What does it take to build and ship software in the AI era?

Like any job, there are three aspects to it:

• Knowledge: Things you know, industry best practice, etc.
• Ability: Innate ability. You have good visual taste or you do not.
• Experience: Things you've done.

But, building software isn't a single job. It's a complex enterprise with many aspects. Traditionally, software is broken down into six main jobs (there are others, but these are the big six):

• Product: Deciding what to build and why
• Engineering: Building it
• Architecture: Deciding how it is to build built
• Design: Deciding how it looks and functions
• Marketing: Deciding how to talk about it and delivering the message
• Sales: Defining, acquiring and closing customers

I've worked at places where there are literally six departments with those names. Architecture is normally part of eng and it's common that sales and marketing are related or joined. Many startups don't have a product team, the founders do that work. However, these are six distinct jobs even if they aren't distinct people or job titles.

What about the world of AI?

Well, in an AI world, all those jobs still need to be done. You may farm out some of them or all of them to AI, but they all still need to be done.

This begs the question: which ones to farm out and why?

For me, this is a skills issue. I know that AI is "average" or better in most of these things. That is to say, it's competent at architecture but a really skilled architect would be better than the AI would be. As a one man shop, I would love to hire an amazing expert to do my visual design but I can't afford that for my side project. So, I use AI for that because it's better than I am even though it's not as good as the best designer I've ever worked with.

As you begin your journey, you need to take a deep look at yourself first. What are you good at and where do you need help? If you try to get AI to jobs that you're amazingly good at, you will probably be underwhelmed. Instead, focus on things you're not good at.

To help my students with this discussion, I created a tool that asks a series of questions and then produces a spider graph. That's my spider graph there at the top. As you can see, I'm strong on product (Sr. Director of PM, etc.) and Architecture (former Principal Architect). But I suck at visual design and marketing. You can use this graph to figure out where AI can help you the most and where it can help you the least. It can also help you decide where you want to improve and skills you may want to acquire.

You can find a link to the tool on our training page.