Breaking Past Slop: Why AI Feels Unreliable, and How to Use It the Right Way

Right now, we are seeing a massive increase in AI-generated content, and it continues to grow by the day. Most of this content is hard to follow and doesn't actually achieve anything by the end. In the end all we have is a frustrated audience and an offended author when they are called out.

This benefits nobody and wastes our precious resources. Not just time and attention, but the literal electricity and water required to run these models. If you’re going to use AI, it should produce value, not generate low-quality, copy-pasted content that burns resources for nothing.

So why exactly is AI so unreliable?

AI is objectively dumb and doesn't actually know what it's talking about. It's essentially stitching together predictive text, and presenting in a way that sounds intelligent. It will never say "I don't know" unless prompted. Despite this, AI has been sold as a complete replacement for your work, thinking, and involvement in your business. This is a bad attitude to have and will surely lead to the destruction of your reputation and monetary losses.

I personally discard between 40-60% of the content in each AI output. I don't think of it as a replacement, but as an assistant whom I've hired to take some of my mental load. Its purpose is to fix grammar, hold onto my thoughts, and spark ideas that my busy mind is too preoccupied to think of.

Just because it's unreliable doesn't mean it's useless.

AI is incredibly useful for people who already know what they’re doing: experts, intuitive thinkers, and fast problem-solvers. It gives you raw material you can shape without slowing down. That combined with your own expertise is extremely powerful.

In the past if we needed inspiration, we would have looked to books and internet searches for snippets of information. This was extremely time consuming and you'd need to manually compile the information into notes that could get lost. Even though AI isn't smart, its speed cannot be matched. One of its greatest strengths is its ability to surface and consolidate massive amounts of information and hold it in its memory for you.

In this age of AI, finding inspiration and brainstorming is easier than ever. You can go through 10 ideas in the same time you might have started just one. This is incredibly powerful for fast-paced minds, especially if you already have expertise to guide and filter what it gives you.

Additionally, if you stay within the same AI app for all of your workflows, that AI can remember all of your ideas and progress and bring them forward when they are needed. Beyond speed, AI's other strength is continuity.

So not only can AI compile massive amounts of outside information, it acts like an extension of your working memory - a place where ideas don't get forgotten.

If you are someone who's mind works faster than you can keep up with, AI is invaluable. I have severe ADHD and my notes would always get lost. I'd never remember to go back to the ideas and the best ones would never reach fruition. With AI, I can ask it to list ideas I've had that are relevant to what I'm working on, and it'll offer advice on the best ways to implement them. If you want structured prompts and workflows that use this kind of personalized recall, that’s exactly what I teach in Lucidium.

All in all, AI is less reliable than your own brain, but can fill in for the parts that are overworked. You don’t need complicated “prompt engineering”, you simply need a realistic understanding of AI’s strengths and limits, and how to work with them intuitively. Breaking past the hype and learning to use it this way will give you more reliable and high-quality results that actually convert.

Have you relied on AI in the past? How did that affect the quality of your work?