The hottest Engineering Substack posts right now

Planes take off and land at speeds between 250 and 310 kph. In places like Bangalore, the altitude makes the takeoff speed a bit faster because the air is thinner.
When cruising, planes fly at around 39 to 41 thousand feet and can travel at speeds of about 800 kph. From up there, you can even see highways on clear days.
Planes begin to descend over 200 kilometers before landing. Once they go below 10,000 feet, it usually takes just over 9 minutes to touch down.

Publishing a serious poetry collection and performing publicly opened up deeper creative authenticity and stronger emotional connection with audiences.
Work moved into large, cross-team engineering projects where coordination, trust, and leadership matter more than raw coding, while tools like observability and agentic coding expanded technical impact.
New coaching and workshop formats supported others' transformation, and clear 2026 plans aim to scale artistic performances, a poetry reading tour, more coaching cohorts, and refined consulting and technical focus.

Automated status messages can be contradictory, for example saying a device is repaired while warning the keyboard or screen may not work, which confuses customers.
If engineers or technicians don’t use and test the system themselves, they won’t notice silly or harmful messages that slip through.
This gap between automation and human oversight can cost time, trust, and business; either update the system or make sure people regularly use and correct it.

Using passive language in reports hides who actually did what and makes it hard to hold anyone accountable.
Paperwork and process fixes are useful but not enough; if root cause analysis ignores human mindsets and norms, the same failures will recur.
Leaders need to watch how people really behave, name specific actions and responsibilities, and enforce accountability to change harmful cultural habits.

Many products end up with absurd, unusable features because no one on the team ever pays attention to real users or real-world use.
Make the customer’s needs omnipresent: short release cycles, engineers talking to customers, and seeing real usage expose design problems quickly and stop bad decisions spreading.
Create a culture where anyone can flag absurdity by encouraging psychological safety and cross-functional responsibility so problems get fixed instead of ignored.

Get a weekly roundup of the best Substack posts, by hacker news affinity:

Treat mistakes like tracer bullets: run cheap, fast experiments that will often be wrong so you get immediate feedback and learn quickly.
Don’t stop at technical fixes — use failures to change culture and processes so you root out over-optimism, whitewashing, and the normalization of bad practices.
Build disciplined rapid-learning loops: ask why repeatedly, do frequent post-mortems, and update norms so teams converge on the right results and make mistakes harder to repeat.

Vague goals and prompts cause complexity to explode, so define clear objectives, boundaries, and success criteria before asking for reviews.
AI will mirror the complexity you give it, so act like the expert: do the hard thinking internally and ask the AI for focused, constrained help.
Complexity is contagious and avoidable — interrupt runaway design early by questioning whether a system should exist, simplifying the problem, and realigning on the real objective.