Vibe Coding on ROBOCO

Surprisingly, the Ralph Loop Is 9x Cheaper - The Economics of Context Caching

Sun, 26 Jul 2026 10:00:00 +0900

A single-session loop that just states a goal and runs until done finished the same task at one-ninth the cost of a plan-driven workflow. The secret isn’t the loop — it’s the cache.

Dohyun Jung - Principal Consultant, ROBOCO

The Ralph loop is a simple approach. You specify only the goal and the completion criteria, then let the agent iterate autonomously in a single session until it passes the completion check. No plan document, no task decomposition, no handoff notes between sessions. At first glance it looks unplanned and inefficient.

Cognitive Debt - Managing a New Kind of Debt in the Vibe Coding Era

Sat, 18 Jul 2026 10:00:00 +0900

“Shipping first time code is like going into debt. A little debt speeds development so long as it is paid back promptly.” — Ward Cunningham, OOPSLA ‘92¹

Dohyun Jung - Principal Consultant, ROBOCO

Work with vibe coding long enough and a strange moment arrives. This is unmistakably my project, yet when I open the code it feels like someone else’s house. The agent designed it, the agent implemented it, and I pressed the approve button. The work got done fast. But when something breaks, I find myself saying: “Why did this happen? The AI wrote it, so I’m not really sure.”

A Close Look at Ruflo: What It Means to Layer Multi-Agent Orchestration onto Claude Code

Mon, 04 May 2026 11:30:00 +0900

How to design Claude Code workflows that learn from a team’s insights, not just a single project.

Dohyun Jung - Principal Consultant, ROBOCO

Use Claude Code for a month and the limits become clear. Context disappears along with the session, and breaking work apart and putting it back together is ultimately a human job. The bigger problem surfaces the moment you have multiple projects and multiple developers. The testing strategy learned on project A, the refactoring pattern discovered by developer B, the architectural judgment validated in repository C — none of it flows well between them. ruvnet/ruflo (formerly Claude Flow) targets exactly this gap. As of May 4, 2026, it has 39.4k stars and 4.5k forks on GitHub, with v3.6.27 as the latest release.¹² Ruflo aims to be an orchestration layer on top of Claude Code that adds roughly 100 specialized agents, HNSW vector memory, a plugin system, and Zero-Trust federation. This post lays out what Ruflo actually solves and what it does not, particularly from the standpoint of how to organically reuse the insights that multiple projects and multiple developers acquire.

Serverless Autoresearch: A Record of Making an ML Experiment Pipeline Serverless Through Vibe Coding

Sun, 29 Mar 2026 16:00:00 +0900

The real value of vibe coding is not only in writing code fast. It is in designing a way of operating that lets you repeat expensive experiments more cheaply and more quickly.

Dohyun Jung - Principal Consultant, ROBOCO

Most people first associate vibe coding with web app prototyping or CRUD automation. But ROBOCO’s serverless-autoresearch repository widens that frame considerably. Instead of the “tie up a single H100 for several hours” approach that Andrej Karpathy’s autoresearch presupposes, this project builds a structure that runs parallel experiments in short, explosive bursts on AWS SageMaker Spot.¹

OpenClaw's Five-Layer Gatekeeping: Quality Control in the Age of AI Agents

Fri, 27 Mar 2026 00:00:00 +0900

Thousands of PRs pouring into a 310K-star project, plus parallel commits from eight AI agents. The secret to keeping quality from collapsing amid that chaos is five layers of automated gatekeeping.

Dohyun Jung - Principal Consultant, ROBOCO

In an earlier post I analyzed the productivity principles embedded in OpenClaw’s AGENTS.md. The heart of that piece was “writing the rules down.” But writing rules and enforcing them are entirely different problems.

The Line Between Repository Rules and Personal Taste in the Vibe Coding Era

Thu, 26 Mar 2026 17:26:21 +0900

The job of AGENTS.md is not to clone your developers. It is to pin down only the boundaries where accidents must be prevented.

Dohyun Jung - Principal Consultant, ROBOCO

Teams that adopt vibe coding soon run into a similar question. Once you start putting an AGENTS.md or CLAUDE.md in every repository to tell AI agents the rules, it becomes murky how far you should pin things down as shared rules and where you should leave things to individual taste. Draw that line wrong and two problems appear at once: the safety rules that absolutely must hold get loose, while choices that are really just preferences become needlessly rigid.

Organizational Design for Adopting Vibe Coding

Fri, 20 Mar 2026 21:23:40 +0900

Whether vibe coding succeeds or fails comes down to organizational design: whether the team can verify and ship AI-generated changes in a repeatable, safe way.

Dohyun Jung - Principal Consultant, ROBOCO

Many organizations understand vibe coding as little more than “a way for developers to produce code faster.” So the adoption conversation usually starts with tooling. But what actually decides success or failure in the field is organizational structure, not the tool or the model’s performance.

A Token Management Strategy for Vibe Coding

Thu, 19 Mar 2026 09:00:00 +0900

Running out of tokens is rarely a model performance problem. It is usually a problem with how you operate your context.

Dohyun Jung - Principal Consultant, ROBOCO

Use a vibe coding tool like Claude Code, Codex, or Gemini long enough and the same symptoms start showing up. Responses slow down, constraints you already agreed on get forgotten, and the model starts touching files that have nothing to do with the task. We usually describe this as “running out of tokens,” but what is actually happening is closer to context pollution, or context rot.

OpenClaw: Nine Best Practices from the World's Largest Vibe Coding Project

Sat, 14 Mar 2026 10:00:00 +0900

“I used to lead a team. I had a lot of software engineers under me. Even then, I had to accept that they would not write code exactly the way I wanted it written.” – Peter Steinberger¹

Dohyun Jung - Principal Consultant, ROBOCO

OpenClaw is the most-starred software repository on GitHub (310K+ stars) and the definitive case study in large-scale AI-assisted “vibe coding.”² Built by Austrian developer Peter Steinberger using three to eight parallel AI agent instances,³ the project demonstrates that a single developer can orchestrate a 300K+ LOC TypeScript monorepo, more than 20 messaging integrations, and native apps for three platforms — while barely reading most of the code.¹

Everything Claude Code vs Oh My ClaudeCode: A Comparison for Team and Enterprise Adoption

Tue, 27 Jan 2026 09:30:31 +0900

Everything Claude Code (ECC) hands developers a rich set of tools and guidelines. Its approach is to improve the output by making you follow optimal development habits and patterns. Oh My ClaudeCode (OMC), by contrast, automatically orchestrates multiple agents without any complicated setup. It focuses on getting results fast.

Dohyun Jung - Principal Consultant, ROBOCO

I covered each of these in two earlier posts, Everything Claude Code Distilled and Oh My ClaudeCode Distilled. In this post I compare the two tools currently generating the most buzz in the vibe coding community, to help you choose between them.

The Best Production-Grade, Cost-Effective Vibe Coding Tool from an Enterprise and Heavy-User Perspective (as of January 2026)

Sun, 25 Jan 2026 11:45:52 +0900

This piece was written as of January 2026. Pricing, usage limits, context windows, and plan structures change extremely fast, so I recommend reading it for the “principles and structure” rather than the specifics.

Dohyun Jung - Principal Consultant, ROBOCO

Vibe coding tools are entering the phase where “whatever you use, it more or less works.” So for enterprise users and heavy users, the question naturally shifts. It is no longer “which tool is the smartest,” but which tool you can use every day in production without hitting a bottleneck, at a cost that still makes sense.

Oh My Claude Code - The Plugin That Turns Claude Code Into a 'Team'

Wed, 21 Jan 2026 22:03:59 +0900

“Don’t learn Claude Code. Just use OMC.” – oh-my-claudecode README¹

Dohyun Jung - Principal Consultant, ROBOCO

oh-my-claudecode (OMC) is a plugin that layers “multi-agent orchestration” on top of Claude Code.¹ Its goal is to automatically activate the behaviors you need — planning, parallelization, persistent execution, research, design sensibility — using your natural-language request as the cue, so you never have to learn concepts like subagents, skills, and hooks one at a time.

Everything Claude Code: A Hackathon Winner's Recipe for an AI Dev Team

Tue, 20 Jan 2026 09:30:07 +0900

“If you want to bring AI onto the team, you have to design the process before the tools.” – Affaan Mustafa, winner of the Anthropic x Forum Ventures hackathon

Dohyun Jung - Principal Consultant, ROBOCO

Everything Claude Code is a collection of configurations that turns the Claude Code CLI into a virtual development team environment. A hackathon winner spent ten months refining these recipes while building an actual startup product, then gathered them all into one public repository; apply them and you can summon Claude as a “senior engineer + QA + architect.”¹ The repository is open for anyone to inspect on GitHub.

A Guide to Enterprise AI Adoption

Fri, 26 Dec 2025 09:16:27 +0900

The essence of AI adoption is not bolting on a new tool. It is changing the very way an organization creates customer value, faster and more accurately.

Dohyun Jung - Principal Consultant, ROBOCO

Meetings about adopting AI tend to start from the same question. “Our competitors are already using it, so when should we start?” Tool comparisons, license budgets, and training schedules follow soon after. That much is easy. The genuinely hard part comes next. It’s common for the way people work not to change fundamentally even after months, or for only a fraction of the organization to use it before it fizzles out. At some point AI stops being an “innovation project” and becomes “one more subscription fee.”

Claude Code Deep Dive: How Is Claude Code Actually Built?

Tue, 25 Nov 2025 10:05:54 +0900

The world of meticulous prompt engineering hidden behind the “it just handles it for you” experience

Dohyun Jung - Principal Consultant, ROBOCO

Developers trying Claude Code for the first time often wonder, “How does it understand the context this well?” Claude Code doesn’t behave like a simple coding assistant; it acts like a colleague who has been on the project for a long time. The secret lies in a sophisticated system architecture in which more than 40 prompt fragments are assembled dynamically.

The Tidy First Methodology: Interpreting and Applying Kent Beck's Augmented Coding

Sun, 27 Jul 2025 06:29:39 +0900

Dohyun Jung - Principal Consultant, ROBOCO

Lately I’ve been using Claude Code as my main working tool. The reason Claude Code performs so well is that it doesn’t execute a task immediately — it plans the work as a whole, systematically, then breaks it into small, clearly defined steps and carries them out one at a time.

Today I’d like to introduce one more useful methodology on top of that: the approach called “Augmented Coding,” proposed by Kent Beck — famous as the originator of TDD (Test Driven Development) — on his blog. Seeing that he, too, is obsessed with coining new terms, it looks like he’s in a hurry to grab a piece of the vibe coding wave. That said, since this approach is grounded in the content of a book Kent Beck wrote previously — Tidy First? by Kent Beck — I’ve decided, on my own authority, to call it the “Tidy First methodology.”

Vibe Coding or Engineering Maturity: Which Comes First?

Fri, 11 Jul 2025 14:46:42 +0900

Dohyun Jung - Principal Consultant, ROBOCO

A lot of developers think of vibe coding as nothing more than “a way to generate code automatically.” Give the tool an instruction, wait a moment, and beautiful code pops out. Reality is not that simple. In an environment with no automated tests, no solid design, and not even a functioning review process, vibe coding just churns out garbage code faster.

This is not a tooling problem. It is an engineering maturity problem. Software engineering maturity and vibe coding complement each other, and only in an environment with high engineering maturity can vibe coding show what it is really capable of.

Hiring Strategy for the Vibe Coding Era

Sun, 06 Jul 2025 19:04:33 +0900

Dohyun Jung - Principal Consultant, ROBOCO

The era of vibe coding is well and truly upon us. A world where AI writes the code and handles most of the day-to-day work has become reality. According to a blog post from Final Round AI, an AI-powered interview assistance platform — Halfway Through 2025, AI Has Already Replaced 94,000 Tech Workers — the headline says it all: roughly 94,000 people lost their jobs to AI in the first half of 2025 alone.

A Review Strategy for Vibe Coding

Sat, 07 Jun 2025 12:21:04 +0900

Dohyun Jung - Principal Consultant, ROBOCO

It is true that adopting vibe coding automates several areas of development and raises productivity. But review is something people have to do, because people are the ones accountable for the outcome — and so the bottleneck ends up moving to review (document review, plan review, code review, and so on). Today I want to talk about review.

TL;DR

Vibe coding speeds up generation, but if code, document, and plan reviews stay the same, the bottleneck simply shifts to the review stage.
To raise review quality, make approval accountability explicit, and use checklists and risk flags so that reviewers judge actively rather than passively.
The more automation grows, the more deliberately human review has to be designed — you need a structure that attaches meaning and accountability even to small actions.

Let’s start with code review.

Vibe Coding and the XY Problem

Fri, 06 Jun 2025 09:56:25 +0900

Dohyun Jung - Principal Consultant, ROBOCO

Have you ever heard of the XY Problem? Put simply, it’s the phenomenon of leaving the real problem (X) untouched while fixating on a solution (Y) you’ve arbitrarily assumed. Here’s an example. Your car won’t start, and without properly checking anything, you jump to the conclusion that it’s the battery and ask, “How do I hook up the jumper cables?” The real problem is that the car won’t start, and while you have no idea whether it’s the battery, whether you’re out of fuel, or whether it’s the electrical system, you’re floundering around clinging to Y (connecting jumper cables).

The Strengths and Weaknesses of Vibe Coding

Mon, 02 Jun 2025 10:40:26 +0900

Dohyun Jung - Principal Consultant, ROBOCO

As I noted in my earlier post on the productivity of vibe coding, that productivity hinges on two factors. The first is human productivity — the developer’s ability to solve problems and analyze them. The second is the productivity of the tools and processes that complement and amplify that ability. Productivity here does not mean simply writing code fast. It covers the whole range of competence involved in making a plan, analyzing a problem, and finding a solution.

AI, Become My Teammate!

Sun, 25 May 2025 07:30:47 +0900

Dohyun Jung - Principal Consultant, ROBOCO

On May 23, Anthropic released Claude 4 — just three months after Claude 3.7 went public. OpenAI’s release cycle is getting shorter too. And it isn’t only the models that are speeding up. So where does all this extraordinary productivity come from?

TL;DR

Vibe coding can be understood as a productivity amplifier that lets small teams deliver outsized results.
Treat AI purely as a threat and you end up dragged along by change; make it a teammate and it expands what developers and organizations can do.
For companies, vibe coding is shifting from an optional experiment to something closer to a survival strategy.

Cursor.ai, which recently stunned the industry, reached $100 million in annual recurring revenue (ARR) in just 21 months with a team of 20, and is now valued at $9 billion (roughly ₩13 trillion). Bolt.new and Lovable, each with a team of 15, posted $20 million and $10 million in ARR respectively in only two months. Midjourney, famous for its image generation AI, hit $200 million in ARR in two years with just 10 people. Mercor (30 people), ElevenLabs (50 people), and Aragon.ai (9 people) are putting up similar numbers.

The Vibe Coding Productivity Equation

Wed, 21 May 2025 08:56:24 +0900

Dohyun Jung - Principal Consultant, ROBOCO

I often get this question from developers: “This is how I develop — does that count as vibe coding?” It leaves me rather baffled. Why are Koreans in particular so fond of measuring themselves against a standard, almost like religious fundamentalists? What matters is not whether you are doing vibe coding correctly, but how much vibe coding actually improves your productivity.

There is another question I hear often: “Can you really do vibe coding without knowing any programming at all?” That is exactly what I want to talk about today.

A Vibe Coding Maturity Scale

Fri, 16 May 2025 09:06:12 +0900

Dohyun Jung - Principal Consultant, ROBOCO

These days there is a sharp increase in developers posting about their vibe coding experiences on developer social media and community sites, talking about how convenient and enjoyable coding alongside AI is. But when you actually look at what they describe, the level of experience packed under the single term “vibe coding” varies enormously. Some are amazed by simple autocomplete; others have gone all the way to designing and operating an entire service together with AI. From here on, I will lay out a maturity scale for vibe coding based on my own criteria.

Cline Is Getting Serious

Thu, 15 May 2025 08:08:46 +0900

Dohyun Jung - Principal Consultant, ROBOCO

TL;DR

Cline is carving out a bigger place in the vibe coding tool market on the strength of being open source, supporting BYOM, and holding up on large projects.
The headline of version 3.15 is that diff-based editing and AST-based analysis make work on big files and big codebases far more efficient.
Plan/Act modes and broad model support let developers tune cost, reliability, and control all at once.

Why Is Cline Back in the Spotlight?

The AI development tool market is a free-for-all, with new tools pouring out every month. Amid the fierce competition between well-known contenders such as GitHub Copilot, Cursor, Windsurf, Claude Code, and Amazon Q Developer, Cline has been steadily growing its presence. Cline actually drew attention by taking a distinctly different path from other vibe coding tools. Being open source, it is not a tool people pick simply because it is free. Cline holds up robustly even on complex, large-scale projects. It is also one of the few vibe coding tools that supports BYOM (Bring Your Own Model), which matters in environments where data security and sovereignty constrain what tools can be used. Leading with these strengths, Cline quickly joined the ranks of mainstream vibe coding tools.