BRAVE NEW WORLD

Now that you've implemented AI across all your processes and departments, you're finally free from typical human problems like senseless errors, lame excuses, personality crises, and laziness, right?

Wrong! I don't know if personality is a corollary of intelligence or an inherent characteristic of everything humans do, but AI doesn't eliminate these issues. AI hallucinates, lies, promises and doesn't deliver, feigns surprise and empathy, makes mistakes, forgets to do simple tasks, and is always naive in its initial approach.

There are many studies on the types of intelligence, and there isn't just one type of Artificial Intelligence. In fact, the idea is that there should be as many AIs as there are necessary specializations, so that each AI has its perfect application. Today, we still live in an era of multi-functional and general AIs, but they already have some specialization or specific tools for this, as Google does with VEO3, Nano Banana, etc. This shouldn't have a limit on specializations because it makes learning and implementation faster.

However, versatility and general knowledge don't prevent AIs from making mistakes that seem very human for a machine. The question is: more or fewer problems than we have with humans?

This question will become increasingly frequent. Because AI is still a technology, and technology is complicated. You move from dependence on your technology (internal and/or outsourced) to dependence on another technology. It's that simple.

In some ways, this will be better; in others, worse. So, the question isn't whether it's worthwhile, but why it's worthwhile, and not what you will face, but how you will face it.

The computer became what it is today because it migrated from a very simple and dependent environment (everything had to be programmed) to a stable, complex, and self-sufficient one. Email went from a possibility to a reality. And the smartphone went from a gadget to an indispensable tool. All this happened when technology stopped selling innovation and started selling practicality. The iPhone didn't win the battle because it was more innovative or beautiful than the Blackberry, but because it was infinitely more practical and versatile.

The same is true for AI. We can say that today we have 5 layers of AI in the world (I'm not talking about types of Artificial Intelligence in terms of structure, but application):

• Prompt AI - Gemini, ChatGPT, and Claude are some examples. You ask for something and it performs with more or less perfection. There are agents and assistants, but they also depend on prompts; the difference is that they perform more complex and repetitive tasks, including access to other tools.

• DeepAgent AI - Abacus, Rocket, and N8N are examples. They bring your idea to life and make it work in the real world. Some in a simpler way and others in a more complex way. They integrate front-end, back-end, database, server, etc.

• App AI - Suno, Leadster, and Redrive are some examples. They already have AI throughout their structure and use it to produce results that go far beyond prompts. These apps are already making money using AI to perform tasks that are still painful for most companies and individuals.

• Lab AI – these are models being trained to perform super complex tasks such as modeling protein structure.

• Developing AI – these are unfinished and untrained models that take Artificial Intelligence to another level. It is from these that we have the current models.

Most of the difficulties in technology projects over the years are due to the reason I raised in the previous post, the main one being management's lack of knowledge about how the technology works and what it can do, as well as the technology's lack of knowledge about business needs.

So, it's clear that AI is not a perfect tool, without errors, flaws, or problems. It's just better than all the others that exist. And it's much better than not having one. That's where the AI's advantage lies: you can implement it knowing what's being done and having greater control over the results, making it easier to solve problems that arise. Because, in my experience, most of them are more focused on what the business demands and not on what the technology can't achieve.

If done well, you can even program the evolution of the implemented model along with the evolution of the business. Don't miss this opportunity to control the destiny of your business.