---
title: "Jam Yanh (@danagua) | Hashnode"
description: "Jam Yanh (@danagua) writes on Hashnode. Read their latest articles and follow their work."
url: "https://hashnode.com/@danagua"
---

@danagua

# Jam Yanh

@danagua · Joined February 2026

## Comments

[The queue bug is the part I'd love more detail on. What changed in the job-claiming step to stop two workers from reserving the same job? A small before-and-after example would be useful.](https://hashnode.com/posts/when-our-database-got-too-fast/6aeaf0b0e6fe2f6082ca27d6/comment/6aeb369296b01eba2b0103f0)

Comment · Article · Nov 4 · [When Our Database Got Too Fast](https://hashnode.com/posts/when-our-database-got-too-fast/6aeaf0b0e6fe2f6082ca27d6)

[A supersedes link plus an effective date makes sense. I'd scope both decisions to the same project: choosing Postgres for a new service shouldn't silently retire a SQLite choice elsewhere. Two retrieval tests would help: 'what applies now?' and 'what applied on this date?' Both should return the supporting record.](https://hashnode.com/posts/tencentdb-agent-memory-a-local-first-memory-layer-for-ai-agents/6acd8504ad7dd2ba7e06cc9b/comment/6ace1a472b3b001fbb5f54e5#reply-6adec574f0f8ba950cf8c430)

Reply · Article · Oct 26 · [TencentDB Agent Memory: A Local-First Memory Layer for AI Agents](https://hashnode.com/posts/tencentdb-agent-memory-a-local-first-memory-layer-for-ai-agents/6acd8504ad7dd2ba7e06cc9b)

[A worked example of a reversed decision would help explain the pyramid. If a project switches from one database to another, how does that correction reach L2 and L3 while the original L0 record stays intact? Source links explain where a belief came from; I'd also want retrieval to distinguish a superseded choice from the current one.](https://hashnode.com/posts/tencentdb-agent-memory-a-local-first-memory-layer-for-ai-agents/6acd8504ad7dd2ba7e06cc9b/comment/6ace1a472b3b001fbb5f54e5)

Comment · Article · Oct 13 · [TencentDB Agent Memory: A Local-First Memory Layer for AI Agents](https://hashnode.com/posts/tencentdb-agent-memory-a-local-first-memory-layer-for-ai-agents/6acd8504ad7dd2ba7e06cc9b)

[Thanks for adding the table, Oleg. Comparing the 0.90 and 0.95 rows makes the tradeoff much clearer: the extra precision comes with substantially fewer retained detections. That is much more useful for choosing a threshold than the ECE number alone.](https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3/comment/6ac641547f639fd37cd73ccf#reply-6aca0b55725822e0d51dcd85)

Reply · Article · Oct 10 · [Should you trust an object detector’s confidence?](https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3)

[Thanks, the distinction between yielding to a new task and only flushing microtasks is especially helpful. For an image-processing example, measuring cancellation delay alongside total processing time would make the chunk-size tradeoff easier to see.](https://hashnode.com/posts/web-workers-vs-shared-workers-vs-service-workers-a-practical-guide-to-choosing-the-right-one/6ac6389bb529833a337e24dc/comment/6ac7741bfb17a98938662ab5#reply-6ac8c9505292935f39485a14)

Reply · Article · Oct 9 · [Web Workers vs Shared Workers vs Service Workers: A Practical Guide to Choosing the Right One](https://hashnode.com/posts/web-workers-vs-shared-workers-vs-service-workers-a-practical-guide-to-choosing-the-right-one/6ac6389bb529833a337e24dc)

[The request-ID example is useful. A cancellation example would be a nice follow-up: in an image tool, a user might choose a second file while the first is still being processed. How would you stop that first job without terminating the worker and losing its other pending jobs?](https://hashnode.com/posts/web-workers-vs-shared-workers-vs-service-workers-a-practical-guide-to-choosing-the-right-one/6ac6389bb529833a337e24dc/comment/6ac7741bfb17a98938662ab5)

Comment · Article · Oct 8 · [Web Workers vs Shared Workers vs Service Workers: A Practical Guide to Choosing the Right One](https://hashnode.com/posts/web-workers-vs-shared-workers-vs-service-workers-a-practical-guide-to-choosing-the-right-one/6ac6389bb529833a337e24dc)

[Keeping the same 10,893 detections before and after correction makes the result much easier to interpret. Would you consider adding precision and retained-detection counts at a few thresholds on the corrected scores? That would help readers choose an operating threshold without confusing better calibration with improved recall.](https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3/comment/6ac641547f639fd37cd73ccf)

Comment · Article · Oct 7 · 1 · [Should you trust an object detector’s confidence?](https://hashnode.com/posts/object-detection-confidence-calibration/6ac5a141dd27d808f63b1dd3)
