<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Blogs on Lucca Di Benedetto</title>
    <link>https://205edb25.luccadibenedetto.pages.dev/blog/</link>
    <description>Recent content in Blogs on Lucca Di Benedetto</description>
    <generator>Hugo</generator>
    <language>en</language>
    <lastBuildDate>Wed, 27 May 2026 12:00:00 +0200</lastBuildDate>
    <atom:link href="https://205edb25.luccadibenedetto.pages.dev/blog/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Why I refactored my benchmarking framework to be agent-first</title>
      <link>https://205edb25.luccadibenedetto.pages.dev/blog/benchctl-agent-first-refactor/</link>
      <pubDate>Wed, 27 May 2026 12:00:00 +0200</pubDate>
      <guid>https://205edb25.luccadibenedetto.pages.dev/blog/benchctl-agent-first-refactor/</guid>
      <description>&lt;p&gt;I just merged &lt;a href=&#34;https://github.com/luccadibe/benchctl/pull/3&#34; target=&#34;_blank&#34; rel=&#34;noopener noreferrer&#34;&gt;PR #3&lt;/a&gt; in &lt;a href=&#34;https://github.com/luccadibe/benchctl&#34; target=&#34;_blank&#34; rel=&#34;noopener noreferrer&#34;&gt;benchctl&lt;/a&gt;, a benchmark orchestration tool I have been developing on my own.&lt;/p&gt;&#xA;&lt;p&gt;I realised that developing a UI, plotting and data analysis features is not really a good focus for this project.&lt;/p&gt;&#xA;&lt;p&gt;Now, with AI agents, that type of code is pretty cheap to produce. One could even argue that the orchestration code is also quite cheap to produce. But there is a fundamental difference in both. The orchestration code is something that actually repeats itself / something that I reimplement every time I do a new serious benchmark. The rest (plots, data-analysis) tends to be very domain-specific and it&amp;rsquo;s very hard to add actually good, useful primitives there that I can guarantee will be used.&lt;/p&gt;</description>
    </item>
    <item>
      <title>How I optimized our HyperFaaS Docker Compose setup by 99.6%</title>
      <link>https://205edb25.luccadibenedetto.pages.dev/blog/hyperfaas-docker-compose/</link>
      <pubDate>Mon, 11 May 2026 18:00:00 +0200</pubDate>
      <guid>https://205edb25.luccadibenedetto.pages.dev/blog/hyperfaas-docker-compose/</guid>
      <description>&lt;p&gt;HyperFaaS’s Compose files model workers, leaves, and routing controllers at different scales. The layout that shipped before &lt;a href=&#34;https://github.com/3s-rg-codes/HyperFaaS/pull/62&#34; target=&#34;_blank&#34; rel=&#34;noopener noreferrer&#34;&gt;PR #62&lt;/a&gt; (&lt;a href=&#34;https://github.com/3s-rg-codes/HyperFaaS/commit/92f3b2a2ddb8be5abf23ba511dfa5bad52e51dc8&#34; target=&#34;_blank&#34; rel=&#34;noopener noreferrer&#34;&gt;&lt;code&gt;92f3b2a&lt;/code&gt;&lt;/a&gt;) was &lt;strong&gt;honestly terrible&lt;/strong&gt;: every replica that needed a custom image went through Compose’s &lt;code&gt;build:&lt;/code&gt; path with its own tag, so a “full refresh” meant &lt;strong&gt;many separate Docker builds of the same Go binaries&lt;/strong&gt; (for example eight worker images on the large profile, each a full compile), and the Dockerfile did not yet use BuildKit cache mounts for &lt;code&gt;go mod&lt;/code&gt; and &lt;code&gt;go build&lt;/code&gt;. It worked, but it was terribly slow.&#xA;We have three setups: small, medium and large. Each setup has their own YAML file with a different number of workers, leaves and routing controllers.&lt;/p&gt;</description>
    </item>
    <item>
      <title>Emoji Embeddings</title>
      <link>https://205edb25.luccadibenedetto.pages.dev/blog/emoji-embeddings/</link>
      <pubDate>Mon, 11 May 2026 17:04:29 +0200</pubDate>
      <guid>https://205edb25.luccadibenedetto.pages.dev/blog/emoji-embeddings/</guid>
      <description>&lt;h1 id=&#34;automatic-emoji-suggestions-for-a-tiny-private-lists-app&#34;&gt;Automatic Emoji Suggestions for a Tiny Private Lists App&lt;/h1&gt;&#xA;&lt;p&gt;I have a small lists webapp that I use privately with my partner. It is just where we keep recurring lists, shopping notes, reminders, and shared household things.&lt;/p&gt;&#xA;&lt;p&gt;At some point I noticed that lists feel nicer when items have emoji. &lt;code&gt;tomatoes&lt;/code&gt; looks better with 🍅 next to it, &lt;code&gt;coffee&lt;/code&gt; looks better with ☕, and &lt;code&gt;train tickets&lt;/code&gt; looks better with 🚆 or 🎟️, etc.&lt;/p&gt;</description>
    </item>
    <item>
      <title>Recommendations</title>
      <link>https://205edb25.luccadibenedetto.pages.dev/blog/recommendations/</link>
      <pubDate>Tue, 23 Dec 2025 17:47:32 +0100</pubDate>
      <guid>https://205edb25.luccadibenedetto.pages.dev/blog/recommendations/</guid>
      <description>&lt;p&gt;I realized that I never really wrote down a list of all the different creators / people in the tech scene that I follow, so this is it.&lt;/p&gt;&#xA;&lt;h2 id=&#34;people&#34;&gt;People&lt;/h2&gt;&#xA;&lt;ul&gt;&#xA;&lt;li&gt;&#xA;&lt;p&gt;&lt;a href=&#34;https://www.youtube.com/c/neetcode&#34; target=&#34;_blank&#34; rel=&#34;noopener noreferrer&#34;&gt;Neetcode&lt;/a&gt;. He is mostly known for his videos on leetcode problems, but he also has many great videos on other topics like AI and system design. He can explain leetcode solutions really well. Recommended if you are interested in learning about actual programming, data structures and algorithms.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
