to them after your first run-through.

One truth before you begin: Brilliant programmers who can't explain what they built are far less effective than average programmers who can. The ability to communicate is a technical skill.

10 Principles for the CS Presenter

Pull from expertise in neuroscience, storytelling, and the TED stage — filtered for the unique challenge of presenting a project you co-built with artificial intelligence.

🔥
1
Lead With "Why," Not "What"

Before you show a single screen, answer the question your audience is already asking silently: "Why should I care about this?" Open with the problem that motivated the app — a frustration, a gap, a moment of curiosity. Make the audience feel the need before you present the solution. Once they feel the problem, every feature you demo carries emotional weight.

★ Simon Sinek — Start With Why
📚
2
Structure as a Story, Not a Report

The most memorable presentations follow a narrative arc: What is (the status quo and its friction) → What could be (your app's possibility) → How we get there (the demo and the learning). Audiences are wired for story. When you present features as bullet points, the brain stores them in working memory and discards them. When you wrap them in narrative, they stick.

★ Nancy Duarte — Resonate
🎮
3
Demo First. Impress Early.

Steve Jobs never spent the first ten minutes explaining what the iPhone was going to do. He showed it. Show your app doing its most impressive thing within the first three minutes. You'll capture attention when it's at its highest, and the rest of the talk becomes explanation of something the audience already wants to understand. Always have a backup — a screen recording, screenshots — for when live demos break.

★ Carmine Gallo — The Presentation Secrets of Steve Jobs
🐛
4
Slay the Curse of Knowledge

You have spent weeks with this code. You know what "the state object" means, why you chose that API, and what a callback is. Your audience may not — and even if they do, they haven't lived in your codebase. The "Curse of Knowledge" is the cognitive bias that makes experts forget what it felt like to not know something. Fight it consciously: define your terms once, use analogies, and never assume shared context.

★ Chip & Dan Heath — Made to Stick
🤖
5
Own Your AI Partnership Publicly

You didn't write every line — and that's the point. The most intellectually honest and genuinely interesting thing you can do is explain how you collaborated with AI. What prompts worked? Which outputs were wrong and how did you know? What did the AI teach you inadvertently? Showing that you can critically evaluate, debug, and assimilate AI-generated code is a far more sophisticated skill than writing it yourself. Make that case out loud.

★ Ethan Mollick — Co-Intelligence
🧤
6
Silence Is Not Weakness — It's Power

The instinct under pressure is to rush. Filling every gap with "um," "like," and "so yeah" signals anxiety to your audience and obscures your competence. A deliberate two-second pause after a key point lets the idea land. Research from Harvard psychologist Amy Cuddy shows that even your own perception of your competence changes when your body language and pace project calm. Slow down. Breathe. You earned the room.

★ Amy Cuddy — Presence
🖼
7
Your Slides Serve You. You Don't Serve Them.

A slide packed with bullet points tells the audience to read, not listen. The moment they read, they stop hearing you. Follow the principle of "one idea per slide" — use visuals, screenshots, and short phrases as cues, not scripts. If you can read your entire slide aloud without adding anything, the slide is doing your job and you're redundant in the room. Be the insight; let the slide be the anchor.

★ Garr Reynolds — Presentation Zen
🔍
8
Anticipate the Hard Questions

Your teacher will ask about implementation decisions. Your peers will ask "did you think about doing it this way instead?" The worst Q&A responses are defensive ones. Prepare three categories: (1) Questions you know cold, (2) Questions you partially know (say what you know, name the gap), (3) Questions you genuinely don't know yet (say "I don't know, but here's how I'd find out" — this is an intellectually mature answer).

★ Adam Grant — Think Again
😂
9
Humor Is a Technical Skill

The most effective tech communicators — from Richard Feynman to Sal Khan to the best TED speakers — use humor deliberately. It is not frivolous. Laughter releases dopamine, which improves memory retention in your audience. You don't need to be a comedian. Self-deprecating honesty about a bug that stumped you for three days, a prompt that returned something hilariously wrong, or a feature that "works on my machine" lands exactly right with a CS audience. Lean into it.

★ Carmine Gallo — Talk Like TED
🏃
10
Practice Out Loud. Every Time.

Running through your talk silently in your head produces a false sense of readiness. The mouth, the voice, and the cognitive load of speaking are entirely different systems. Stand up. Say it out loud. Time it. Record yourself once — watching playback is uncomfortable, but it is the single highest-ROI practice technique available. Your voice will change, your pace will slow, and your transitions will sharpen in ways that silent rehearsal never achieves.

★ Research Consensus — Cognitive Load & Verbal Processing

The Acronym: SPARK CODES

Yes, it's corny. That's why it works. Ten principles — one phrase you already know how to spell. Because you're CS students, and you literally write code.

S
P
A
R
K
C
O
D
E
S
S
Story Over Bullet Points

Structure your talk as a narrative arc, not a feature list. Audiences remember stories; they forget slides.

#2
P
Pause on Purpose

Deliberate silence is power. Let your best ideas land before rushing to the next one.

#6
A
Anticipate the Q&A

Prepare for the hard questions before you walk in. Know what you know, know what you don't.

#8
R
Run the Demo First

Show your most impressive moment within three minutes. Earn attention before you explain.

#3
K
Kill the Curse of Knowledge

You know too much. Remember what it felt like to not know it, and explain from there.

#4
C
Claim Your AI Collaboration

Own the partnership publicly. Explain what the AI got right, what it got wrong, and how you knew the difference.

#5
O
Open With Why

Before a single feature, answer the silent question every audience has: "Why should I care about this?"

#1
D
Design Slides That Serve You

One idea per slide. If the audience reads your slides, they stop listening to you. Be the insight.

#7
E
Earn Laughs Intentionally

Humor is a technical skill. A well-placed moment of honesty about a ridiculous bug is worth ten polished slides.

#9
S
Say It Out Loud to Prepare

Rehearse verbally, every time. The mouth is a different system than the brain. Your talk doesn't exist until you've spoken it.

#10
“
"The single biggest problem in communication is the illusion that it has taken place."
George Bernard Shaw — cited in virtually every communication research curriculum

Survival Guide: When the Questions Come

Q&A is not an attack — it's evidence that people were listening. Treat each question as a gift. Here's how to handle the most common scenarios.

🔴 "Do you actually understand the code you used?"

Don't deflect. Walk to a specific function or block and explain what it does in plain English. You don't need to have written it — you need to be able to read and reason about it. That's a professional skill.

🔴 "Why didn't you just use [different technology]?"

Acknowledge the alternative: "That's a valid approach. I chose this because..." Then explain your reasoning — constraints, what you knew at the time, what the AI suggested. Decision-making under uncertainty is worth defending.

🔴 "What would you do differently if you started over?"

This is a gift. Answer it honestly and specifically. Reflection demonstrates maturity. "I'd start with a cleaner data model" or "I'd spend more time on the prompt architecture first" signals genuine learning.

🔴 "What was the hardest bug to fix?"

This is your moment to be human and relatable. Tell the story. Name the bug, describe the confusion, describe how you broke through. The room will be with you — everyone in CS knows that feeling.

🔴 "What did the AI get wrong?"

Have two examples ready. Being specific here proves you were the critical thinker in the collaboration, not a passive executor. "The AI gave me a function that looked right but had an off-by-one error in the loop condition" is a strong answer.

🔴 "I don't understand what you just said."

Don't repeat the same words faster and louder. Re-approach from a different angle or use an analogy. "Let me put it another way..." is a sign of a strong communicator, not a weak one. Clarity is a skill you practice.

How This Page Was Made

This page was generated with AI assistance — and that process is itself a model for how we use AI in this course: not to replace thinking, but to accelerate it. Below is the exact prompt that produced the first draft of this content, followed by notes on what was evaluated, kept, and refined by a human teacher and developer.

USER PROMPT → GitHub Copilot (Claude Sonnet 4.6) • May 5, 2026
Claude, you are a passionate speech teacher and technical writer. You understand that to be successful, even 'nerds' and 'computer geeks' must be able to capture an audience's attention when giving presentations. They must be able to share sometimes complicated materials with poise, confidence, clarity and yes, even humor, even if they are otherwise brilliant programmers and developers. This site is aimed at high school students who are taking CS courses and are using AI to help them develop apps. Since the end product is no longer a program as such, it's actually the starting place, they must be able to assimilate the code and then present the app to their peers and share the 'big ideas' they learned during the process. These 'Spark Speeches' are typically in the 15-30 min range. They will be fielding questions from their teacher and peers alike. Please create a gettingStarted.html app that features the top recommendations you have for giving such a talk. Pull from your knowledge of effective communication from experts in the field of technology, communication and psychology. Please include this prompt in the discussion so people can see how we are utilizing AI in our development of this app.
🤖 What the AI Contributed
The AI drew on its training across thousands of books, academic papers, and instructional resources — synthesizing voices from Sinek, Duarte, Gallo, Cuddy, Heath, and Reynolds into a coherent framework aimed at a specific audience. It generated the tip text, Q&A scenarios, visual layout structure, and HTML/CSS code matching the existing site design system — all in a single pass.
🧠 What the Human Contributed
The prompt itself required contextual expertise: knowing which audience (high school CS students), which constraint (AI-assisted build process), and which tone (encouraging but rigorous) would produce useful output. Evaluating the result, checking expert attributions for accuracy, judging whether the advice is genuinely actionable for this specific context, and deciding what to publish — that judgment is irreplaceable and belongs to the teacher and student, not the model.
💡 The Bigger Lesson
AI doesn't remove the need for expertise — it changes where expertise is applied. A poor prompt from an uninformed user would have produced generic, useless advice. A precise, contextually rich prompt from someone who knows their students, their curriculum, and their goals produces something genuinely useful. The skill being developed in this course — prompting, evaluating, assimilating, and presenting AI output — is exactly the skill the professional world is now competing to hire. You are practicing it right now.
Simon Sinek — Start With Why Nancy Duarte — Resonate Carmine Gallo — Talk Like TED Carmine Gallo — Presentation Secrets of Steve Jobs Amy Cuddy — Presence (Harvard) Chip & Dan Heath — Made to Stick Adam Grant — Think Again (Wharton) Garr Reynolds — Presentation Zen Ethan Mollick — Co-Intelligence George Bernard Shaw