Winners Announced:
Congrats to the "Built with Google Gemini: Writing Challenge" Winners!
...
Some comments have been hidden by the post's author - find out more
For further actions, you may consider blocking this person and/or reporting abuse
A writing challenge ! :-D. DEFINITELY trying this challenge <3
Good luck!
Thank you !
Quick question: if my article is fully in English and the app interface is also in English, but the AI-generated insights are in Ukrainian for its target audience, is it still eligible for prize consideration?
@ben @jess cc
@curiousvlxd not a problem!
You and me both!
Can I write about several projects in one post? I've been using Gemini CLI a lot these days in older projects
Yes, definitely!
Awesome 😍 😍😍
Oh wow, how exciting!!! What a cool prize!!!
I'm really looking forward to reading everyone's posts. Good luck!
I'm thrilled we're able to offer such a cool prize for a writing challenge - this is a first!
On last minute: dev.to/pengeszikra/the-royal-gazet...
this is just a vibe rush under 2h
Another LLM challenge, another idea from the past (1987), that will be interesting.
This is a nice framing for a challenge - especially the emphasis on reflection instead of just showcasing outputs.
What I find interesting about using Gemini (or any LLM) in these projects is that the most valuable part usually isn’t what you build, but where it breaks: prompt boundaries, evaluation gaps, tooling friction, and the moments where the model’s strengths force you to rethink your own architecture.
Treating the write-up as a post-mortem rather than a demo recap feels like the right direction. Curious to see submissions that go beyond “here’s my app” and actually unpack the engineering decisions behind it.
@lhabacuc couldn't agree more. I think you're going to really like one of the new writing features we're piloting internally and are hoping to launch in a few weeks. 👀 Surfacing this exact decision flow is critical to understanding software these days.
This definitely made me curious 👀
Does this visualization focus more on transparency of the reasoning process, or on supporting more structured writing?
I saw overestimate the due date, so my program will be goes to another challenge. That was too complex for this rush. In my head live a bit longer timeframe, never mind.
@ben, @theycallmeswift, @jess I had a small query -- do the tags (like "#") attached to submissions affect judging?
I realized I forgot to include the
#devchallengetag in my submission, so I updated the tags on my post today (Kaizen). Before that, among the other tags, it had a different tag,#productivity. And because of the update, the post now shows that it was edited today (Mar 8). Just wanted to check if that’s okay and won’t affect the judging. Thanks and sorry for the direct tag!I posted! Check it out here folks: Fusing NASA Data with AI: How I Built Cosmodex and Wont the MLH Data Hackfest
My submission doesn't show up in Latest Entries, so posting it here :)
I Built a Gemini-Powered History Narrator 6 Months Ago. Here's What I'd Tell You Now.
Nikoloz Turazashvili (@axrisi) ・ Feb 28
I think the template linked here has the wrong tags -- swap out
mlhreflectionsforgeminireflections(that's what the template link from dev.to/challenges/mlh-built-with-g... uses)I just added geminireflections. weird. I used the template when creating this article :)
let's see if it pops up. thanks!
Built 25+ Gemini-powered projects — from a Chrome extension with Gemini Nano to a full multi-agent storytelling system on Cloud Run with ADK (with 10ms routing logic).
The post has something personal, something technical, and a look at what's next. If you read it to the end — thank you, and I'd love to hear your thoughts in the comments. 🙏
G.E.M.I.N.I.: I Had 3 Hours of Electricity. I Shipped Anyway.
And congrats on joining MLH! 🎉 Bridging online knowledge-sharing with hands-on building — that's exactly what this community needed. Thank you to the organizers for this challenge.
Good luck everyone!🍀
This was a really nice writing challenge. My first dive into LLMs wasn't as a coding partner or a chat bot, it was as a tool to help my automated matching system better understand the films that are on at cinemas. Gemini made that really simple to try out.
Since then, I've grown to use LLMs more and more (generating code, documenting systems, managing priorities), but I want to keep that foundation: the right tool for the right job, not LLMs for everything. You can read my writeup of how I use Gemini in Clusterflick in Three Things I Learned Using LLMs in a Data Pipeline
when will be the results revealed ?
supposedly today later, by US time I guess :)
they just put notification that it's delayed to 19th March :)
Hello, I just wanted to let you know I don't think the link in the template is working. Just letting y'all know so projects can be found!
Introduction
Programming is not just solving problems—it’s a constant battle with the human brain. Developers spend 20–30% of their time not on logic, but on syntax traps: where to put a bracket, how not to mix up variables, how not to forget task order, how not to drown in 300 lines without a hint. Cognitive load piles up—the brain holds at most 5–9 items at once (Miller’s rule), while code demands 15–20. Result: bugs, burnout, lost productivity, especially for beginners.
Modern languages (Python, JavaScript, C++, Rust) offer tools for performance—async, lambdas, match-case—but none for the brain. There’s no built-in way to say: “do this first, then that”, “this matters, this is noise”, “roll back five steps”, “split into branches and merge later”. It all stays in your head—and it breaks.
We propose a fix: seven universal meta-modifiers—symbols added to the core of any language as native operators. Not a library, not a plugin, not syntactic sugar. A new abstraction layer: symbols act as a “remote control” for the parser, letting humans manage order, priority, time, and branching without extra boilerplate.
$ — emphasis, | — word role, ~ — time jump, & — fork, ^ — merge, # — queue, > / < — resource weight. They don’t break grammar: old code runs fine, new code breathes easier.
The concept emerged from a live conversation between human and AI: we didn’t run it on a real parser, but already used the symbols as meta-commands to describe logic. This isn’t a test—it’s a proof-of-concept at the thinking level.
The goal of this paper: show these seven symbols aren’t optional—they’re essential. They cut load by 40–60%, slash errors, speed up learning. Not for one language—for all. In five years, any coder should write “output#1-10 >5” without pain. This isn’t about us—it’s about a civilization tired of fragile syntax.
With Gemini, I was expecting a coding challenge or a hackathon, but its interesting to see an Article Writing challenge using an LLM, does that mean the prize goes to the writer, or the LLM (which wrote it)? food for thought 🤔
Just posted, it was a lot of fun writing. Great way to reflect!
Thanks for sharing! Excited to participate.
all i want is just to win a free one year gemini ultra plan ngl...
submitted!
This is exciting, thanks. Will definitely participate.
Well, I’ve submitted my entry. Is there a Discord server for this challenge where I can get updates?
Can't wait to see everyone's project writeups
its challenging , meh
Trying....
Gonna participate in the challenge for sure!
a question @jess @ben , can I write about a project that I built in google AI studio for this challenge?
@saloni0512 you may!
Only Gemini cli project or can we also mentioned the project that we make using Google ai studio
I can't see my submission in the latest uploads, so putting the link here just in case! dev.to/bananapocalypse/brickthis-4c8a
Oh wow
nice
Just sent my submission
All the best everyone