Hacker News: thienannguyencv

New comment by thienannguyencv in "Ask HN: How do you offload all coding to AI?"

thienannguyencv — Wed, 25 Mar 2026 02:15:54 +0000

"Offloading all coding" is perhaps a misleading expression. Those who say they no longer write code are often describing a change in what kind of work they do, not that they've stopped writing code entirely. They spend more time on technical specification, architectural decisions, considering differences, and figuring out when the model misinterprets intent—and less time on actual code typing.

Your brownfield instinct is right though. The productivity gap between "fixing it yourself" and "require → plan → evaluate → deploy → evaluate" only narrows when the task is large enough to justify the cost incurred, or when you're running parallel agents. For a bug requiring only two lines of code, the cost of context switching alone can negate the return on investment (ROI).

New comment by thienannguyencv in "The more AI I used, the worse my code got"

thienannguyencv — Tue, 24 Mar 2026 16:18:26 +0000

The real loss isn't discipline—it's the feedback loop. When you write code by hand, the friction in the process forces you to internalize it. You sense when something is wrong before you can say why. AI eliminates that friction, so you read the result instead of writing it, and reading is always easier than it should be. The specification/architectural solution works, but it's an external constraint replacing an internal one. A more profound solution might be to proactively find ways to bring the friction back-writing tests before reading the AI's output (maybe with AI support), retyping critical sections (with AI or not), anything that forces the writer to engage in the loop.

New comment by thienannguyencv in "Our commitment to Windows quality"

thienannguyencv — Sat, 21 Mar 2026 11:06:43 +0000

The irony here is that their "commitment to quality" is essentially an admission they never delivered. And half of what they promised—like moving the taskbar—are features that already exist in Windows 10. They removed them, ignored complaints for years, and now they say they're "listening to feedback."

Similarly with "reducing unnecessary Copilot integration." They added it everywhere before, users hated it, and now removing it is considered a feature.

This isn't a commitment to quality. This is just a fix for years of treating the operating system as an advertising platform.

New comment by thienannguyencv in "I'm OK being left behind, thanks"

thienannguyencv — Sat, 21 Mar 2026 06:04:45 +0000

The query parameter issue is a pattern I see a lot. The model has thousands of examples of "how HTTP requests usually look" from its training process. When your input data conflicts with the pattern, the training data takes precedence.

Interestingly, the model doesn't "know" that it's ignoring you. From its perspective, it has retrieved a "meaningful" pattern—virtual parameter names that probably fit common conventions it saw during training. Your actual request simply... wasn't documented.

New comment by thienannguyencv in "A sufficiently detailed spec is code"

thienannguyencv — Sat, 21 Mar 2026 04:12:00 +0000

"regurgitate human intelligence from training data" - exactly. And the tricky part is when your actual code contradicts that training data.

Model saw thousands of examples of "how to implement X". When your codebase does X differently, training data wins. You can see it happen: point out the conflict, and a model that's reasoning would shift gears - ask questions, acknowledge tension. But model in retrieval mode just reiterates. Same confidence, same explanation, maybe rephrased.

That's why "I completely understand this time" keeps happening in AI responses. From model's view, nothing to check - the pattern it retrieved already "makes sense."

In short, if you're not doing something completely new, something that AI will almost certainly do better than you, then you're safe. Otherwise, you'll have to put in a considerable amount of effort to get the AI to cooperate, or you'll have to do the most difficult tasks yourself, simply because the limitations described above haven't been addressed.

New comment by thienannguyencv in "Be intentional about how AI changes your codebase"

thienannguyencv — Fri, 20 Mar 2026 15:59:41 +0000

This very matches my observation. The error isn't due to incorrect code—it's code that looks specific to your system but is actually generic patterns applied from the training process. The structure is correct, the logic is sound, it just doesn't interact with what your source code actually does.

Harder to catch because nothing is factually wrong. You have to ask: could this output have been produced without actually reading my codebase?

New comment by thienannguyencv in "Be intentional about how AI changes your codebase"

thienannguyencv — Fri, 20 Mar 2026 15:50:43 +0000

Yes, it scored 84% in GPTZero's AI test, but it was still "good enough" to pass HN's anti-AI test.

New comment by thienannguyencv in "Launch HN: Canary (YC W26) – AI QA that understands your code"

thienannguyencv — Fri, 20 Mar 2026 15:25:03 +0000

This benchmark measures whether tests are relevant, coherent, and have good coverage. But there's a more subtle type of error: AI creates tests that look specific to PR but are actually generic patterns mapped from the training data—correct test structure, reasonable assertions, but not actually interacting with what this specific piece of code does.

How do you differentiate between ""understood the code and generated a targeted test" and "recognized this looks like an auth flow and produced a standard auth test template"? The latter might still pass your coherence/relevance metrics while missing the actual exception.