If coding has been solved, why does software keep getting worse?
Many developers and users feel everyday software is becoming buggier, slower, and more hostile to user needs, even as tools like AI-assisted coding supposedly make “coding a solved problem.” Commenters largely attribute this decline not to AI itself but to misaligned business incentives, product cultures that reward feature churn over quality, and ever-growing complexity layered on fragile foundations. Some point to open-source and carefully engineered systems as counterexamples, arguing that true improvements require valuing robustness, UX, and long-term maintenance rather than short-term velocity and hype.
Scope of the Problem: Is Software Actually Getting Worse?
- Many commenters resonate with “nothing works”: buggy banking sites, car infotainment, mobile apps, OS updates that users now dread.
- Others argue this is mostly visibility and nostalgia bias: software and platforms are vastly more complex, so small bugs are more frequent, while catastrophic crashes are rarer.
- Some say flaws have shifted: fewer hard crashes, many more “paper cuts” and UX annoyances.
AI, “Coding Is Solved”, and Quality
- Strong pushback on the idea that “coding is solved”: LLMs help write snippets, but don’t understand large systems, trade‑offs, or architecture.
- Several see AI as a force multiplier: good engineers get somewhat faster; weak engineers produce much more low‑quality “slop” more quickly.
- Others say management hears “coding is solved” as “we can ship faster with fewer humans”, not “we can build better software”.
Incentives, Management, and Product Culture
- Repeated theme: incentives favor shipping features over quality, robustness, or UX.
- Agile/MVP culture and non‑technical product owners are blamed for “happy‑path only” builds that accrete bugs and technical debt.
- Some describe large orgs where no one truly owns outcomes, teams glue around bad architectures, and success is measured by slideware and burndown charts, not working software.
- View that companies optimize for shareholders and short‑term metrics; as long as users don’t churn, there’s little incentive to invest in quality or support.
UX Regressions and Enshittification
- Focus stealing (e.g., chat apps popping to front and capturing keystrokes) is a lightning rod; some note Linux/KDE has tunable focus‑stealing prevention, others say major OSes ignore this.
- Complaints about intrusive pop‑ups, dark patterns, AI nags, cute but useless error messages, and constant UI churn (e.g., mobile OS redesigns).
- Security UX (e.g., repeated logins, 3‑D Secure flows) is seen as overcorrected and user‑hostile, though some argue it’s better than past insecurity.
FOSS, Platforms, and Alternatives
- Many advocate Linux/FOSS (especially KDE, NixOS, etc.) as relatively stable, user‑respecting refuges, though others report plenty of bugs and rough edges there too.
- Gaming and certain consumer hardware (e.g., new GPUs) remain practical blockers for some trying to abandon Windows/macOS.
Deeper Causes and Prospects
- Comments point to exponential growth of mediocre developers, deep dependency stacks, and “tower of Babel” architectures built on poorly understood libraries.
- Several argue software will only improve if market incentives change or if some companies successfully differentiate on quality, UX, or support—and are visibly rewarded for it.