NEW YORK, April 14, 2025 – The term “technical founder” takes a lot of time to have a specific connotation in the world of ​startup – typically a codex that could ​build software from the ​ground. But in 2025, this archetype evolved. Thanks to the rise of the “vibe code”, a term popularized by the co-founder OpenAI Andrej Karpathy, artificial intelligence removes the traditional definitions of technical exploit. Vibration coding allows ​no ​encoder to build applications, websites and tools that simply describe what it wants ​in natural ​language, with the AI processing the code. As a ​result, investors, startup incubators and risk capitalists change ​the eyes of the founders ​with broader and less conventional skills.

According ​to Business Insider, the risk capitalists who swore allegiance to the idea of a technically qualified founder – the one who could ​hit JavaScript and purify server errors ​- now make room for a new type of innovator. These are people ​who may not know how to write ​a single line of code, but are competent to manipulate artificial intelligence tools to bring the software to life. What lacks syntactic fluidity consists of knowledge of the field, business ​and strategic use of major language models such as ​ChatGPT and ​Claude. The traditional encoder does not ​disappear, ​but they are no longer the only game in the city.

What is Vibe ​coding and why is it important?

Vibration coding is ​the software generation process by giving instructions in natural language to AI. If you want ​to build a task management application, an online game or a complex data panel, you simply describe it ​to an artificial intelligence assistant, who produces an operating ​code. If the output is not correct enough, ​they will be iterated – not plunge into the code, but refine ​its indications. Think of directing a chef with words instead of entering the kitchen yourself. As noted by ​PCWorld, many users have created ​everything from Pomodoro timers to full ​productivity in this way, without writing a single line ​of code.

This ​change in the way the software is ​built has important wave effects. The founders no longer need years ​of coding experience to create digital products. With models such as the OpenAI ​GPT-4 and the Antropics Claude 3.5 sonnet, the entrance barrier has decreased considerably. As the Times pointed out, even beginner users ​can create websites or functional games by inciting AI ​and copying the resulting ​code into publishers such as VSCode or ​TextEdit. This democratization of development enables more entrepreneurs to send software and attract investment.

How do investors respond to ​this change?

Investors adapt ​quickly. According to partner ​Oxxx Bob Thomas, the priority ​is no longer whether a founder can write code, but if they have a deep vision of the industry, they interrupt. “If you are an ex-accountant with 20 years in the field, ​you might be more likely to build the next ​fintech unicor than a brilliant but agnostic domain encoder,” said ​Business Insider. As AI tools continue to ​evolve, capitalists ​at ​risk place more emphasis on the fluidity of the domain and the ability to orchestrate the vision of the product ​through AI.

This does not ​mean that technical knowledge is obsolete. Priya Saiprasad of Touring Capital ​noted that the technical founders are still in ​demand, but expectations have changed. “It’s less about building each function yourself,” ​he explained, “and more about optimizing user engagement, retention and ROI.” Essentially, the role of the technical founder is being transformed from ​coder to strategist, someone who understands architecture, the adjustment of the product market and how to use AI to ​work faster and smarter.

What skills ​remove ​the traditional encoding experience?

Being “technical” in 2025 is not only meeting Python ​or ​JavaScript. This is system-wide thinking and architectural foresight. As BGV Yashwanth Hemaraj says, startups now need founders who can serve as software architects, imagining how components interact and scale. The real asset is the ability to purify, adapt and iterate – quickly. In many ways, this promotes small and agile equipment. Hemaraj says: “Instead of hiring ​30 to 40 engineers, you can now get a ​similar exit with ​a 10-person team when combined with good artificial ​intelligence tools

First-stage ​investors, ​particularly the financing of seed companies and startups chaired, are looking for founders who are comfortable with artificial intelligence tools and can use them to achieve an ​efficient product market. Zoe Qin from Dawn Capital added that although ​they do not choose the ​line of use of AI, the depth and sophistication of ​the product must speak for themselves. Founders must show that they can ​use AI to build products that solve real business problems.

What can you really ​build with Vibe ​Coding?

As it is, a lot. One user, writing for How-To Geek, detailed his ​experience using ChatGPT to build a productivity application that combines Pomodoro timers ​with time ​lock ​and distraction tracking. Starting with ​a notice ​describing your vision, you have a working ​HTML file in less than two minutes. Over an hour, they refined it through successive impulses, fixing design problems, cutting timers and even adding a logger ​of distraction. All this was done without manually changing a single line of code.

Another person, speaking with The Times, described how she built a fully functional website using Claude AI after ​years of WordPress ​fighting. With ​a single player and a screenshot of their old site, they had a new HTML ​/ CSS / ​JavaScript version in operation in a few hours. ​Inspired by the ease and joy of the process, ​they continued to ​create several tools, including RPG dice reels and puzzle ​games, many in ​languages they had never studied. It was not just about productivity, but it revived his love for programming.

Can you really jump into the ​code?

The short answer is: for many projects, yes. But there’s a cave. As users have repeatedly pointed ​out, vibe ​encoding is as good as ​its instructions. If you ​cannot clearly explain what you want the software to do, AI will fight to deliver. In addition, purification ​and refinement always ​require a level of technical ​intuition, even if you do ​not handle the ​code directly. A good coder vibration knows how to drive it effectively, identify when something is off, and guide AI to the right solution.

Workwaves like “Work With Apps”, which allows the ChatGPT desktop application to interfere ​directly with editors like ​TextEdit, help simplify the process. But even these tools have cranks. If an application is left too long in ​the background, the connection falls, requiring manual reconnection. In short, although you do not need to code, you need patience, clarity and will for this.

Is Vibe Coding the future of software development?

Vibration coding is not without limits. It is not ideal for very complex ​systems, and certainly has not replaced the need for professional developers. But ​as a gateway ​for ​nocoders, and as a productivity ​engine for experienced engineers, it’s a game ​change. It allows ​founders to try ideas faster, iterate cheaper, and move more commonly ​in the product’s life cycle.

Importantly, it also changes the balance ​of power. Innovation ​is no longer the only area that ​can speak ​between square brackets and semi-colons. Now, a founder’s ability ​to communicate, conceptualize and collaborate with AI can be as important – if not more important – as its coding ​cuts. ​It is ​a subtle but profound ​transformation, and as IA tools continue to improve, defining ​what makes a great founder ready to continue to evolve.

So, if you are an experienced developer or a complete beginner, ​vibe encoding offers a look at a more inclusive and creative future. He won’t do anything for you, but I ​could help you build what ​you always wanted to create.