Back to all discussions

  •   about 2 months ago

If Technical Execution is 40%, how is it evaluated without testing code?

Genuine question for the community, not a complaint.

The scoring rubric weights Technical Execution at 40% — the single heaviest criterion. It asks: "Does the project demonstrate quality application development? Is the code of good quality and is it functional?"

The latest update states that judges "are not required to test your project live" and will be "reviewing your demo video."

Meanwhile, Google is retiring Gemini 3 Pro on March 9. Projects built on Pro will lose live functionality. The official advice is to avoid editing your submission.

So here's the situation:
A project with clean architecture, real error handling, and a working fallback system looks the same in a 3-minute video as a prompt wrapper with a polished UI. The only way to tell them apart is to run the code or read the repo.

If judges evaluate Technical Execution primarily from videos, the 40% weight rewards presentation, not engineering. That's a different competition than the one described in the rules.

I'm not writing this because my project is affected — I built on Flash + Gemma with a hybrid architecture specifically to avoid single-model dependency, and my demo still runs. I'm writing this because the evaluation method should match the weight it carries.

A video shows what something looks like. Running code shows what something is.

— Che, Solo developer, Project RE, Taipei Taiwan

  • 5 comments

  •   •   about 2 months ago

    A polished promotional video is not the same as a technical demonstration.
    In less than 20 seconds it is usually obvious whether someone is showing a real system or just a UI wrapped around prompts. When a project is real, the demo naturally exposes signals of engineering work such as system behavior, architecture decisions, error handling, or operational flow.

    Experienced judges tend to recognize these signals very quickly. A marketing style video focuses on appearance. A technical demo focuses on how the system actually works.

  •   •   about 2 months ago

    I agree experienced judges can read signals from video. That's not the issue. The issue is structural: the rules require entrants to produce quality, functional code — then tell judges they don't have to verify it. Whether judges can tell from video is about their skill. Whether the rules require them to look at code is about the process. My question is about the process, not the judges.

  •   •   about 2 months ago

    agree, Chieh. I'm also a solo developer with time constraints, so I chose to focus more on the technical execution than the demo video because of the rubric. Had I known the requirement that the project be functional would be dropped, I might have spent more time on the demo video. But I get that they had way more submissions than they anticipated.

  •   •   about 2 months ago

    Don't you have your github repo linked to your submission? A quick glance on your project can tell more than any video.

  •   •   about 2 months ago

    Beck — exactly. If the rubric says 40% Technical Execution, it's rational to prioritize engineering over production value. The rubric shaped our decisions. If the evaluation doesn't match the rubric, those decisions were made on false premises.
    And to your point about volume — the submission period ran from December 17 to February 9. Almost two months. Registration numbers are visible in real time. They had nearly 60 days to watch the count climb to 35,000+ and adjust the judging process accordingly. Volume on this scale isn't a surprise — it's a planning failure.

    Almin — yes, my repo is public and linked. That's precisely the point: every participant was required to submit a public repo/live demo. The question isn't whether code exist — it's whether the process requires judges to look at them. Right now, the answer is no.

Log in or sign up for Devpost to join the conversation.