New database splits 420,000 live podcasts by hosting type to surface independent shows

·1 views

A new podcasts collection on NicheDB categorizes live podcast feeds into two groups: those hosted on commercial platforms and those self-hosted on a creator's own domain. Analysis of a Podcast Index bulk dump from August 2026 found 421,928 live feeds, with 93.8% concentrated across just 307 domains — anchor.fm alone accounting for nearly 98,000. The remaining roughly 13,400 domains represent the self-hosted, independent half of the medium. An early technical flaw caused the commercial source to surface only 12 shows instead of hundreds of thousands, because recent directory entries skew heavily toward independent submissions rather than reflecting the full catalogue. After a fix that walks the entire listing before switching to incremental updates, the commercial source grew from 12 items to over 30,000 across two runs, with both halves now publicly browsable at nichedb.dev.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Geiger Tool Lets Users Monitor All AI Agents and Their System Access

A new open-source tool called Geiger has been released on GitHub by Atomburstofficial. The tool is designed to give users visibility into every AI agent running on their machine. It also shows what system resources, files, or data each agent is capable of accessing. This addresses growing concerns around transparency and control over locally running AI processes. The project was shared on Hacker News, where it attracted initial community attention.

0 comments Read more at Hacker News

ProgrammingDEV Community ·

Developers Build Zero-Dependency Java JSON Parser in 72-Hour Hackathon

A team participating in the ZeroDepsHack 2026 hackathon built JValue, a fully functional JSON parser and serializer in Java, using no third-party libraries whatsoever. The project relied solely on the Java standard library, including hand-written recursive-descent parsing, a custom test runner, and a 15-command CLI built without any external frameworks. Developers also implemented a JDK-only file fetcher to download test corpora, replacing common tools like curl and JUnit. The team documented 12 specific substitutions where third-party libraries were replaced by standard library equivalents, honestly noting the tradeoffs involved. The exercise was intended to test whether modern Java development could function without the dependency ecosystem most projects take for granted.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

AI Code Review Test Reveals Which Tools Actually Read Your Rules Files

A developer designed a controlled experiment to determine whether AI code review tools genuinely read project-specific rule files like CLAUDE.md and AGENTS.md, or simply rely on general best practices. Two tools, CodeRabbit and Qodo, were tested against a small Cloudflare Worker API that deliberately violated two rules: one widely-known security bad practice and one obscure internal naming convention. Both tools flagged the security violation, but that test proved little since any standard scanner would catch it regardless of file access. Only Qodo successfully identified the camelCase route naming violation — a purely internal convention with no external reference — and traced its functional impact on real requests. The experiment highlights a meaningful gap in AI code review reliability when enforcing custom, repo-specific conventions at scale.

0 comments Read more at DEV Community

ProgrammingDEV Community ·

Five Common Myths That Blur the Line Between AI Models and Actual Servers

A technical FAQ published on DEV Community identifies five widespread misconceptions that cause developers to conflate AI language model outputs with real server-side execution. The piece clarifies that a free model provides token prediction only, with no bundled compute or server infrastructure attached. It also explains that environment variables set inside a chat or markdown code block do not affect any running process on an actual machine. The author recommends a set of shell checklists — including hostname, pwd, and export verification commands — to confirm whether a task was genuinely executed on a server rather than merely described by the model. The article was prepared as part of outreach for MonkeyCode, a platform the author uses for free model access and server resources.

0 comments Read more at DEV Community