Last Year I Thought Hacktoberfest Was Stupid. This Year, I Built Look Again💀

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 1: Touch Grass

So last year, an online dev friend introduced me to Hacktoberfest, I checked it out. A month-long event, open-source contributions, pull requests… and my first thought was: "huh, that's it? This sounds ridiculously easy. What kind of competition is this? 🥱"

I made two PRs, i think🤔, and went back to doing my own thing.

Fast-forward to 2026.

This time, I came prepared. I knew how PRs worked, I knew GitHub, and I was convinced I had this whole thing figured out.

DigitalOcean, you dig? 😶😂

but Hacktoberfest 2026 came with a different kind of challenge as we all knoww!!

The Week 1 challenges theme was Touch Grass - building with open-source AI to help people spend less time on screens and more time experiencing the world around them.

and now I had an actual question to solve: what could I build that would make sense for this theme?

I didn't want to build just another chatbot or add AI to an application for the sake of it. I wanted to build something that actually fit the challenge, that's how I started exploring an idea: what if AI could give people a reason to put their phones away, look around, and notice things they might otherwise overlook?

and that idea became Look Again. 🌱

What I Built

Look Again is an AI-powered curiosity companion that gives you a question to take away from your screen and into the real world.

The idea isn't complicated.

You open the application, choose how you want to observe your surroundings, and receive one AI-generated question, maybe the question encourages you to notice an overlooked detail, observe a change, find a connection, pay attention to people, or approach something familiar with a little more curiosity.

Then you put your phone away.

You explore the question at your own pace, when you're ready, you return and describe what you noticed.

Look Again helps turn your reflection into a Field Note - a small record of an observation you can revisit later.

The experience follows five steps:

1. Choose: Select an attention mode.
2. Discover: Receive a focused AI-generated question.
3. Observe: Step away from the screen and explore.
4. Reflect: Describe your experience in your own words.
5. Remember: Save the reflection as a Field Note.

That's the core idea.

I'm not trying to make people spend more time inside Look Again. I'm exploring whether a small interaction with software can encourage someone to pay attention to something outside it and I deliberately didn't build the experience around streaks, leaderboards, or endless missions.

Imagine creating an application to help people spend less time on their phones and then making them addicted to completing its daily challenges.

Congratulations, you've invented the problem you were supposed to solve , wth brooo 💀

Instead, Look Again gives people a starting point and lets them take it from there.

The application doesn't need to be the most interesting thing in the room.

Demo

Project demo: LOOK-AGAIN

Code

GitHub repository: REPOSITORY

The project brings together a React frontend, an Express backend, AI-powered generation, and journal persistence, the repository is where you can explore the implementation and how the different pieces fit together.

How I Built It

Okay, enough talking about curiosity. Let's get into the engineering. 😂

Look Again uses React and Vite on the frontend, Node.js and Express on the backend, SQLite for journal storage, and Google's Gemma model through the Google GenAI SDK.

Here's the stack:

React + Vite - Frontend application
Tailwind CSS - Styling
Motion - Interface animations
Lucide React - Icons
Node.js + Express - Backend API
Google GenAI SDK + Gemma - Prompt and Field Note generation
SQLite via sql.js - Journal storage
Render + Vercel - Backend and frontend deployment

The AI layer

I integrated Google's Gemma model through the Google GenAI SDK for two main tasks:

• Generating attention prompts: The model produces a question based on the selected attention mode and relevant context.
• Structuring Field Notes: The model helps turn the user's reflection into a more organized note.

The idea is that the AI shouldn't invent the observation, it shouldn't pretend it experienced the user's surroundings or write a fictional reflection on their behalf.

The user provides the actual experience. The model helps shape the interaction around it.

The AI is the guide, not the protagonist.

The trade-off is that AI generation depends on network access and the availability of the external service. API credentials, rate limits, and usage costs also need to be considered.

Why Does Open Innovation Matter?

I didn't want to use AI just to put an AI label on a project, i wanted to explore how an open-weight model could support a particular experience: helping someone approach their surroundings with fresh curiosity.

Gemma gave me a way to experiment with generated prompts and reflections while concentrating on the product itself.

Open-weight models and open-source tooling give developers more freedom to experiment, learn, and make informed technical decisions.

There's an important distinction, though: Look Again uses hosted inference, so the current implementation isn't entirely offline. The model integration depends on network access and the external service.

But that doesn't diminish the value of the experiment. It makes the architectural trade-offs clearer.

For me, open innovation is about being able to explore what technology makes possible and then decide how to use it meaningfully.

Prize Categories

Best Use of Gemma - Integrated Google's Gemma model through the Google GenAI SDK to generate attention prompts and help transform users' reflections into meaningful Field Notes.

What I Want to Explore Next

Look Again is an experiment, and there are still things I'd like to improve, i want to explore how Field Notes could connect observations across different moments, making it easier to revisit an experience and discover patterns over time.

also want to strengthen persistence and deployment reliability.

Most importantly, I want to find out whether the idea actually works for people.

Do the prompts lead to observations they wouldn't otherwise have made? Does reflecting on those observations feel meaningful? Can the experience encourage curiosity without becoming another obligation?

Those questions matter more to me than adding features simply to make the application bigger.

One Last Thing, DigitalOcean...

Last year, I came here thinking, "PR count? Easy."

This year, I came back wanting to win and ended up building something that made me think differently about the purpose of an application.

I still want to win, by the way. Let's not get too philosophical here. 😭😂

But the interesting part of this challenge has been having a reason to build something beyond the usual portfolio checklist.

Look Again explores a simple idea: technology doesn't always need to demand more of our attention. Sometimes, it can give us a reason to direct that attention somewhere else.

A question. An observation. A Field Note worth remembering.
and then, the freedom to close the app.

Happppy Hacktoberfest!✌