Skip to content

Methodology · Resources

Vibe Coding Manifesto

Methodology3 steps

Basic principles of working with AI: the developer who writes code becomes an architect of AI systems.

Methodology

What it is

Generative artificial intelligence is changing programming from the ground up. In the past, a developer's value was measured by remembering syntax and assembling logic by hand. Today what matters is the ability to steer the outputs from AI, choose among them and join them together. Vibe Coding is a complete framework for this new working culture. It moves from technical purism to practical "product thinking". While traditional methods such as Agile or Scrum manage processes in teams, Vibe Coding focuses on strengthening the abilities of an individual working with AI.

3 steps

Steps

  1. Step 1 · Imperative 01

    Stop writing syntax. Start defining logic.

    Syntax (the rules for writing code) is handled by the language model (LLM). The human's job is to determine the app's main logic and its architecture.

  2. Step 2 · Imperative 02

    Write code only when AI fails.

    Writing code by hand serves only as an emergency brake. When a developer writes code by hand, it means the process or the tool has failed.

  3. Step 3 · Imperative 03

    Speed is metric #1.

    Tradition stresses clean code, testability and resilience. The manifesto puts speed first: "Perfect is the enemy of shipped".

Principles

Principles

  • Flow over Friction: Keeping a state of full concentration ("flow") matters more than fighting with tools. Keep the pace, do not argue with technology.
  • Iteration over Perfection: Perfection makes no sense when you can have the result generated again at any time (reroll). Quickly trying more versions is better than a perfect first attempt.
  • Code is Disposable: Code is a consumable. There is no emotional attachment to the lines written. If it does not work, generate it again.

The equation of the new identity

Human = AI Manager. AI = Programmer. This dissolves the traditional division of roles. The developer is no longer a craftsman assembling code line by line but the manager of a team of AI agents. The human supplies context and taste, the AI supplies syntax and brute force.

Vibe Factor: f(V) = (e × r × c) / (n × AI(k))

The success of this approach is described by the concept of "The Vibe Factor™". The equation models how well programming with the help of AI works: f(V) = (e × r × c) / (n × AI(k)). The numerator holds what helps (higher is better): e (Experience), experience measured by the number of failures solved; r (Reliability), the reliability of testing and monitoring operations (observability); c (Context), depth of understanding of the field. The denominator holds what harms (lower is better): n (Attempts), the number of attempts needed while adjusting the prompt; AI(k), the illusion of time saved. The equation implies that Vibe Coding is not for beginners without knowledge of the field (c) and experience (e). If the values in the numerator are low, f(V) drops below 1 and AI creates more technical debt (a mess that will be expensive to fix) than benefit.

Historical anchor: William Hickson

William Edward Hickson (1803 to 1870) was a British educational writer who popularised the verse: "If at first you don't succeed, Try, try, try again." With generative AI, generating a new variant of code (reroll) takes seconds and costs nothing, so Hickson's "Try again" gains a new, technical meaning. Do not fix mistakes (don't debug), try again (reroll/regenerate). His name on the manifesto is therefore no accident: it links modern technology with the old wisdom of perseverance.

Critical analysis and risks: The security paradox

Vibe Coding promises speed, but the data show serious security shortcomings. Coding with agents solves the task functionally in up to 61% of cases, but securely in only 11%. AI often creates code with hardcoded API keys, without limiting the number of requests (rate limiting) and prone to SQL injection (an attack through input fields).

Critical analysis and risks: Erosion of skills

Critics warn of "brittle Frankensteins": apps that look functional but inside are chaotic and unmaintainable. When developers stop "writing syntax", they risk losing the technical knowledge needed for complex problems that AI cannot handle.

Critical analysis and risks: When Vibe Coding works

Success does not depend on blindly following the "vibe" but on experienced professionals (AI Managers) being able to steer, check and put together the outputs of their AI collaborators. It requires high values of Experience (e) and Context (c) in the Vibe Factor equation.

Outcome

Outcome

A shift from a craftsman who assembles code line by line to a manager of AI agents. A marked speed-up, but success depends on an experienced professional who steers and checks the outputs.

Start here

15 minutes is enough for you to know whether it makes sense.

You tell us what is bothering you. We tell you whether we can help, roughly what it costs and where I would start in your place. No presentation.

Address
Legerova 1820/39, Praha 2

Your plan:

AI Setup Day, from 1,396 EUR

The calendar is run by Cal.com (a third party) and loads only after you click.