The crucial point of the above illustration is that, from a current perspective, AI tools assist workers. However, as they surpass Accuracy and Creativity, AI with specific capabilities will eventually take over these tasks. This could lead to a change in organizational structure, as depicted in the image below.

Here, business growth is measured not by the number of personnel but by computing power. In other words, the decision-making structure narrows down to 1-2 people, with AI handling the rest. Does this mean AI will completely replace humans? Not entirely. This is just one perspective on AI, and viewing AI from multiple angles is beneficial.

Today, AI proponents outnumber skeptics. For instance, there was a live poll at the ADFEST 2024 held in Thailand in March. Is AI a friend or foe? An overwhelming 92% of respondents answered that AI is a friend. Today's AI is far from perfect and does not accomplish everything at 100% efficiency. From the perspective that AI is imperfect, I'd like to introduce the word "Slop." Recently, when someone asked Google, "How do I get the cheese to stick to the pizza?" AI humorously responded, "Use glue to stick the cheese." Such instances of providing unnecessary or inaccurate information are known as "Slop," implying 'cheap junk.' Consequently, more discussions and judgments about the quality of AI outputs will arise.

#MAIN BODY – How to Enhance Performance with AI…

So, how can we enhance performance using AI, which seems poised to replace and befriend us yet is not 100% perfect? It's about finding your own 'washing machine' and using it efficiently.

Even today, to operate a washing machine, we need to load it with laundry from the basket, manipulate the buttons, and transfer the washed clothes to the dryer or hang them out to dry. Although a washing machine might seem inconvenient, it's a valuable tool in the pre-washing machine era. Our work is no different. Tasks that AI can handle should be left to AI, while the rest we do ourselves. Comparing work to a washing machine might seem a bit unrefined, so let's depict it in a more sophisticated manner:

Complex processes and automation exist, but humans still play a role. In such a workflow, if efficient AI tools are available, we should use them.

Who are the people achieving these three kinds of results? Naturally, it's those engaged in the relevant tasks. No one can do your job better than you. You can achieve all three kinds of results by distinguishing between parts of your work where AI can be utilized and parts where it can't and by refining the 80% AI output to a 100% result.

If you categorize your entire workload from A to Z, determine the C areas suitable for AI application, and use AI tools while accumulating effort, personal growth and notable performance in the respective tasks will be apparent.

Of course, there will be counterarguments. How can we possibly learn and try out all these tools? As a fellow human, I, too, wish for the emergence of a miracle-working AI Super App. But we must keep learning whether a super app or Artificial General Intelligence (AGI) comes along. Don't worry too much, though. While we may not get everything, assistants like NVIDIA's "Project G-Assist," which guides us on what to do and how to do it, will undoubtedly appear.

We always have tasks to accomplish. When the time comes to complete these tasks, having an AI assistant like the one shown in the image above can help us learn and achieve much more quickly than we currently do.

Jim Keller, CEO of Tenstorrent and often called the father of AI semiconductors, once said, "If you don't know something, ask. From a young age, I always asked questions if I didn't understand something. Some people refrain from asking questions to appear smart, but that way, they learn nothing. It's a common mistake among young people. Don't try to look smart; try to learn."