Wesley D. Snow, Co-Founder and President of Ascendix Technologies. 30+ years in software development, a pioneer in AI for Real Estate.
We have a Teams channel where AI enthusiasts on our team gather to swap new tools and gripe about the ones that broke. I’m not the most technical person on that channel. So mostly, I’m there for the jokes.
Lately, almost all of them are about vibe coding—describing what you want to an AI and letting it write the software for you. It’s gotten mainstream enough that OpenAI released a special keyboard for vibe coding.
Half the humor is watching someone brag about shipping something in twenty minutes that any real developer would recognize on sight as a house of cards. I scrolled past dozens of these jokes without thinking much about them. Then I encountered one of those while scrolling through search results.
One Person Can Look Like An Entire Company
I came across an AI staging tool built for real estate: a vibrant, interactive interface, headlines written to sell, the kind of site you’ve seen a thousand times. It looked polished enough that I could picture pitching it to one of our clients. So I scrolled down to the contact form and found the “parent company.”
It was a second website hosting a whole lineup of AI tools for real estate, each with its own name, logo and set of glowing headlines. But when I looked closer, every one of those “products” traced back to a single LinkedIn profile with a clearly AI-generated photo.
AI now lets one person’s weekend effort pass for a funded, staffed software company. Every so often it’s how real estate brokers end up buying half-baked software and trusting their client data to nobody in particular.
I won’t share the links. Most of those sites are blocked now, gone entirely or their domains have been resold to Taiwanese casinos.
A Chatbot’s Top Pick Isn’t A Recommendation
I found that tool because ChatGPT suggested it as a top choice, and that worried me more than the scam itself.
AI recommendation tools are good at surfacing what’s marketed well and search-optimized. Security, scalability and long-term support rarely factor into the ranking. A slick landing page and a cheap subscription can outrank a serious, well-built product every time.
That’s why AI likely will never replace consultants and software developers. Someone needs to go down the rabbit hole a chatbot won’t, test the tool under real conditions, read the actual documentation instead of the marketing page, find out who’s actually behind it and how long they’ve been at it, and check what happens to your data if you connect it to the solution.
And while exploring the limitations and prevalence of vibe coding, curiosity got the better of me.
I Vibe-Coded My Own CRM. My Security Team Said No
After seeing how many of these one-person operations were passing themselves off as real companies, I figured it couldn’t be that hard. If a scammer with an AI avatar could fake a whole company, surely I could build something real for myself. So I tried it.
It was fairly easy. I vibe-coded my own CRM in an afternoon.
Then I asked our security team to look through it. Patiently and persistently, they explained to me, like I was five, why I shouldn’t use it and mustn’t connect it to anything real. The diagnosis: among other things, weak login protection, no encryption on client data and no way to tell afterward who had gotten in or what they’d changed. A good-looking wrap, but no better than those scams.
The Risk Is Skipping The Expert
Vibe coding makes custom development less time-consuming and expensive. When paired with human expertise, rigorous testing and continuous improvement, you can get a high-quality solution at a reasonable price within a realistic timeline.
Custom development still takes time and resources, but they are more manageable than those wasted just to find out six months in that the tool you bought can’t scale, won’t pass a security review or was never really supported by anyone in the first place.
A knowledgeable team can vibe-code responsibly, because they know exactly where the shortcuts are safe and where they aren’t.
Or they can vet the “hot new tool” a client found. Many red flags arise with a deeper web search and finding the people behind the solutions. Some issues surface during security audit and user testing. At the end of the day, a tool recommended by a consultant is more trustworthy than the ChatGPT top pick based on AI-generated website materials.
I still check that Teams channel, mostly for the memes. They’re funny because they’re true, and they’ll stay true for as long as buyers keep judging software by its wrap instead of what’s underneath. Meanwhile, my own CRM is still sitting there, disconnected from everything, looking great.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?