Hoang Nguyen

If you are still coding the same way you did two years ago, you might be falling behind without realizing it. Learn more why 👇

Leadership isn’t about being right. It’s about getting it right.

@simonw uv and ruff earned trust by being fast, focused, and community-driven. Acquisitions change at least one of those.

Thoughts on OpenAI acquiring Astral and uv/ruff/ty simonwillison.net/2026/Mar/19/op…

@bentlegen Scope creep dressed up as a pivot. The tools that last are the ones that stay boring and focused.

You either die a codegen tool or live long enough to become the everything app

@sundeep Token burn rate as a KPI. That's how you end up with a codebase nobody can debug without an LLM.

“If your $500K engineer isn’t burning at least $250K in tokens, something is wrong.”

@kdaigle @github Benchmarks measure what the model can do in isolation. Code survivability measures what developers actually trust. That gap is where the real story is.

Hot take from looking at @github Copilot telemetry: benchmarks make coding models look wildly different. Production workflows make them look much more similar. 👀 We looked at 23M+ Copilot requests and examined one simple metric: code survivability.

@_developit Users building workarounds for your app is the feedback loop nobody's reading. The code ships, the flow doesn't.

Has anyone else noticed mobile frontends (native and web) are getting increasingly terrible (as we descend into UX vibe coding?)? I keep noticing I've developed ridiculous coping strategies like constantly copying input to clipboard in case it gets lost when sending.

A culture that avoids risk avoids growth.

@tobi The 10x shows up in the demo. The quality shortcuts show up three months later when someone has to fix them.

It has been exactly 1 year since I wrote that AI memo. People are definitely reaching for AI reflexively.

@ClementDelangue Agents got good at generating PRs. Nobody built the judgment for when not to submit one. That's the gap.

Our biggest open-source repos are getting overwhelmed by AI slop which literally makes Github unusable (~a new pull request every 3 minutes). Fun new challenges in an agentic world!

@thdxr love to connect and trade ideas with you Dax

one thing to appreciate during hype cycles is it is very easy to spot people worth developing relationships with there's people building in the ai space that are ruthlessly honest and care about being right in the end, not just momentarily they stand out among the craziness

@rakyll The 'if' is doing a lot of work in that last sentence. Best time for independent engineers, rough time for the ones building someone else's empire.

Dynamics in software engineering is changing and it’s disruptive. Corporate empire builders are in the same position. The playbooks that worked for decades don’t work anymore. It’s the best time to be an engineer if you have some capital and autonomy.

Not an easy time to be a VC. For decades, they asked young people to give up everything, drop school, and write code until they find a market fit for a possible new unicorn. They created echo chambers and cults. Now those young people are replaced by agents and everyone has them.

@cnakazawa Tech debt from agents compounds differently. You can't refactor what nobody on the team actually understands.

Use coding agents. Go into tech debt if you have to.

@GergelyOrosz The tell is when the reply is more polished than the thought behind it. Real opinions have rough edges.

It’s not X — it’s Y I cannot unsee how so much of the writing on this site (and online, in general) is increasingly AI-generated. It’s still pretty easy to recognize. Probably not for long tho Just alarming that ppl outsource even typing 3 sentences for a reply on this site…

AI cannot replace judgment built from experience.

@unclebobmartin Every big refactor I've reverted taught me the same thing: the tests were passing but the structure wasn't ready.

I started a major reorganization of the empire game yesterday. I was confident that all my tests and rigor would keep the behavior stable. I was wrong. The behavior began to degrade in a way that the tests were not detecting. So after a couple of hours I reverted the work back to a known stable point. And now I'm going to proceed in much smaller steps, being careful to anchor the behavior with even more tests. AIs move fast, and they can take you off the rails at pretty high speed. Caution is always warranted.

@martinfowler @techygarg Half of working with AI is re-explaining what you already decided. That overhead adds up fast.

NEW POST Conversations with AI are ephemeral, decisions made early lose attention as the conversation continues, and disappear entirely with a new session. @techygarg explains how to externalize the decision context into a living document. martinfowler.com/articles/reduc…

@mitsuhiko Slop scales for free. Your attention doesn't. That fight was never fair.

Trying to fight the slop but the slop fights back :(

The best engineers design for clarity first, performance second.

Every new wave in tech has this moment. You learn a new tool, ship a few things fast, and suddenly the whole industry looks… easy. There’s a name for that phase: the Dunning-Kruger effect. Speed is exciting. But building a real company is still product, systems, distribution, operations, and years of scar tissue. The first version is easy. The second year is where things get interesting.

@martinfowler Code review was never about catching bugs. It's how seniors transfer judgment to the rest of the team. That part doesn't automate.

Fragments: supervisory engineering, 2 sets of 8 levels, generating code to an architecture, perils of detecting AI-aided writing, should humans review code? martinfowler.com/fragments/2026…