Using AI the Smart way with Suyash

Joining us in the Bay Area this Summer, Suyash brought his expertise in AI and ML to the students in our 2025 Bay Area Entrepreneurship Program.

We sat down with him to explore the smartest ways to utilize AI and the best approaches to leveraging the tools.

AI as a coding assistant

There are more AI tools than ever to help you code. Whether that is a copilot that offers suggestions and fixes for code, all the way to a vibe-coding solution where you build through text prompts, there are immense opportunities available to youth entrepreneurs.

When using AI as a code helper, Suyash offers an excellent illustration to help you use the tool in the best way possible.

View yourself as a developer manager and the AI tool as a low-level intern.

Complexity is your responsibility, not AI’s

AI tools can be great at simple tasks, but as the complexity grows, so does the need for you to understand the complexity yourself.

If I ask AI to help produce a simple button on a page, I can probably use that prompt to easily make something workable quickly. However, if I aim to build an online marketplace with a front-end and back-end for my clients, as well as a comprehensive back-end for myself to manage, there will be problems.

First, it is essential to understand that there are limitations to what AI can do. Remember, think of AI as a low-level intern. I am not going to ask an intern to architect an entire system. And, in our example above, AI might even be the wrong tool to use.

However, if you are reasonably sure AI is the right tool for my project, then I need to begin breaking down the project into bite-sized pieces that my *intern* (AI) can help me accomplish.

You need to be able to clearly define what you want the system to do and how it will work. If you have this info, you can use AI to accelerate the process. However, if you are unclear, this will cause problems quickly.

Active management matters

You can’t just prompt the system and leave it be. Use print statements throughout the process to begin to understand what process AI is using. Just like a good manager, you will need skills in debugging to know when mistakes are made and how to solve new challenges created by using the system.

You can learn to be a manager somewhat on the fly because you will have the system to help you. This is an excellent opportunity to learn and grow, but it can be frustrating as the system may not perform exactly as you hope it will. The benefit of working with and learning from the system is that it allows you not to be an expert in everything, but instead to focus on how to manage the system effectively.

This requires a different way of thinking.

How to get good at AI

Build.

No, seriously, get out there and start to build something small. Unlike coding languages that have defined rules and terms, which you can study, when utilizing AI tools, it is better to learn on the job.

Additionally, with systems constantly changing, what you find true today may not remain so in a month. Start by building small things and figure out what works and what doesn’t. This will teach you to be a “manager” while learning the limitations of the system.

Whether you are using a code helper or vibe-coding tools, figuring out the system through practice will help you be able to create more complex apps and products in the future because you will be able to determine what it is capable of, and probably more importantly, what the system can’t do.

Get feedback and go again

At LaunchX, our philosophy has always been bent toward velocity. Get your product out in the world, go great, and get feedback. Coding in this way is no different.

Create your small app and publish it to the world. Ask people to give you feedback. Post it forums where others are interested in vibe-coding, then start chatting to figure out if there are better ways to approach the problem.

Suyash’s advice? Build as you learn, learn as you build.