ai in engineering

Bilingual

AI doesn't generate working products, that's still your job
2 months ago
- AI tools can rapidly create working prototypes, but the distance from a prototype to a production-grade system remains significant.
- Building production-grade software requires judgment, system design, error handling, security, scalability, and observability—areas AI does not automate.
- Learning computer science is more valuable than ever because it provides a mental model to evaluate AI-generated code for flaws like performance issues or race conditions.
- The demand for mechanical coding is declining, but engineers with deep understanding who use AI as a force multiplier will reach new productivity heights.
- Engineers who rely on AI as a substitute for understanding will be left behind, unable to fix, scale, or explain their systems.
- Success comes from using AI as a tool on top of strong fundamentals, not as a replacement for them.
AI, Vim, and the Illusion of Flow
2 months ago
- Vim enhances flow by making text editing seamless, not by increasing speed, and it directly translates user intent into output without distortion.
- AI tools promise similar fluidity but introduce a 'depth problem': their output appears polished but may not accurately reflect the operator's intent or skill, making it hard to spot flaws in code.
- The gap between 'looking done' and 'being right' is growing, as AI prototypes lack edge cases, error handling, and other critical details, leading to unrealistic expectations.
- AI fosters an 'empathy gap' where surface-level contributions across disciplines create false confidence in understanding others' jobs, undermining organizational respect for depth.
- To address these issues, we need to set clear boundaries for AI output, invest in education on evaluating AI work, build cross-disciplinary empathy, and protect personal flow by ensuring AI enhances rather than replaces problem-solving.
Can AI build a jet engine? JARVIS Challenge tests AI copilots
2 months ago
- The JARVIS Challenge at MIT tested if AI can accelerate the design-build-test cycle for a small gas turbine engine, with undergraduate teams building engines in four weeks.
- AI tools, accessed via the Parley platform, helped with tasks like summarizing textbooks, software tutorials, vendor sourcing, and design trade studies, but faced limitations like hallucinations and lack of physical understanding.
- Teams found that engineering judgment and expertise were crucial; while AI assisted, human oversight was needed to trust, challenge, and translate AI outputs into working hardware.
- Younger students used AI more frequently, while senior teams relied on deeper experience; the winning team was initially resistant to AI, trusting fundamentals and teamwork instead.
- Manufacturing and vendor relationships emerged as rate-limiting steps, with personal connections proving more reliable than AI-sourced vendors for tight timelines.
- The challenge demonstrated that AI can compress engineering cycles, but education and hands-on experience remain vital for developing the judgment to effectively lead AI tools.
You're Not a Better Engineer Because You Type Git Commands by Hand
2 months ago
- AI agents handle tedious engineering tasks like commit messages, PR descriptions, and code comments, saving significant time.
- AI ensures consistency in naming conventions, Git hygiene, issue tracking, and release workflows, reducing errors and conflicts.
- LLMs benefit from structured codebase context (comments, rules, paths), enabling more fluid AI-assisted interactions and challenging inconsistencies.
- Agents manage Git operations (branch creation, PR preparation, label updates) and automate delivery checks, turning rituals into efficient sequences.
- Delegating repetitive tasks to AI prevents burning money and life on mechanical steps, allowing focus on solving actual problems.

ai in engineering

AI doesn't generate working products, that's still your job

AI, Vim, and the Illusion of Flow

Can AI build a jet engine? JARVIS Challenge tests AI copilots

You're Not a Better Engineer Because You Type Git Commands by Hand