The Problem Isn’t AI. It’s Lazy Creators

There is a clear backlash against software written by AI, something we hear from some developers and consumers alike. The reality however, is that most if not all software will be partly written by AI, it’s inevitable and even irresponsible to not do so.

The problem isn’t using AI; it’s the laziness of shipping a product without thinking through and owning its decisions, denying users the care and craftsmanship they deserve.

Let me say it another way, the problem with products partially or entirely written by AI. It’s lazy creators who accept its decisions instead of exercising their own judgment, owning the details, and making a product worthy of their users’ time.

Here, “lazy” doesn’t mean using a shortcut, working fast, or not knowing how to code. It means avoiding the work of deciding what the product should be and whether each choice serves the person using it. At the end of the day, the way you build your product doesn’t matter to me (as long as it’s safe for my devices/data) but if you didn’t spend the time and effort to think through your product decisions and UX, I don’t want to deal with your slop. Speaking of slop, the internet is full of it and you probably don’t want to read a long essay on the topic so let me give you bullet points.

1. Responsibility: AI isn’t responsible for a generic, thoughtless, uninspiring product. The person who ships it is.
2. Authorship. Making something is more than getting it to work. It’s choosing, questioning, testing, rejecting and refining until the product reflects your true intent. You can be the author without having written a single line of code, but only if you truly designed and shaped the experience.
3. Respect. Shipping without that care is disrespectful to users, who receive a product its creator hasn’t taken responsibility for. That’s my definition of slop.
4. Positive example: recently the cheap $50 FM1 mini synthesizer community created hundreds of embedded music software projects for the device. It shows that someone who can’t code can still be an author: the creativity and judgment are theirs, even when AI handles implementation. Some of those projects are really awesome!
5. Opportunity. AI gives us more capacity. The challenge is to invest it in greater breadth, depth and attention to detail, not just ship the first working result. It’s also an amazing opportunity to be innovative, let’s not waste it.

In conclusion, my thinking can be summarized as “you can delegate the code, not the judgment”.