<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Agents on Moises Vega</title><link>https://mvega.dev/tags/agents/</link><description>Recent content in Agents on Moises Vega</description><generator>Hugo</generator><language>en-us</language><copyright>Moises Vega</copyright><lastBuildDate>Sat, 15 Aug 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://mvega.dev/tags/agents/index.xml" rel="self" type="application/rss+xml"/><item><title>Building Agentic Systems in Go with Ollama</title><link>https://mvega.dev/posts/go-agents-ollama/</link><pubDate>Sat, 15 Aug 2026 00:00:00 +0000</pubDate><guid>https://mvega.dev/posts/go-agents-ollama/</guid><description>&lt;p&gt;I worked through Joel Boursiquot&amp;rsquo;s GopherCon 2026 workshop, &lt;a href="https://github.com/jboursiquot/gc26buildingagenticsystems"&gt;Agentic
Systems the Hard Way&lt;/a&gt;.
Zero third-party dependencies. The framing that stuck: Go is the control
plane; the LLM is the decision engine. Here&amp;rsquo;s the core of it.&lt;/p&gt;
&lt;h2 id="what-an-agent-actually-is"&gt;What an agent actually is&lt;/h2&gt;
&lt;p&gt;Model + tools + loop. The model never executes anything — it emits a
wish (&amp;ldquo;call &lt;code&gt;search_docs&lt;/code&gt; with these args&amp;rdquo;), your Go code validates and
runs it, feeds the result back, and repeats until the model answers
without asking for tools. Deterministic Go around a nondeterministic
model. That&amp;rsquo;s the whole trick.&lt;/p&gt;</description></item></channel></rss>