The Substack CEO's interview on the "AI slop problem" names a discomfort most of us have felt but few have articulated. We are not just drowning in low-quality content; we are being asked to treat it as an acceptable byproduct of progress. That framing is convenient for platforms that profit from volume, but it does a disservice to the people actually trying to build and use these tools. We have seen the same tension play out in our own coverage, where Talking to My AI Clone Taught Me to Question the Tech forced us to sit with the unease of interacting with something that mimics understanding so convincingly. The problem is not that AI generates text; it is that we have stopped demanding a reason for that text to exist.
The interview rightly pushes back on the idea that "more content" is a neutral goal. When every tool is optimized for output, the incentive shifts from answering a question to filling a void. We see this in the practical advice we have given on Unlock LLM Training: A Practical Guide to Distributed Algorithms, where the focus is on rigor and reproducibility, not on generating endless variations of the same insight. The slop problem is not a technical failure; it is a design choice. If a system is built to produce ten articles where one would do, it will. The responsibility falls on the humans who set those parameters, and on the rest of us to refuse to accept noise as a fair trade for convenience.
What makes this interview worth our attention is that it moves the conversation from "AI is bad" to "AI is being deployed with the wrong intentions." That is a distinction our readers can act on. When you are evaluating a tool, ask whether it was built to help you think or to help you scroll. The former is worth your time; the latter will happily take it. We have made this point before in Verify Your AI's Understanding: A Simple Check for Tax Season, where the lesson was simple: verify before you trust. The same instinct applies here. Do not just consume what an AI produces because it is fast. Ask what it is optimizing for, and whether that matches what you actually need.
The open question the interview leaves us with is not whether we can solve slop, but whether we have the collective will to demand better. The concrete point to watch is the next time a platform announces a new "creation" feature. Count how many of those features help you make something you could not have made before, versus how many just help you make more of the same. If it is the latter, we are not being served; we are being used as the product. That is the takeaway to quote: the real problem is not the AI, but the absence of a reason. Bring the reason back, and the slop will take care of itself.