When everyone has AI and the company still learns nothing
Enterprises are pouring millions into AI coding tools like Copilot and Claude, but many engineers report that faster individual productivity isn’t translating into real business gains. Commenters describe organizational bottlenecks, perverse incentives, and hostile cultures where sharing AI workflows can feel like training your own replacement, while management chases token-based ROI and surveillance-style metrics. There is broad agreement that AI can be a powerful personal accelerator, but without changes to process, incentives, and shipping practices, companies risk bloated codebases, lost institutional knowledge, and little lasting advantage.
Impact on developer work and collaboration
- Several comments worry that AI reduces peer interaction: devs ask models instead of colleagues, weakening mentorship and “answer people” who deeply know the codebase.
- Others see AI as a strong aide for navigating large, complex codebases and quickly explaining unfamiliar parts.
Productivity gains vs organizational bottlenecks
- Many note that coding speed was never the main bottleneck, especially in large enterprises.
- Real delays come from infra, compliance, testing, approvals, and release processes; AI just piles more unshipped code at the gate.
- Some describe “two timelines”: official engineering with slow governance vs. fast “vibe-coded” side systems built by non‑engineers or data scientists.
Knowledge sharing, incentives, and workplace dynamics
- Strong thread on misaligned incentives: devs feel no reason to share AI workflows or internal tools if it brings extra support burden without pay, and may even threaten their job security.
- Others counter that hoarding tools is toxic and that sharing has historically driven promotions and career growth.
- Underlying theme: companies treat employees as disposable, so many respond by treating companies as adversarial.
Code quality, technical debt, and risk
- Multiple developers report subtle AI‑introduced bugs and worry that people overestimate AI accuracy.
- Fear that AI will bloat codebases and documentation, worsening maintainability and locking firms into particular vendors.
- Concern about “vibe‑coded” prototypes becoming unmaintainable production systems, especially when built by non‑engineers (e.g., finance teams).
Measurement, ROI, and token costs
- Skepticism that expensive AI subscriptions will show clear ROI once investors demand hard numbers.
- Attempts to tie AI gains to story points are seen as easily gamed; some expect AI usage metrics to morph into employee surveillance (“tokenmaxxing”).
Broader employment and cultural concerns
- Widespread anxiety that AI will justify layoffs, eliminate junior roles, and erode institutional knowledge.
- Others argue AI mainly amplifies capable engineers and that resisting it increases layoff risk.
- Some note that while companies “learn nothing,” AI vendors quietly accumulate the real organizational knowledge.