Live Nov 18 · Cohort 3Nov 18, 19:00 MSK — free hands-on AI agents live workshop · Live Cohort 3 starts Nov 23Join Nov 18 Live →Cohort 3 (Nov 23)
ReleaseInsight · On RadarHabr AI→ original

The Best Code Didn't Win: What the Dev-to-Dev Hackathon Final Says About Agentic Engineering

At the Dev-to-Dev hackathon, second place went to a project with immaculate code: 65 tests, full typing, protection against vulnerabilities. But a different project won, the one that better captured the essence of agentic engineering. The takeaway: in 2026, writing high-quality code and building agentic systems are already different professions.

AI-processed from Habr AI; edited by Hamidun News
The Best Code Didn't Win: What the Dev-to-Dev Hackathon Final Says About Agentic Engineering
Source: Habr AI. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceHabr AI ↗
Editor-in-Chief
Key metric #1Interactivity and responsiveness matter more than internal architecture
Key metric #2The ability to adapt matters more than typing everything
Key metric #3Fast iteration matters more than full test coverage

At the Dev-to-Dev: Agentic Engineering Challenge final, the project with the best code finished second. It had 65 unit tests, full typing, protection against path traversal, and Docker with a non-root user: an architecture that shows a serious engineer at work. And that is exactly why it did not win.

When code quality isn't enough

The second-place project is one you rarely see at hackathons: a clean codebase, test coverage, and security you had to dig through the details to find. Its author wrote not just code, but code that could be shipped to production tomorrow.

But the hackathon measured something else. The organizers measured agentive engineering, the very thing this high-quality project lacked. The winning entry understood better than any other what they were looking for.

Agentic engineering is a different language

Agentive engineering is a young term (barely a couple of years old), and every team reads something different into it. But this year the final showed clearly that agentic systems demand different priorities. Instead of code purity, the ability to adapt on the fly. Instead of full test coverage, readiness to work with incomplete information. Instead of perfect architecture, iteration and learning from mistakes. It turned out that:

  • Interactivity and responsiveness matter more than internal architecture
  • The ability to adapt matters more than typing everything
  • Fast iteration matters more than full test coverage
  • Working with uncertainty matters more than type guarantees

A split in engineering

The second place is not a judging error. It is a signal of a deep split in how we think about code. Classical engineering: types, tests, documentation, an architecture that keeps 65 tests in check and follows every best practice. Agentic engineering: a quick hypothesis, integration with the outside world, readiness to be wrong and to learn, the ability to think under uncertainty.

This is not only about code. It is about two different worldviews on what good software is.

What this means

For developers, it means the mindset is changing. Classical engineering skills remain valuable, but the agentic paradigm requires a completely different approach to design. A global split in how we write systems is not far off.

Story

Key Questions Answered

When code quality isn't enough?

The second-place project is one you rarely see at hackathons: a clean codebase, test coverage, and security you had to dig through the details to find. Its author wrote not just code, but code that could be shipped to production tomorrow. But the hackathon measured something else. The organizers measured agentive engineering, the very thing this high-quality project lacked. The winning entry understood better than any other what they were looking for.

What are the key details on Agentic engineering is a different language?

Agentive engineering is a young term (barely a couple of years old), and every team reads something different into it. But this year the final showed clearly that agentic systems demand different priorities. Instead of code purity, the ability to adapt on the fly. Instead of full test coverage, readiness to work with incomplete information. Instead of perfect architecture, iteration and learning from mistakes.

What are the key details on A split in engineering?

The second place is not a judging error. It is a signal of a deep split in how we think about code. Classical engineering: types, tests, documentation, an architecture that keeps 65 tests in check and follows every best practice. Agentic engineering: a quick hypothesis, integration with the outside world, readiness to be wrong and to learn, the ability to think under uncertainty. This is not only about code. It is about two different worldviews on what good software is.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

Test Frontier LLMs & Autonomous Agents in One Workspace

Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…