February 23, 2026
When vibes beat code
HackEurope 2026: A short rant on AI and hackathons
AI stole the show—devs cry foul as shiny demos beat real code
TLDR: A HackEurope rant says flashy AI demos beat working projects, sparking a brawl over whether hackathons are now beauty contests. Comments split between “AI killed creativity” and “play by the rules,” with a surprise exploit link adding chaos—raising stakes for anyone building beyond buzzwords.
HackEurope 2026’s post dropped like a caffeine bomb: the author says front‑end flash beats real code, sponsors ghosted entire countries, and AI stamped its name on every winning pitch. Cue the comments. selridge cracked, “AI murdered this guy’s submission,” turning the rant into a meme about robot judges. crowfunder backed it up with a war story: pre–large language model (LLM) era vs post‑LLM era hackathons now reward two‑minute demos over working apps. fxwin pushed back: “You sound bitter—AI was the theme,” arguing this has always been the game.
The project itself? A safer “app store for code” that watches how packages behave before you install them—think a smarter gatekeeper for npm—using behavior signals instead of just known vulnerabilities. Commenters loved the idea, hated the odds. The real drama: Is AI killing creativity or just exposing who’s better at pitching? Memes flew about “vibe‑coded UIs,” “Figma‑first judging,” and “two‑minute TED Talks.”
Then dtf dropped a link to a high‑severity Chrome exploit, sending a brief panic and adding to the chaos: was this a demo or a detour? The room split between “hackathons are beauty pageants now” and “play the game, win the prize.” The tea is hot, the vibes hotter.
Key Points
- •Hackathon judging favored polished front-ends and AI-themed projects, with limited time to explain complex ideas.
- •Track sponsors and prizes were shared across three countries, and sponsor presence varied by location.
- •The author’s MVP is a secure package registry replacing npm that tests packages for malicious behavior.
- •Behavioral telemetry is collected via eBPF and analyzed using baselines plus AI or historical data to flag anomalies.
- •Planned expansions include reproducible builds, multi-ecosystem support (PyPI, Maven, Cargo), source-tracing, HTTPS decryption via eCapture, and honeypot data.