Notes From The Real World

I've been exploring beyond AI Twitter lately...

Talking to more potential clients and working on more real-world projects.

And what I've come to realize is that when you spend too much time on AI Twitter, you start to overestimate the average person's familiarity with AI.

Most people know far less about AI than I initially thought.

This realization was kinda disheartening at first.

But after pondering on it for a while, I realized that it also means the opportunity for those of us who are immersed in AI is way bigger than I originally thought.

So please allow me to share the biggest lessons I learned by working with more clients and also the exciting opportunities that come with it.

1. Most People Are Still Clueless to What AI Can Do

Every day people still don't understand AI's capabilities.

Chances are, they have only played with ChatGPT a couple of times and only use it for basic use cases.

Of course, they recognize its power, but still don't know how it will all play out outside of the most simple workflows.

Most people are at least in one of the three categories below:

• Plain clueless;

• Don't know what are AI's strengths and weaknesses;

• Overwhelmed by the possibilities and unsure how to prioritize.

For example, while talking to a potential client about an automation solution for him, I suggested that we should start by first creating a prompt...

Then he looked at me completely confused: "What do you mean by prompt?"

This is more common than you can imagine.

Many concepts that are second nature to us are still completely foreign to the average person. Don't forget that.

2. Good Prompt Engineering Feels Like Magic To Most People

Most beginners in AI:

• Struggle a lot with hallucinations and incorrect outputs;

• Don't know how to use information outside of GPT's knowledge cutoff;

• Cannot get outputs that consistently abide by the quality standard guidelines of their tasks;

Most people get outputs that are often random, not reusable, and require lots of back and forth. (The complete opposite of what automation strives to achieve)

However, if you master prompt engineering and automate a task they thought impossible, they'll look at you as if you are some kind of wizard.

3. Start developing the skills you need to thrive in the coming era of AI

The skills needed to excel in the upcoming AI era are not as complex as some might think.

Most clients aren't looking for fine-tuned models or some crazy out-of-the-box automation.

They mostly need help with creating databases, writing prompts, scraping data, and automating stuff.

More often than not, you don't even need to learn how to code if you are creative enough with no-code tools.

The good news is, you can learn all of these skills online.

Just choose a good Udemy course and you'll be ready to start learning them. Heck, even I have a prompting guide myself.

However, you must be willing to invest time and effort to acquire these skills if you haven't already.

There is simply no way around it.

4. Most Clients Still Don't Know How To Ask For What They Want

Since many people are still new to AI's capabilities, it's not surprising they struggle to articulate their ideas and goals.

And a big part of your role in these projects is to help clarify their thoughts.

By doing this, you'll earn their respect and trust, as they'll recognize your expertise and knowledge in the field.

For example, there's a lot of confusion between fine-tuning and RAG (Retrieval-Augmented Generation). Clients often mix them up, but they're different.

RAG (mostly) involves adding external documents to prompts, not altering the model's statistical parameters.

Fine-tuning is more complex and expensive, requiring extensive data, retraining, and changes to the statistical parameters of the model.

When clients refer to fine-tuning, what they are usually referring to is the ability to use information outside of the knowledge cut-off of the AI model.

A task that can be easily achieved through RAG.

But if you take their word at face value, you will often end up building the wrong thing.

You are the expert here.

So take some time to clarify with clients exactly what they need to avoid misunderstandings and overengineered solutions.

5. People Still Don't Know What Tools To Use

Most people don't know which AI tools are the best for different tasks.

They are so far removed from the AI bubble that they can't tell which tools are good and much less which ones are emerging.

For instance:

• Chat With PDFs tools are still mind-bending for a lot of people;

• Lots of people still use ChatGPT for research, and never heard of perplexity or anything similar;

• Almost everyone is still unaware of Grok;

Believe it or not, knowing which AI tools can be applied in every scenario is valuable knowledge.

Lists of AI tools go viral on Twitter every single day for this very reason.

Now, I hope you have noticed the common theme of all of the lessons, and that is: The average person is still clueless about AI in general.

And it shouldn't come as a surprise to you, since AI is our world, not theirs.

They might have used ChatGPT a few times but likely haven't thought deeply about prompt engineering nor looked for ways to automate and create systems in their work.

Which means the knowledge gap between most of us AI weirdos and the average person is wide.

We still have a long way to go in teaching others how to leverage AI and helping businesses become more efficient.

And, if you're new to AI and you're the "average person" I'm referring to, learning about AI shouldn't take too long.

You can improve your AI skills relatively fast, and then you can start experimenting with it and creating automation workflows.

It's not too late to get started.

I firmly believe the future belongs to those who are learning about AI, and it's up to us to seize this opportunity.

Remember, it's been just over a year since ChatGPT was released, and we're still in the early stages of the AI revolution.

PS: Need help implementing AI in your business?

Let's talk about how you can leverage AI.

• AI Automation;

• Prompt Engineering;

• AI for Content Creation & Social Media Growth;

• And much more.