Back to Articles

Fast Software and Its Costs


Quick Points

  • People still use the internet along with computers
  • The contact information you gather is for a real person, not just a piece of data
  • Quick money always has a cost hidden somewhere
  • Regaining trust is a long and difficult road

Outline

  1. People Still Use The Internet
  2. The Cost of Free Software
  3. Selling Those Lists
  4. The Danger of Vibe Coding Public Tools
  5. Why Does This Matter?

People Still Use The Internet

Even before the current AI era of LLMs, the internet was full of bots, crawlers, etc. But now with LLMs, it can almost feel like the people using the internet are an afterthought. So many people are trying to make a quick buck that they build or copy a website, make sure it works somewhat, and then launch it. Only if it has traction will they refine it and clean it up. But this can lead to some dire circumstances. How many of us have gotten the dreaded email from some company about “a data breach has occurred”, or ended up on an email or SMS list we never asked to be on, or have read stories of Vibe Coded websites where the email addresses of their waiting list or customers were in plain text for the world to see. While it’s true that lots of Internet traffic can be attributed to bots or crawlers and that the barrier to entry for building software is lower than ever, people and their data are still a major part of it. While data breaches or vulnerabilities will always exist due to humans being humans, I believe many companies could do well to remember that the data they collect or sell represents real living people and needs to be protected as such.

Failing Fast

The argument I have heard the most for this approach to building software/websites is that “it prevents wasting time/resources on things that won’t work in the end”. As a business owner, I understand the want to fail fast and protect your resources. Resources are limited, we can’t do everything 100% correct all the time, and we do need to make a living. But I believe there is a difference between doing something quick and doing something sloppy.

Quick is great. It forces you to define an MVP, think about the problem you are really trying to solve for your customers, and cut everything else out. Then you can focus on doing that MVP as properly and as fast as you can. No one likes sloppy: whether it is poorly defined requirements that contradict one another, file structures that aren’t maintainable, or poor handling of sensitive data. Sloppy drives everyone crazy who comes across it. So when you are using your favorite LLM, tread the line between quick and sloppy very carefully.

The Cost of Free Software

Turns out our grandparents and parents were right: there is no such thing as a free lunch. With all the free software and applications available, it’s easy to forget this fact. After all, all we need to do is make an account with our email and confirm we received the follow-up email. That’s not so bad, right? For many, it’s not. It is a cost well worth paying. Until they end up on an email or spam list full of phishing links. While there are plenty of valid reasons for a company to want you to have an account, the number of apps and websites that now require this, even for the simplest of transactions, is starting to boggle the mind. And all because the business model is to turn their users into an audience to market to. Again, marketing to your audience is not bad in and of itself. But if that is your business model, the data of your users and who you give it to is your responsibility.

Selling Those Lists

Just to reiterate, selling contacts and lists of contacts is not the issue. “I know a guy” and “let me introduce you to…” are phrases that have been around for a long time. They are just now done at a much larger scale and are much less personal. But with this abstraction from names, faces, and handshakes to First Name, Last Name, Email Address, it is easy for companies and their employees to forget a simple fact: these are people. Do these lists contain bots and fake information? Many do, especially lists that are not high quality. But they still include real people. If this business model is one you engage with, I challenge you to ask yourself, “Would I want to sell my own information to this person/company?” If the answer is no, why would you sell someone else’s?

The Danger of Vibe Coding Public Tools

I want to be clear, I am not against A.I. I use it as well and it is a phenomenal tool. But it has further abstracted away the idea that people use our websites and our applications. You don’t have to search far to find examples of Vibe coders or “A.I Visionaries”, who were so focused on building their application and making money that they forgot that they work with real people. Their customer list or other personal data is stored in the html for all to see. Some of these tools have even bitten their creators when they hardcode their API keys and ended up with $100s, $1000s, or even $10,000s of unintended charges. Or had their databases dropped because of a lack of permission management. If you have an idea for an app and want to collect people’s data, well done and good for you! But think of that and do your research as you go. As a matter of fact, use your favorite LLM to do so. Even a simple prompt of:

I am collecting user's email address and personal info on my website, Scan my code and look for places where this data is not protected and update it to follow current security best practices

This will go a long way in not only helping you build your app, but keeping your customers safe.

Why does this Matter?

There is another universal human fact that, once trust is broken, getting it back is extraordinarily difficult. If you have grand visions for a SAAS product or a website, for it to succeed, your users will have to trust you. And if your product is affected by a bad data breach or a vibe coded ‘Whoopsy’, you may immediately lose a large part of your user base. If you slow down just a little bit, take the time to do it as right as possible, and keep the safety of your user’s data in mind, keeping that trust will be much easier.