Is Software Getting Worse?
Many engineers argue that everyday software feels slower, buggier, and more bloated despite massive hardware gains, blaming incentives that reward rapid feature shipping, ad-driven “free” apps, and disposable, web-based tooling over efficiency and reliability. Others counter that overall reliability and capabilities have improved dramatically since the floppy disk era, pointing to robust operating systems, security advances, and mature tools, and noting that complex modern requirements (globalization, accessibility, sync, media, security) inevitably add overhead. Across both views runs a common theme: markets and management optimize for visible features and growth, while performance, simplicity, and long-term maintainability are undervalued until failures or user frustration become too costly to ignore.
Is Software Getting Worse? (Overall Sentiment)
- Many argue “yes”: more bloat, slower apps despite faster hardware, unfinished “permanent beta” releases, and disrespectful behavior (tracking, ads, forced changes).
- Others say “not really”: crashes and OS instability were worse in the 80s/90s; modern systems, compilers, and tools are more robust and accessible.
- Several note that new software is often bad, while long‑lived projects (OS kernels, compilers, major apps) have steadily improved.
Performance, Bloat, and Resource Use
- Frequent complaints about chat/email clients (Slack, Teams) using gigabytes of RAM and heavy CPU just to show text and images.
- Some defend modern chat tools as genuinely more complex (history sync, media, threading, moderation, APIs, globalization, accessibility).
- Others say that doesn’t justify current performance; older clients (AIM/ICQ/MSN/IRC) offered similar functionality with tiny resource footprints.
- Resource abundance (RAM, CPU, bandwidth) is seen as enabling sloppy design; some teams intentionally constrain dev environments to catch inefficiencies early.
Economics, Incentives, and “Quality”
- Strong theme: organizations reward feature delivery and growth, not speed, reliability, or resource efficiency.
- PMs and management often prioritize visible features (and resume bullets) over refactoring or performance work.
- Users rarely choose software for efficiency; they pick “good enough” tools with needed features, or what IT mandates.
- Free/ad‑supported models and “attention harvesting” push toward surveillance, tracking, and engagement over craftsmanship.
- Several note that “quality” is multi‑dimensional (performance, reliability, UX, features, privacy) and highly subjective.
Complexity, Security, and Distribution
- Platforms and stacks are more complex: web apps, app stores, multiple device types, localization, accessibility, security layers.
- Security hardening and global text support add overhead but are considered necessary tradeoffs.
- Internet distribution shifted incentives: shipping patches is cheap, so 1.0 releases can be much buggier than the CD/floppy era, when support and patch distribution were expensive.
Nostalgia vs. Real Regression
- Some see nostalgia and selection bias: we remember the best old tools, forget the bad ones.
- Others point to concrete regressions: formerly lean tools (e.g., lightweight clients, small utilities) becoming bloated, subscription‑tied, or ad‑ridden.
- Broad agreement: most software has always been mediocre; what changed is scale, incentives, and how visible the bloat has become.