Everywhere online, I see two main types of people: those who rapidly build with AI tools and those who overthink and take too long to perfect their work.
People are creating SaaS products using AI tools in as little as a day, a week, or a month. I often see posts on Reddit like “spent a week building and now making $80k/month” or “got my first clients after just a day of building.”
Soon, we’ll probably see posts like “built this app while on the toilet and now I can retire my family.” While many of these claims are likely exaggerated, some are genuine, and it’s tempting to jump on this bandwagon.
But where’s the quality? Quick builds are fine for testing ideas, but how can users rely on them long-term? And how do you improve something coded with AI-generated spaghetti code?
It seems some people either don’t know how to code and find learning too time-consuming, or they love coding so much that they can’t stop refining.
Where are the people who spend a few months building an app, then a few more improving it based on feedback, and finally focus on marketing a solid product? These are the ones who know how to code and value their time.
I’ve been working on alfred-assistant.com for the past few months, and while I think it’s a reasonable approach, these success stories make me question if I’m wasting too much time.
I encounter many people who build projects without truly understanding their purpose or target audience. This issue isn’t limited to those using AI tools many software developers face this challenge too.
The key to success is understanding who your target audience is and what problems you’re aiming to solve for them.
Convincing someone to spend their hard-earned money is much more difficult than most people realize.
I agree with quickly building your MVP and then refining it based on feedback. AI’s impact on coding was expected, and it seems beneficial—speeding up development and reducing costs, though it may lead to quick failures when things don’t work out.
If someone is claiming $80k MRR right away, use your judgment. It’s either completely unrealistic, or you should replicate their project and improve on it to compete.
I usually take about two weeks to develop an MVP and complete the whole project within a month. Currently, I’m working on a cold outreach tool that generates revenue for anyone selling services, whether digital or physical. It lets you directly reach your target audience and DM hundreds of people daily on Instagram, which is excellent for generating leads and filling your calendar quickly.
The development of easychatdesk.com took me between 12 and 16 months, and I just put it in beta a week and a half ago. The path was not without its challenges. Since I also own a web development agency, which requires a lot of time, it wasn’t a constant project. If you work on something, though, make it sophisticated, start with a small number of features, then over the next few months, quickly add more. Something constructed in seven days won’t be a reliable product.
You build a product quickly because you need to see if people will visit your site, sign up, or buy your product. If they do and then complain or drop off after signing up, and you can identify issues with the onboarding or product, it’s a sign that investing more time might be worthwhile.
Fail quickly if needed many ideas do fail. If your product has strong product-market fit, you’ll attract early users despite its shortcomings. The goal is to determine as fast and cheaply as possible if your product is worth pursuing. Sometimes you can test this without a full product, using an email list, design partner, or letter of intent, but generally, you should build a quick MVP to get users and gather feedback.
I believe that individuals don’t really care about creating quality items, thus they don’t take the time to do it.
Like a lion in a cage, they would rather concentrate on “EARNING MONEY”.
Bravo for having the mindset to put in the effort to create high-quality items; that’s what will reduce attrition. Fans will follow you, not just one-time buyers.
Those posts make my eyes glaze over. My SaaS was launched in 2008, and I’m still working on it, albeit after putting it off for a few years. Comment Image I don’t even make $8k a month, let alone nearly $80k. However, as I grow to know my consumers, I’m happy to build it gradually. (Btw, your AI assistant seems intriguing; I used Motion for a while, but it wasn’t functional for my purposes.)
I can see the advantages of employing AI to do proof-of-concept work because I’m a developer. However, in my experience, most AI-generated code requires some kind of modification before being included in a production build.
Regardless of the concept behind them, I just do not see how complete AI solutions advocated on this sub have any actual value; with software, value is derived from the intellectual property of the codebase. AI can “do it” repeatedly if it can do it for a single person. Other than the original concept, where is the IP?