We Just Built More in a Weekend Than We Could Have in Months. Here's How

In a single weekend, a third-party partner program that would have taken weeks to plan came together in 36 hours. An assessment tool that once cost $50,000 and took five months to build was prototyped in seven minutes. A year's worth of YouTube data transformed into polished, visualized dashboards with almost no human interaction.

None of this required a team of developers. None of it required permission. And none of it would have been possible 12 months ago.

To illustrate what's now within reach for the every-day professional using AI tools Lovable and Claude Code, I spoke with SmarterX and Marketing AI Institute founder and CEO Paul Roetzer on Episode 190 of The Artificial Intelligence Show.

AI Builds a Partner Program

Roetzer’s challenge: A new employee was starting Monday to focus on a channel partner program for SmarterX’s AI Academy. A longtime team member had been exploring the concept for months but he wasn’t clear on the program's structure, pricing, or go-to-market approach.

So he found himself waiting for takeout on a Saturday night and thought: “I should solve something on my list this weekend because Sunday's crazy." 

"So I go into my Co-CEO GPT and I type: We're going to launch a channel partner program for AI Academy," Roetzer says. "Can you help me think through the design of the program? I want to keep it simple. To start, let's take it one question, one step at a time to plan the program."

What followed was a strategic planning session that spanned 25 minutes at the restaurant and a few hours Sunday morning. Roetzer chatted with it about pricing and commission models and it produced more than 50 steps with insights and decisions.

The key was Roetzer's experience and expertise: 16 years in the HubSpot partner ecosystem, having been the company's first agency partner back in 2007. He watched that program grow to more than 4,000 partners accounting for more than 40% of HubSpot's revenue.

"Anyone could give that prompt," he says. "Very few people had 16 years in the HubSpot ecosystem going through all the pains they went through, sitting in many of the early meetings as they were constructing and architecting that program."

By Sunday evening, he had a strategy brief, a draft press release announcing the program, and a complete chat thread documenting every decision. He plans to share these with his team so they can see the chain of thought and debate any choices before moving forward.

Seven Minutes to a New Tool

The partner program wasn't the only recent breakthrough for Roetzer using AI. On Thursday evening, while waiting for his family to get ready for his wife's birthday dinner, he had about seven minutes of downtime.

He decided to try Lovable, the AI-powered full-stack app builder that lets you describe what you want in plain English and get a working, deployable app or website in minutes.

He told it he wanted it to build an assessment tool and briefly explained the concept. Lovable walked him through several decisions: Who's the audience? How are you going to structure it? What access do they need? Do you want to collect emails?” 

In those seven minutes, he built a functioning assessment tool with 26 questions and the ability to generate PDF reports with visualizations.

A decade ago, Roetzer built a similar tool at his agency that took approximately five months and at least $50,000 in development costs. Building this way on Lovable wouldn't have been possible 12 months ago, he said.

Claude Code for Non-Developers

Meanwhile, I'd personally been thinking about Claude Code for weeks, unsure how it applied to someone with no development background. The breakthrough came when I stopped thinking of it as a coding tool.

Instead, I realized Claude Code is really just a large language model with access to external tools, to borrow from a helpful explanation I found online. You prompt it like a chatbot, but instead of just responding, it can go do things such as call tools, use file systems, browse the internet, and execute code.

For my test, I exported a year's worth of our YouTube data to a folder, directed Claude Code at it, and asked it to produce a comprehensive analysis brief and a strategy to grow our audience 10x. I also told it to produce any other documents it thought would be helpful.

Then I let it do its thing.

For about 15 minutes, Claude Code worked in the background analyzing data, writing code, and browsing the internet for context. It occasionally asked for permissions, which I granted. When it finished, it had produced five fully designed HTML dashboards displaying visualized YouTube analytics.

I got the analysis and strategy I asked for, plus a quick reference dashboard, monthly trend data, and video ranking charts I hadn't explicitly requested but that it determined would be useful.

The analysis was good, even better than anything I could possibly do in the same amount of time. And Claude Code’s level of autonomy and decision-making was impressive.

Software 3.0 Is Here

Roetzer saw a tweet that captured what we're experiencing. Yuchen Jin, co-founder and CTO of Hyperbolic Labs (formerly of Octo AI, acquired by Nvidia), wrote last week:

"We're entering Software 3.0, a world where everyone can build personalized software and ship new features in minutes, and the cost is zero. Software used to be built by companies with lots of engineers. Now it's built by individuals."

That's the shift. The obstacles that once blocked ideas from becoming reality are gone.

"Everything becomes solvable," says Roetzer. 

Now teams just need to spend time brainstorming about problems to solve, products to build, and ideas to bring to life, he says.

Think about all of the data sitting unused in most organizations (things such as media buy reports, customer analytics, performance metrics, etc.) that never gets analyzed or used for insights. Now imagine using a coding agent to transform that data and make it usable.

"We're just entering this world where anything becomes possible," says Roetzer. "Whatever you can imagine, you can do or will be able to do in the near future."

How Do You Compensate for Remarkable Innovation

This capability explosion creates an uncomfortable question for organizations: How do you compensate people who can have 10x or 100x impact?

"It becomes hard to even measure,” Roetzer says.

What happens when someone's idea triggers a new product line, eliminates expensive software, or creates real-time actionable data that generates millions in value? Traditional compensation structures weren't built for this.

It's a topic that came up in SmarterX’s AI Talent Council meetings and will be addressed in their upcoming talent report. The answer isn't clear yet, but the question is becoming urgent.

The amazing thing is that even if AI development stopped today, the technology exists now to transform the workplace. The only question is who will figure it out.

“For people who choose to use it for good in their careers, you have an amazing runway ahead of you,” Roetzer says.

He added: "I feel like I'm living through this privileged period of human history where we get to reimagine everything. And I just want to inspire other people to reimagine stuff."