Podcast Summary

Overview

Episode: AI Assisted Coding: From Designer to Solo Developer – Building Production Apps with AI With Elina Patjas
Podcast: Scrum Master Toolbox Podcast: Agile Storytelling from the Trenches
Host: Vasco Duarte
Guest: Elina Patjas (Founder of Lexi, AI-powered study tool)
Date: November 24, 2025

In this bonus episode, Vasco Duarte sits down with Elina Patjas, a designer-turned-solo-developer, to explore how AI-assisted coding is transforming the software development landscape. Elina shares her journey from educational technology design to building production-ready apps like Lexi with the aid of AI tools. The conversation delves into paradigm shifts in software development, the empowerment of individuals, practical workflows, potential pitfalls, and predictions for a future where software is increasingly built (and possibly ephemeral) on demand.

Key Discussion Points & Insights

Elina’s Background & Motivation

• From Designer to AI-assisted Developer

  • Elina describes 15+ years in Finnish edtech as a designer and product manager.
  • Motivation to help her own children practice reading led her to experiment with AI coding tools, which rapidly expanded into developing Lexi, an AI-powered study app for kids.
  • [02:11] "If I can do this, so what else can I do?" – Elina Patjas

• Addressing Real User Needs

  • Finnish system is excellent, but not all kids get help at home; Lexi aims to bridge this gap.
  • Struggled in traditional teams where only "10% of your original vision" survived the process.
  • Building solo with AI allowed Elina to realize her complete vision at speed.

The Experience of AI-Assisted Solo Development

• From Vision to Execution, Without Compromise

  • AI tools shift the paradigm: less "writing code" and more "building products."

  • Designers like Elina, who "hate making compromises," are empowered to create exactly what they envision.

  • Rapid feedback cycles; features can ship within a day of a user request.

    "Now I am able to kind of execute the thing that I want, like 100% the way I want it, and I can do it really fast and I get the feedback immediately."
    —Elina Patjas [08:41]

• Example:

  • A student requested a feature for editing study sets. By next morning, Elina implemented and shared the new feature – something unthinkable in larger, conventional teams.

    "That makes me really happy because I'm actually able to do what I have always wanted to do as a designer."
    —Elina Patjas [09:47]

Approaches to AI-Assisted Coding

• Vibe Coding vs. Structured AI Development

  • "Vibe coding" is trendy but can be misunderstood as sloppy.

  • Elina’s method is structured:

    1. Designs full feature workflow and plan in natural language.
    2. Writes detailed specs, then collaborates with Claude (AI assistant) for actual coding.
    3. Reads and reviews every line of code the AI produces—but doesn't write code manually anymore.

    "I start doing the planning together with Claude... Claude is also like my planning partner and then it is my executional partner."
    —Elina Patjas [18:28]

• Workflow Details

  • Keeps AI’s context file (Claude.md) rigorously updated.
  • Breaks implementation plan into sprints/tasks, explicit prompting, frequent checkpoints.
  • Treats Claude as a "highly skilled companion"—communication and precise prompting are key.

Evolution of Tools

• Cursor to Claude
  • Started with Cursor for learning coding, transitioned to Claude for higher-level collaboration.
  • Switch driven both by evolving approach ("I actually don’t want to write the code myself") and tool change (pricing, capabilities).
  • Uses AI both in IDE (Zed) and command line; can run multiple AI processes in parallel.

    "At some point during this year, I realized that I actually don’t need to do this because it will do it so much faster than me anyway."
    —Elina Patjas [19:28]

Important Segments & Timestamps

The Paradigm Shift: Designer to Developer

• [02:11–06:06] — Elina’s journey and initial experiments with AI tools.
• [06:06–10:45] — Frustrations with traditional teams; empowerment as a solo developer.

Workflow & Best Practices

• [12:00–17:34] — Difference between “vibe coding” and Elina’s structured approach.
• [15:00–19:01] — Step-by-step workflow: planning, writing specs, detailed prompting with AI.

Lessons Learned & Challenges

• [22:10–24:46] — Benefits: uncompromised vision, speed, rapid feedback.
• [23:18–24:46] — Pitfalls: if you “let any AI tool run rampant,” things break; need constant engagement and control.

Communication: With AI vs. With Human Teams

• [25:14–28:51] — Faster, less frustrating feedback and iteration with AI; communication skills still vital but easier via writing.

  "The iteration loop is so fast because when you kind of fail to explain something, you get the feedback so fast that you can actually say, no, no, no, this is actually not what I wanted."
  —Elina Patjas [25:14]

The Future of Software Development

• [28:58–35:30] — Prediction: towards intent-based, ephemeral apps; one person able to accomplish what once took a team.

  "I think that kind of the whole paradigm is changing... we are going to the direction where people are, like, the software gets built in an instance... you don't even need apps as we know them right now."
  —Elina Patjas [29:44]

• [32:51–35:30] — Reflections on the potential and scariness of this change; designers and technologists must adapt continually.

Notable Quotes

• On breaking free from constraints:

  "There are people who hate making compromises... now I am able to execute the thing that I want, 100% the way I want it."
  —Elina Patjas [08:41]

• On fast feedback:

  "You can clarify the thing that you needed like really fast. So the feedback loop is so fast that it makes it so much easier."
  —Elina Patjas [25:14]

• On the role of AI as a collaborator:

  "I treat the AI as this highly skilled companion that you need to be able to communicate yourself super well and be on top of everything."
  —Elina Patjas [23:31]

• On the evolving future:

  "I can see that basically good AI tools can't do the, like, the actual design, like the user interface design or UX design pretty well already... At the same time it's scary, but at the same time I'm really excited."
  —Elina Patjas [35:30]

Practical Advice & Resources

• How to Start AI-Assisted Coding?

  • Don’t rely on tutorials or books that date quickly—just start building something you need and learn through direct interaction with the tools.

    "The best resource is just your own frustration with some existing tools... Just open whatever tool that you're using... and start interacting and discussing..."
    —Elina Patjas [36:08]

• Tools Discussed:

  • AI Assistants: Cursor, Claude
  • IDE: Zed

• Connect with Elina / Try Lexi:

  • Product: [Lexi app](show notes)
  • Social: LinkedIn, Twitter (but her focus is where her users are: TikTok, WhatsApp, YouTube)

Key Takeaways

• AI-assisted coding enables designers and product thinkers to build and iterate production software, often without manual coding.
• Success depends on structured planning, strong communication (in prompts), and treating AI as a skilled but literal collaborator.
• Short feedback loops and direct execution of vision are major benefits, but care must be taken to avoid "runaway AI" coding without oversight.
• Software’s future may be “on demand,” ephemeral, and intent-driven, challenging traditional ideas of teams and product maintenance.
• Start experimenting with current tools—learning by building your own solutions is the fastest way forward.

End of Summary