An update on Popform, vibe coding, and the future of building.

We fired everyone.

Since starting Popform, we’ve had a rotating cast of 5 software engineers helping out. Some in a full-time capacity for as long as 9 months and some part-time helping build key pieces of the app in a few short sprints.

Today, we have none. It’s just myself and my co-founder… and AI. Now, we’re adding features, smashing bugs, and deploying faster than ever before.

Quickly: We’re personally onboarding users into Popform’s full beta this week and next. Everything is free while in beta, and unlimited e-signatures will likely always be free.

If you need a Docusign killer for contracts, proposals, and agreements go to www.popform.ai.

What changed?

Up until February, I was mainly responsible for product vision, UX design, product management, and customer acquisition.

I had never looked at our code, I didn’t know how to commit or submit a pull request, and I didn’t even have the right setup to run Popform on my machine locally — fuck, I barely knew what any of that even meant.

Then, while catching up with a friend, a CTO at another startup, a breakthrough happened.

He shared that they had let go of all their contractors and that his non-technical founder had picked up vibe coding. They are seed stage with a healthy round, but their VCs were grateful nonetheless that they had eliminated burn, extended runway, and unlocked more time to find product market fit as their idea evolved.

But I was hesitant to pick up vibe coding.

I’m good with HTML & CSS, I picked up basic Javascript long ago, but my knowledge of modern frameworks and React is very limited. I thought my baseline knowledge needed to be higher before I could even start vibe coding, so instead I was enduring the long grind of taking a React course in an attempt to become useful on the technical side of things.

He told me to stop learning. Download Cursor, get set up, and start building.

The next morning I went over to my co-founder’s house and had him set up my machine’s local environment and teach me the basics of Git and GitHub.

That night, I had contributed my first lines of code to Popform and added a basic feature or two.

By the end of the week I had made more commits than our contractor.

By the end of the month, it was taking me longer to document new features than it was to just fire up Cursor and build it myself.

Shout out to my friend Sunny who gave me the courage to dive in.

After 2 decades of failed attempts at learning to code functional applications, AI has come along and made it possible. It has blurred the lines between a non-technical and technical founder. All the developers are super scared right now.

This is not hyperbole, it has already changed my life.

One quick note - I can’t vibe code Popform to life by myself. My cofounder Viral is a seasoned engineer who builds and maintains the infrastructure needed to create a full production worthy app. He reviews my code, often has to fix it, and makes sure everything is kosher before we deploy. However, having a mentor and a gatekeeper of quality all in one is the best way to learn while continuing to push everything forward - it’s the best apprenticeship possible.

Zooming out

In hindsight, it’s not surprising at all that we will create software by simply asking for it.

The evolution of programming languages can be easily summed up in one word - simplification. It’s a history of new coding languages, each adding layers of abstraction — each designed to help humans communicate more easily with machines.

Early programmers wrote in machine code or assembly languages that was closely tied to the hardware itself. Over time, languages like C, Java, Python, and JavaScript were developed to be more human-friendly.

Today, the machine-level output of modern code would be unrecognizable to even the most senior developer. Just as a high-level language like JavaScript might appear confusing to someone with no programming experience at all.

If you plot this trend outward — each generation of languages moving further from the machine and closer to human logic — it starts to feel inevitable that we’d eventually land on the most universal language for expressing ideas and building systems: English.

Maybe we got here faster than we thought, but ultimately natural language has become the next layer of abstraction. With AI models now able to interpret and execute instructions written in plain English, we’re approaching a point where the interface for programming isn’t code but conversation.

What’s Next?

Developers are scared. If you know one, give them a hug.

For the past 3-4 decades, only “tech” companies have had the concentration of developers needed to produce enterprise-level software. Even industries like finance and healthcare never attracted the top echelon of software talent.

And everyone else? The IT department.

What will change?

Today, one developer can have the same output as five. One product person can have the output of a developer and a UX person combined.

• Big companies will be built with less people.

• More billion dollar companies, less multi-billion dollar ones

• Developers’ salaries will humbly decrease

• Roles will consolidate - most likely into one product team of generalist builders

• Apps & SaaS could die completely - giving way to agents unconstrained by the boxes drawn by their UX overlords

But, most importantly, development will be democratized to every SMB.

Your laundromat, general contractor, dog groomer - they could all have apps & agents tailored to their specific needs. All built in a weekend or two.

What does this mean for lowercase b2b?

Spend a few minutes on LinkedIn right now and you’ll get some AI whiplash with an entire spectrum of AI bullish and bearish takes from people you know.

Some folks think they won’t need website designers and developers because they can just ask AI for one. Others think they won’t even need a website because everything will be agentic and websites will be done.

Some folks think SEO is dead and all search is moving to AI. Others think they need to be on the frontier of SEO for AI (AIO/GEO/AI SEO - all TBD terms).

No matter what, I know that if you retreat in fear of the future, you’ll get left behind. You will fail.

For lowercase b2b, my agency, I believe anything is possible. I believe in all of the above.

I choose to play with these tools and surround myself with others who are in play mode too. I don’t want to just see what the future holds - I want to shape it. I despise nostalgia and believe that it’s more fun, more productive, and more helpful if you just play!