# Running Median

Running median is a technique for updating the median as each new numerical data point is encountered, efficiently implementable with two heaps.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## OpenRouter clocks Luna fastest, Jev second

DevFeed: [OpenRouter clocks Luna fastest, Jev second](<https://devfeed.tech/articles/openrouter-clocks-luna-fastest-jev-second-84130.md>)

Original publisher: [Read original article](<https://jevdaily.com/issues/2026-10-09>)

Author: Jev Daily

Published: 2026-10-09T17:00:00Z

Content type: news

Language: en

Sources: [Jev Daily](<https://devfeed.tech/sources/jev-daily.md>)

Topics: [Running Median](<https://devfeed.tech/topics/running-median.md>), [Calibrated uncertainty](<https://devfeed.tech/topics/calibrated-uncertainty.md>)

Tags: [decision-models](<https://devfeed.tech/tags/decision-models.md>), [evals](<https://devfeed.tech/tags/evals.md>), [langsmith](<https://devfeed.tech/tags/langsmith.md>), [latency](<https://devfeed.tech/tags/latency.md>), [openrouter](<https://devfeed.tech/tags/openrouter.md>), [someone-measured-it](<https://devfeed.tech/tags/someone-measured-it.md>), [typesafe-ai](<https://devfeed.tech/tags/typesafe-ai.md>), [writing-criteria](<https://devfeed.tech/tags/writing-criteria.md>)

### AI overview

OpenRouter's latency board ranks GPT-6 Luna first at 180 ms for global requests, with Jev behind it; Jev's latency number is not stated. Separately, Shrivu Shankar reports that Jev was more calibrated than 40 other decision models tested on email classification. Harrison Chase describes Jev's use as a LangSmith eval judge, where it gives difficulty and correctness separate typed answers for every trace. Another cited point is that agent trajectories can be assessed across several dimensions, including consideration of diverse design choices.

### Source excerpt

OpenRouter's latency board puts GPT-6 Luna on top at 180ms on global requests, with Jev behind it. Separately, Shrivu Shankar says Jev was the best calibrated...

## Diagnosing high p99 latency caused by Java garbage collection pauses

DevFeed: [Diagnosing high p99 latency caused by Java garbage collection pauses](<https://devfeed.tech/articles/your-p99-is-a-garbage-collector-setting-60525.md>)

Original publisher: [Read original article](<https://ankit-rana.com/logs/59-p99-is-a-garbage-collector-setting/>)

Author: hello@ankit-rana.com

Published: 2026-09-27T00:00:00Z

Content type: article

Language: en

Sources: [Ankit Rana | Mechanical Sympathy](<https://devfeed.tech/sources/ankit-rana-mechanical-sympathy.md>)

Topics: [Go Language](<https://devfeed.tech/topics/go-language.md>), [Running Median](<https://devfeed.tech/topics/running-median.md>), [proguard](<https://devfeed.tech/topics/proguard.md>)

Tags: [cost](<https://devfeed.tech/tags/cost.md>), [garbage-collection](<https://devfeed.tech/tags/garbage-collection.md>), [java](<https://devfeed.tech/tags/java.md>), [jvm](<https://devfeed.tech/tags/jvm.md>), [latency](<https://devfeed.tech/tags/latency.md>), [logs](<https://devfeed.tech/tags/logs.md>), [median](<https://devfeed.tech/tags/median.md>), [observability](<https://devfeed.tech/tags/observability.md>), [performance](<https://devfeed.tech/tags/performance.md>)

### AI overview

The article explains how stop-the-world garbage collection can cause p99 latency to rise even when application code paths are fast. It recommends comparing GC pause logs with request timestamps and explains how allocation rate, object retention, and heap size affect collection frequency and pause duration.

### Source excerpt

A p99 that is many times the median with no matching slow code path is usually a stop the world garbage collection pause, which stops every application thread regardless of what it was doing. The lever is allocation rate rather than heap size, because most collectors scale their work with the number of surviving objects and the frequency of collection follows directly from how fast you fill the young generation. Enlarging the heap reduces collection frequency and can lengthen each pause, which trades a common small delay for a rare large one.

## Rethinking UX for AI-driven alerting

DevFeed: [Rethinking UX for AI-driven alerting](<https://devfeed.tech/articles/rethinking-ux-for-ai-driven-alerting-83496.md>)

Original publisher: [Read original article](<https://www.datadoghq.com/blog/engineering/rethinking-ux-for-ai-driven-alerting/>)

Author: Steve Boak

Published: 2019-01-22T00:00:00Z

Content type: article

Language: en

Sources: [Datadog](<https://devfeed.tech/sources/datadog.md>)

Topics: [Monitoring & Alerting](<https://devfeed.tech/topics/monitoring-alerting.md>), [Physics-guided deep learning](<https://devfeed.tech/topics/physics-guided-deep-learning.md>), [outlier-detection](<https://devfeed.tech/topics/outlier-detection.md>), [Running Median](<https://devfeed.tech/topics/running-median.md>)

Tags: [2018](<https://devfeed.tech/tags/2018.md>), [alerting](<https://devfeed.tech/tags/alerting.md>), [alerts](<https://devfeed.tech/tags/alerts.md>), [anomaly-detection](<https://devfeed.tech/tags/anomaly-detection.md>), [capacity-planning](<https://devfeed.tech/tags/capacity-planning.md>), [critical](<https://devfeed.tech/tags/critical.md>), [ux](<https://devfeed.tech/tags/ux.md>), [ux-for-ai](<https://devfeed.tech/tags/ux-for-ai.md>)

### AI overview

The article describes how statistical alerting methods such as forecasting, anomaly detection, and outlier detection change monitoring UX. It contrasts configured alerts with algorithmic feeds that surface events proactively, and argues that useful feeds need clear explanations, context, and routing to the teams responsible.

### Source excerpt

The introduction of advanced statistical methods is reshaping the UX of alerts

## Regaining Lost Knowledge

DevFeed: [Regaining Lost Knowledge](<https://devfeed.tech/articles/regaining-lost-knowledge-21695.md>)

Original publisher: [Read original article](<https://rhettinger.wordpress.com/2010/02/06/lost-knowledge/>)

Author: rhettinger

Published: 2010-02-06T23:28:30Z

Content type: article

Language: en

Sources: [Raymond Hettinger](<https://devfeed.tech/sources/raymond-hettinger.md>)

Topics: [Running Median](<https://devfeed.tech/topics/running-median.md>), [Skiplists](<https://devfeed.tech/topics/skiplists.md>), [Algorithms](<https://devfeed.tech/topics/algorithms.md>)

Tags: [algorithm](<https://devfeed.tech/tags/algorithm.md>), [algorithms](<https://devfeed.tech/tags/algorithms.md>), [data-structures](<https://devfeed.tech/tags/data-structures.md>), [knowledge](<https://devfeed.tech/tags/knowledge.md>), [lost-knowledge](<https://devfeed.tech/tags/lost-knowledge.md>), [median](<https://devfeed.tech/tags/median.md>), [python](<https://devfeed.tech/tags/python.md>), [running-median](<https://devfeed.tech/tags/running-median.md>), [skiplists](<https://devfeed.tech/tags/skiplists.md>), [sliding-window](<https://devfeed.tech/tags/sliding-window.md>), [wikipedia](<https://devfeed.tech/tags/wikipedia.md>)

### AI overview

The excerpt presents an indexable skiplist as a portable O(log n)-time approach to maintaining a running median in a sliding window. Earlier replies based on the median's small movement and a QuickSelect recipe were described as unusably slow for large windows; a published method offered O(√n), while a tree-based solution offered O(log n) but was messy to implement and depended on the blist extension for a ready implementation. Adding link widths to a skiplist made median indexing logarithmic, according to the author, and the resulting approach scales to large windows and is portable across languages. The excerpt claims this application had not previously been presented, while acknowledging the passage is truncated.

### Source excerpt

Rediscovery of Indexable Skiplists in the solution of the Running Median problem.