Today is the worst AI will ever be.

Expectation Management

Becoming a programmer takes years or decades, and requires both a particular mindset as well as specific skills.
AI can help you bridge the gap between ideas and prototype – turning your prototype into a product might require skills and oversight which might go beyond your current capabilities.
The best way to get a sense for where your own skills and weaknesses are when it comes to building your own apps is through experimentation. AI models are progressing incredibly fast, and in all likelihood you have a lot to catch up with when it comes to the state of the art capabilities of coding with AI.
AI can help you fill in the blanks in any dimension you can conceive of – but language models are often incapable of making judgement calls about which questions to even ask.

Having technical skills pays off at all levels of software development. AI can take you further than you thought was possible, yet the ROI for having a fundamental understanding of all areas you are getting involved with is immense.
Software can break in unexpected places. Most people are used to interacting with code only through the UX, but having a programmer’s perspective of things like user flows, authentication and databases will help you build better apps.

While traditional code is created by characters and lines, AI programming is all about asking the assistant your implement “chunks” which might correspond to a feature, implementation or bug fix. This does not absolve you from the responsibility of having correct code.
The more time you spend programming with AI, the more appreciation you will have for the actual code that is being generated. You might abstract it away, but the computer doesn’t.