Bilingual

AI, Vim, and the Illusion of Flow
11 days 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
15 days 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
21 days 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.

Hasty Briefsbeta

ai in engineering

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