Why senior developers fail to communicate their expertise
Senior software engineers are being pulled between business demands for rapid AI-driven feature delivery and their own focus on stability, risk management, and reducing complexity. Commenters argue that while AI can accelerate experimentation and “speed” loops, it also tends to produce fragile, poorly understood systems, heightening the long-term value of seniors who can design for maintainability, communicate trade-offs, and take responsibility when things break. Many note, however, that incentives, company politics, and a lack of appetite for mentorship or deep domain understanding often prevent this expertise from being heard or effectively transferred.
Perceptions of the Article and “AI Slop”
- Several readers felt the prose had an “AI-written” tone (rhetorical questions, staccato style), which some found distracting; others argued this is just modern copywriting and we must adapt.
- Some explicitly want to “resist AI slop” and said pervasive AI-generated text may be shifting how humans write, even when no AI is used.
- A minority thought the article was engagement bait or oversimplified, while others found it “clarifying” and said it expressed a real, hard‑to‑articulate dynamic.
Role of Senior Developers: Avoiders vs Innovators
- Many agree seniors should reduce unnecessary complexity, avoid premature optimization, and be able to say “no” or “not yet.”
- Others warn against blanket praise for “avoiders”: avoiding change can accumulate tech debt, hurt performance, and leave legacy stacks dangerously outdated.
- A recurring theme: good seniors know when to add complexity or experiment and when to cut scope; context (startup vs cash cow, CT scanner vs CRUD SaaS) is crucial.
Speed vs Scale, AI, and Two Loops
- The speed loop (fast experimentation, AI-assisted “vibe coding”) is seen as optimizing for uncertainty reduction and quick feedback.
- The scale loop (stability, maintainability, risk management) optimizes for reliability, understandability, and long‑term cost.
- Many doubt organizations will actually invest in a separate “stable” system once a fast version exists and makes money; “temporary hacks” tend to become permanent.
- Some argue AI should also be used on the stability side: tests, benchmarks, security reviews, instrumentation, refactors.
PoCs, Rewrites, and Tech Debt
- Multiple commenters report that “proof of concepts” almost always become production systems and promised rewrites rarely happen.
- Opinions split: some say rewrites are often unjustified “for purity”; others note they’re essential once scaling or existential constraints appear.
- Culture and incentives dominate: product and sales often push all‑in feature builds, don’t accept scoped‑down experiments, and underweight long‑term risk.
Mentorship, Tacit Knowledge, and Communication Gaps
- Seniors frequently report that juniors rarely seek mentorship, preferring internet/AI answers; juniors counter that corporate expectations punish visible ignorance and time spent learning.
- Several discuss “world models” or tacit knowledge: core expertise is an internal mental model built via experience, hard to fully transfer via docs or talks.
- Communication costs (“communication tax”) and lack of time/organizational support are cited as reasons seniors fail to share expertise, even when willing.
- Some see AI as excellent at surfacing facts but not at replacing deep domain understanding, judgment, or organizational context—areas where senior devs remain critical.