<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Posts on David Bensoussan</title><link>https://blog.bensoussan.de/post/</link><description>Recent content in Posts on David Bensoussan</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Wed, 05 Aug 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://blog.bensoussan.de/post/index.xml" rel="self" type="application/rss+xml"/><item><title>Six Checks Against AI Slop</title><link>https://blog.bensoussan.de/post/2026-08-05-six-checks-against-ai-slop/</link><pubDate>Wed, 05 Aug 2026 00:00:00 +0000</pubDate><guid>https://blog.bensoussan.de/post/2026-08-05-six-checks-against-ai-slop/</guid><description>&lt;p&gt;&lt;strong&gt;&lt;a href="https://github.com/immich-app/immich/pull/30514"&gt;feat(transcription): automatic video subtitle generation via faster-whisper #30514&lt;/a&gt;&lt;/strong&gt; — closed · 17 commits · 3,035 tests passing · 0 lint errors · &lt;code&gt;auto-closed:llm&lt;/code&gt;&lt;/p&gt;
&lt;p&gt;My AI-assisted pull request to &lt;a href="https://github.com/immich-app/immich"&gt;immich&lt;/a&gt;, an open-source project, was closed within minutes today because an LLM was involved. Their policy forbids it:&lt;/p&gt;
&lt;blockquote&gt;
&lt;p&gt;We ask you not to open PRs generated with an LLM. We find that code generated like this tends to need a large amount of back-and-forth, which is a very inefficient use of our time.
We cannot have any confidence in an LLM, so if it&amp;rsquo;s non-trivial for us to verify it works and you don&amp;rsquo;t show that you fully understand all implications of the change, reviewing the PR is not worth our time.&lt;/p&gt;</description></item><item><title>Zero Overhead</title><link>https://blog.bensoussan.de/post/2026-08-04-zero-overhead/</link><pubDate>Tue, 04 Aug 2026 00:00:00 +0000</pubDate><guid>https://blog.bensoussan.de/post/2026-08-04-zero-overhead/</guid><description>&lt;p&gt;I&amp;rsquo;ve never kept a blog. Every time I considered it, the overhead argument
won: a platform to maintain, a hosting bill, and posts that would have to be
written from scratch.
Not worth it.&lt;/p&gt;
&lt;p&gt;Two things changed this year, so the math finally works.&lt;/p&gt;
&lt;h2 id="who-i-am"&gt;Who I am&lt;/h2&gt;
&lt;p&gt;I&amp;rsquo;m &lt;a href="https://resume.bensoussan.de"&gt;David Bensoussan&lt;/a&gt;, a software engineer.
Most of what I build outside work lives in one monorepo: the apps, the
infrastructure that runs them, and the CI in between. The infrastructure part
is a homelab — two second-hand Debian machines running about 45 services:
photos, media, documents, monitoring, an LLM server on an old GPU, and the CI
runners themselves. The full topology and progress log is at
&lt;a href="https://homelab.bensoussan.de"&gt;homelab.bensoussan.de&lt;/a&gt; if you want the
inventory. Everything is provisioned with Ansible, and every significant
decision gets written down.&lt;/p&gt;</description></item></channel></rss>