How to use coding agents to stand out in AI era

174.9K views
•
October 16, 2026
by
Nate Herk | AI Automation
YouTube video player
How to use coding agents to stand out in AI era

TL;DR

Telling agents what you care about improves output and helps align results with you. The environment you give the agents matters, and rising the floor of performance comes from applying traditional software engineering ideas such as testing and quality controls. Be skeptical of AI suggestions and rely on primary sources when learning.

Transcript

How do you stand out in a world where everyone can build things so quick? >> I think the people who are really successful are the people who are the best communicators, right? So, in my opinion, the best way to get the best work out of agents is to tell them what you care about. >> So, the skill there is being able to just get it all out of your he... Read More

Key Insights

  • Agents produce better results when you clearly state what you care about, reducing misalignment and guiding the agent toward desired outcomes.
  • The human agent gap persists because the model does not know your exact preferences, so ongoing clear guidance is essential for quality outputs.
  • A well cared-for environment, with standardized processes and readable code, raises the performance floor and prevents garbage in garbage out outcomes.
  • Spec-driven development can temporarily degrade results as the system learns new constraints and expectations, making iteration and validation crucial.
  • Studying traditional engineering concepts like testing, checks, and continuous improvement remains relevant when integrating AI agents into workflows.
  • Trust in AI should be cautious and bounded; always seek high-quality primary sources to verify teach outputs and avoid hallucinations.
  • Non-technical people can apply software engineering fundamentals by focusing on goals, feedback loops, and process improvements rather than raw coding skill alone.
  • Adoption success comes from practice, experimentation, and a disciplined approach to harnessing agent capabilities for collaboration and automation.

Explore YouTube Video Summarizer or Get YouTube Transcript Extractor

Questions & Answers

Q: A real search query question about the video (How to, What is, Why does, When should). How to get the most out of coding agents by telling them what you care about

Telling agents what you care about is the core practice that produces better results because it closes the gap between human intent and machine action. When you articulate your goals and priorities clearly, agents can align their outputs with those targets, leading to more useful, reliable, and faster outcomes. This approach also makes feedback loops more efficient, as you can correct direction precisely where it matters most.

Q: What is the main reason a strong development environment matters for AI agents

A strong development environment matters because it raises the floor of performance for all coders, including those using agents. If the codebase and processes are sloppy or poorly organized, even a capable agent will generate suboptimal results. A clean, well-documented environment provides context and safeguards that help agents produce better code and fewer mistakes.

Q: Why did spec-driven development fail for months in this context

Spec-driven development initially led to worse outcomes because the specifications did not capture the evolving realities of using agents. As teams tried to rigidly codify requirements, the agents produced outputs that did not meet true needs, causing a deterioration in results until the process adapted to iterative testing and feedback.

Q: How does the speaker suggest approaching trust and learning from AI teach sessions

Trust in AI teach sessions should be cautious and evidence-based. The speaker recommends fetching high-quality primary sources, providing quotes and attributions, and using teach sessions to access information so you can then read the original material yourself. This reduces overreliance on hallucinations and builds a foundation of verified knowledge.

Q: What opportunities exist for non-technical people in this AI era

Non-technical people can apply software engineering fundamentals by focusing on goals, process design, and feedback systems rather than writing code. By understanding how to frame problems, design checks, and organize collaboration, they can guide agents effectively, contribute to automation strategies, and create value without deep technical skills.

Q: What role does skepticism play in working with AI agents

Skepticism plays a crucial role because AI agents can still hallucinate or misinterpret. By maintaining a critical mindset, verifying outputs, and seeking primary sources, you ensure that what you adopt is accurate and trustworthy. Skepticism keeps you from overtrusting the agent and helps maintain quality control.

Q: How can one stand out when everyone can build quickly with AI

Standing out comes from applying fundamentals, such as rigorous thinking, process discipline, and clear problem framing. By designing effective environments for agents, emphasizing quality checks, and communicating goals precisely, you create value that goes beyond rapid but shallow outputs and demonstrates thoughtful, reliable work.

Q: What is the speaker’s view on learning to code versus using agents

The speaker notes that learning to code is valuable but that agents can take over much of the manual coding work once they mature. The shift is about leveraging coding fundamentals to guide and supervise AI, enabling humans to focus on higher level design, mentorship, and process improvements rather than writing every line of code.

Summary & Key Takeaways

  • In this interview, the value of communication with AI agents is emphasized, highlighting how expressing care and goals leads to better results and faster delivery. It also discusses the importance of a thoughtful development environment and applying age-old engineering practices to AI-enhanced workflows.

  • The discussion covers the shift from hand coding to agent-based coding, the skepticism toward AI, and how to leverage teach moments to access primary sources for deeper understanding.

  • The speaker outlines opportunities for non-technical people to apply software engineering fundamentals, stressing that standing out comes from process, environment, and a clear value proposition when building with AI agents.


Read in Other Languages (beta)

Share This Summary 📚

Explore More Summaries from Nate Herk | AI Automation 📚