Producing great artistic work is incredibly hard, especially when you're doing it alone. Why is it so difficult? In my experience, it's because the work can't speak back to you. In many non-artistic professions, your work gives you immediate feedback. Code tells you when it doesn’t work. Equations don’t balance. Cars don’t start. Accounts aren't profitable. But with artistic work, the feedback is often slow and ambiguous. Sometimes, what seems like an error becomes a unique feature of the piece. This can lead to an unhealthy relationship where you become your own harshest critic.

On the flip side, artistic work can't easily be outsourced. The soul of the art often gets lost in boardrooms and team meetings, where ideas are reduced to mere tasks.

We once believed that the gargantuan titan known as AI would free us from this burden. Unfortunately, we were mistaken. While AI-generated art may look impressive at first glance, it lacks depth. As someone with a scientific background, I find this disappointing. Has AI been a poor investment? The short answer is no.

AI is like the self-checkout at a store. It won’t help you shop, but it speeds up the checkout process, allowing one staff member to manage six, maybe even ten, registers at once—a tool no store knew it needed.

So, what has AI done for us? At its core, AI is a statistical model. Depending on the provider, it has 'seen' everything. If your work is a painting, AI has observed every brushstroke and style ever produced. It can replicate the best techniques, but it can't create anything truly original. It's not unique yet (a nod to I, Robot fans).

So, how can AI help you? It can be your best critic. Create your work and let AI provide feedback. While the feedback may be imperfect—since no one, not even you, fully knows your end goal—it’s there, and it comes at a rapid pace. AI can help maintain consistency in perspective, harmony in strokes, and balance in sound. Though AI hasn’t produced the artist we hoped would do the creative work for us, it has given us the assistants we needed to make that work easier.

What does this mean for you? You can critique each of your strokes individually or as part of the whole. Either ask AI to analyze specific strokes, or have it suggest the next move — so that you can decide whether it is a bug or a feature. This approach is akin to the shortcuts often used in non-artistic professions. Instead of checking each water tap for a leak one by one, you flood the entire system and spot the problem easily. As an assistant, AI can make your work feel as light as a bubble floating through the air.

When using AI, I like to follow two paths: What is the quality of my work so far? Where can my work go?
In my experience, these two questions form the foundation of my creative hurdles. The difficult part is training the AI to be genuinely useful. It must be explicitly told what is required, the overall objective, key details to focus on, the desired level of grammar, and many other specifics. AI is only useful when properly trained; otherwise, it becomes generic, non-unique, and bland.