Someone behind the software
Modern software increasingly feels like a feature-packed but incoherent “Frankenstein’s monster,” leading many developers and users to long for products that clearly reflect the intent, taste, and values of a human creator. Commenters argue that AI-powered “software factories” amplify existing problems—additive feature roadmaps, resume-driven rewrites, and data- or hype-led decision making—because they make it easier to ship more, not to choose better. The consensus is that the enduring value in software now lies less in writing code and more in product vision, empathy for users, and the courage to say no, roles that AI cannot yet convincingly replace.
Human craft and “mind of the maker”
- Many agree that good software carries the imprint of its makers: their worldview, experience, and care.
- Commenters connect this to older ideas: creation as idea → implementation → interaction, where the product is only “complete” when a user encounters the maker’s mind.
- Several compare software to pottery, food, or handcrafted goods: you can sense when someone truly cared.
Product vision, intuition, and leadership
- Strong support for “opinionated” products with a clear point of view, even if that means some users churn.
- Good product leadership is described as rare and centered on:
- Having a coherent vision.
- Being willing to remove or radically change features, not just add.
- Deeply understanding the domain and actually using the product (“dogfooding”).
- Some note that intuition vs data is a false dichotomy: they should inform each other.
AI, agents, and the “software factory”
- Many fear AI-driven “software factories” will amplify a preexisting problem: disjoint, additive features turning products into “Frankenstein’s monsters.”
- Others argue AI is just another implementation tool; the real leverage and responsibility stay with humans deciding what to build.
- Several insist AI lacks genuine values, empathy, and individual taste; others counter that models already embody “taste” as learned heuristics, just not necessarily good or distinctive taste.
- There is skepticism that AI can yet handle product judgment or user understanding, even if it codes competently.
Feature creep, bloat, and enshitification
- Strong consensus that many modern products keep adding capabilities instead of refining core workflows, often to satisfy growth metrics.
- Examples across software and social media: everything converging into “mega-products” or “labyrinths” that lose their original focus.
- Some tie this to organizational incentives, resume-padding, and stock-market pressure rather than user needs.
Solo devs, niche products, and future direction
- Several think solo or very small teams may now have an advantage: AI lets them move fast while preserving a single coherent vision.
- Others worry that as LLMs homogenize outputs, distinctive taste and voice—“someone behind the software”—will be the scarce differentiator.