Summary
The first session brought together a mix of curious beginners and a couple of more experienced users. Connor opened with an overview of the major AI platforms, then let attendee questions drive the agenda. The group explored AI for travel planning, Excel help, email drafting, and bracket building — and wrapped up with a live vibe coding demo that turned a simple idea into a working confetti cannon app.
Questions Asked
- What should I be using AI for — where do I even start?
- Can you give an overview of the different AI platforms and apps that are out there?
- Did the AI tailor its suggestions to you because it already knows you're an educator?
- What's an example of something someone has actually used AI for at work?
- What is a large language model — what does that term actually mean?
- Is AI music generation the same basic idea as text-based AI?
- Can you give more examples of what "generative AI" means?
- Is something like a poem written by AI considered generative?
- Have you used AI to do stock trading?
- Is this the free version of Gemini you're using?
- If AI writes code for you, how would you get that code onto a real website?
- What is "vibe coding"?
Demos
- AI platform overview — walked through the three main players (Google Gemini, Anthropic/Claude, OpenAI/ChatGPT) and why any of them is a fine starting point for beginners
- Beginner use cases via Gemini — asked Gemini to generate use cases for a new user; showed how the model tailored suggestions based on Connor's profile and how to get a more generic result
- Travel itinerary (Croatia) — built a 3-day Split itinerary from a simple prompt; showed how the model holds context across follow-up questions
- Personalized travel planning with an interview prompt — asked the AI to interview Connor with 5 questions before building a Montana itinerary; showed how this approach produces dramatically more relevant results than a one-shot prompt
- Packing list and trip cost estimate — extended the Montana itinerary into a packing list (iterated format from paragraph → itemized → table) and a rough total cost breakdown
- March Madness bracket building — showed two brackets built with different prompts: one optimized to win, one with high risk tolerance; illustrated how framing changes outputs
- Vibe coding: confetti cannon — used Gemini Canvas to build a working confetti cannon simulator with zero hand-written code; iteratively added features (spelling "Connor" in confetti, cannon graphic, team color presets, fireworks mode) based on group suggestions
- Image generation — generated a scenic Montana mountain bike trail image and a caricature of Connor; showed that results vary and aren't always reliable
- Connor's vibe-coded projects — quick showcase of what's possible: a learning hub website, personal portfolio site, a Secret Santa app, and the in-progress Party Quests event app
Key Takeaways
- Start with tasks where you already have some expertise — that way you can actually tell if the AI is giving you good answers
- AI works best as a back-and-forth conversation, not a one-shot query; most good outputs come after several rounds of refinement
- Ask AI to interview you before tackling a complex task — it surfaces the details the model needs to give you something genuinely useful
- These tools update constantly; features that didn't exist last week may be there now
- You don't need to know how to code to build functional software — vibe coding lets you describe what you want and iterate from there
- AI generates the most probable answer, not necessarily the correct one — always verify anything that matters