30 minutes a day. 15 days. Learn AI by actually using it.
Stop saving AI posts. Start using AI.
The first Founding Sprint is completely free, celebrating The Arthi AI Collective crossing 300 community members.
Day 5 highlights
Inside the Sprint
A little of what we tried, made and learned together.
Day 5: Plan a home for your chatbot (0:44): A website can give your existing chatbot a home alongside your portfolio, business and offerings.
Day 5: Start your website with pen and paper (0:38): A quick paper or tablet sketch helps you express your own layout and personality before asking an AI tool to build.
Day 5: Sketch the structure of your website (0:45): Sketch the navigation, product cards and footer before building. Make the structure you want visible.
Day 5: Attach your sketch to the right project (0:27): Attach your sketch to the project you want to extend. Use a separate project when the new website should stand alone.
Day 5: Tell AI what your website needs (0:42): Describe what you sell and what the page needs: products, animation, a portrait and a short introduction.
Day 5: Save your code in GitHub (0:22): Once the first version is built, save its code in a new GitHub repository.
Day 4: Chatbot vs AI agent: what changes? (0:20): A chatbot can answer questions. An AI agent can use connected tools to take action toward a goal.
Day 4: Your AI agent needs a clear goal (0:49): Define the job before adding tools. An email triage agent, for example, has a clear purpose: surface important messages and deadlines.
Day 4: What does an API connect? (0:16): An API connects your application to another service, tool or model. It is the bridge between your chatbot and the tools it needs.
Day 4: The Gmail + Telegram build prompt (0:25): Ask for two clear outcomes: let visitors contact you by email, and send conversation summaries to Telegram. Start by describing the behavior you want.
Day 4: Create your first Telegram bot (0:39): Use BotFather to create a Telegram bot with /newbot. Give it a name and a unique username ending in bot. Keep the token private.
Day 4: Adding API keys is only one part (0:37): A tool integration also needs backend code. In this build, integrations.py holds the email and Telegram connection logic alongside the app changes.
Day 4: Give your coding agent the actual error (0:31): When a test fails, give your coding agent the actual error and the intended result. Keep credentials in private configuration and verify the next test.
Day 4: When your chatbot should hand over (0:46): If a question falls outside your chatbot’s context, it can hand the conversation to you. A Telegram notification is one way to support that human handover.
Day 3: Turn your profile into a personal AI chatbot (0:23): Build a personal AI chatbot around information you choose to share about yourself. Day 3 starts with a profile and a clear goal for the chatbot.
Day 3: A plain-English brief for a profile-based chatbot (0:39): A useful chatbot brief names the source, sets an answer boundary, asks for a simple interface and requests source references. This is the brief used for the Day 3 profile chatbot.
Day 3: APIs explained with a restaurant waiter (0:39): An API is easier to picture through a restaurant: a waiter carries your request to the kitchen and brings the result back. A simple starting point for understanding request and response.
Day 3: Your weather app uses an API (0:34): A weather app can request data from a weather service through an API, then display the response. Your phone does not need to measure the outside temperature itself.
Day 3: Test your chatbot with a question the profile cannot answer (0:41): The local profile chatbot cites a skills answer, then returns “not in the profile” for a missing detail. One useful test, alongside further checks for incorrect answers.
Day 3: Keep your API key out of GitHub (0:36): The Day 3 explanation separates the code from the .env file containing the API key. Check what you upload to GitHub and use placeholders when sharing configuration.
Day 3: Set the API key as a deployment environment variable (0:46): The Vercel walkthrough imports the GitHub project and adds the Gemini API key as an environment variable. After deployment, test the live chatbot before sharing it.
Day 3: Build your chatbot, then share what you learned (0:35): The Day 3 challenge is to build a profile chatbot, move the code to GitHub, deploy and share your learning. Test the live app and keep private information out of anything you share.
Day 2: Stop asking AI to explain. Ask it to build. (0:23): A small change in the prompt turns an explanation into an app-building request. Watch the Canvas prototype appear in our live session.
Day 2: A school timetable app from a simple prompt (0:44): We tried a school timetable idea in Gemini Canvas and explored working days, subjects and CSV export. This is a prototype, with more testing needed before real use.
Day 2: Share your Canvas prototype with a link (0:35): A Canvas sharing link gives you a way to show a prototype and gather feedback before moving into a separate hosting workflow.
Day 2: GitHub explained using Google Drive (0:45): New to GitHub? Here is the beginner-friendly analogy from Day 2: a place to store your code, with repositories you can also explore.
Day 2: Hosting your first app is less mysterious than it sounds (0:28): The basic workflow is easier to picture once you separate the steps: your code, a GitHub repository and a hosting service. This clip explains the idea behind deployment.
Day 2: It deployed. Then the app showed a 404. (0:26): The deployment screen gave us a link, but the live app showed a 404. A useful reminder from our session: open the actual URL and check the result, then inspect the logs.
Day 2: Share your AI learning journey (0:20): Share what you are building and learning. This is the invitation from Day 2 to document your AI learning journey and support each other along the way.
Day 1: AI made my avatar. Then this. (0:28): A live reveal, a surprised reaction, and an imperfect result we can learn from. The avatar in this demo is AI-generated.
Day 1: Your prompt can be simple. (0:22): A reminder from Day 1: you can begin with ordinary language. You don't need a complicated prompt to get started.
Day 1: Start practising today. (0:20): A little encouragement to make time, try something, and build the habit together.
Day 1: One photo. Different angles. (0:41): A storyboard prompt from the session: a useful starting point for creating AI video with a more consistent character.
What happens every day
One useful, current idea: A single AI idea or tool that actually matters right now, not a fixed syllabus written months ago.
Live explanation: A short, plain-language walkthrough, no jargon, no assumed background.
Try or build it together: You open the tool and use it live, in the session, with Arthi and the group.
A tiny daily challenge: One small thing to try before the next session: momentum over homework.
What we might cover
The curriculum is intentionally dynamic: we cover what's useful in AI right now.
Prompting, AI research, Claude, ChatGPT, Gemini, Codex, AI agents, Automations, New AI tools, Real workplace workflows
How the 15-day sprint works
Day 1: Join: Register free, get the meeting link, and show up for day one, no setup, no prerequisites.
Days 1–15: Daily live AI sessions: 30 minutes a day, live. One idea, one try-it-yourself moment, one tiny challenge, every day for 15 days.
Along the way: Earn badges as you go: Show up, and streak/finisher badges unlock automatically. Build something real, and Arthi can award AI Builder recognition.
Day 15: Finish, and decide what's next: Wrap the sprint with what you've built and learned, then decide what you want to learn next, together as a club.
Badges
Founding Member: Awarded to everyone who joins Cohort 001: the very first Sprint.
AI Sprint Finisher: Attend 10 or more of the 15 daily sessions.
15-Day Streak: Attend all 15 sessions: the full sprint, no days missed.
AI Builder: Submit a real project or workflow you built, approved by Arthi.