HackEurope 2026: A short rant on AI and hackathons - Insomniac's Log
HackEurope 2026: A short rant on AI and hackathons
By<br>Antonio Cheong<br>on<br>Feb 23, 2026<br>on<br>Permalink.
HackEurope is over. In many ways, it was a complete shitshow (vibe coded<br>inaccessible UI for participants, lots of delays, miscommunications, and other<br>issues too many to list). But now that the caffeine overdose and sleep<br>deprivation is over, I can say that there were actually some important lessons.
TL;DR:
Front-end is almost everything. There is 0 burden of proof that your project<br>is actually functional or that it has any practical application. As long as<br>it looks cool, investors and non-technical people will eat that up.
Choose your track wisely. Make sure that the track sponsor IS ACTUALLY AT<br>YOUR LOCATION. Most people were under the impression that tracks were<br>per-country when in fact there was a single €1000 prize shared across the 3<br>countries and the sponsor wasn't actually operating in some.
Choose a problem that is easy to explain. There were 2 minutes to explain. It<br>is a losing game whether or not you explain context. Non-technical people<br>will tune out confused regardless. We were extremely lucky with 2/3 of the<br>evaluators actually knowing about open-source supply chain attacks and being<br>excited about our solution.
Follow the trends. All winners had "AI" as a significant part of their<br>solution.
That being said, I personally wouldn't follow my own advice. I went in with the<br>goal of building something that I would want to maintain long term. Not just AI<br>slop (I fucking hate Lovable).
AI encourages conformity and kills creativity
A solid 90% of the projects there were just vibe coded slop. Even the ideas were<br>AI. You can tell when multiple people implemented the exact same idea with the<br>exact same title, description, and implementation.
While people call me a luddite, I do not particularly hate AI as a tool. My<br>problem is that it has significantly lowered the bar for certain project types<br>and therefore incentivize people who would have otherwise built something cool<br>to instead fit into a mold constrained by the capabilities of AI.
A lot of cool ideas are out of distribution from the training data, and those<br>rarely show up at hackathons anymore. The AI says they're "too hard" and people<br>simply avoid these.
The grand winner was an idea to use LLMs for predicting wildfires caused by<br>lightning strikes, and subsequently using LLMs to orchestrate drones to do cloud<br>seeding and prevent wildfires. Cool UI and all, but there was (at least from<br>observation), nothing actually behind it.
In a now edited post by Anthropic's head of Startup Sales, he mentioned that the<br>winning team (LLM cloud seeding) had only 1 software engineer and 3<br>non-technicals. The accessibility is cool to see, but it is not expected at all<br>for any of these projects to exist long-term. Just a marketing stunt to claim<br>that code is now a commodity.
It feels like hackathons used to be a place where real startups are made or at<br>least a proxy for the ability of individuals. Now, with everything being<br>front-end only demos, there is no expectation at all for any follow up, and<br>nothing is said of ability except for pitching and trend chasing.
Some other funny ideas I saw:
Stopping AI prompt injecting by scanning every prompt with an LLM (there were<br>multiple duplicates of this)
Using LLMs to control satellites and move them when a Russian satellite gets<br>too close (winner)
"Palantir for tech teams". "A real-time security guardian sitting silently on<br>every dev's machine, scanning their screen, code, and communications to<br>proactively prevent vulnerabilities."