<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Soham Dasgupta on foojay.io - Friends of OpenJDK</title><link>https://foojayio.github.io/website/today/author/soham-dasgupta/</link><description>Articles written by Soham Dasgupta on foojay.io - Friends of OpenJDK</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Fri, 05 Jun 2026 08:35:05 +0000</lastBuildDate><atom:link href="https://foojayio.github.io/website/today/author/soham-dasgupta/index.xml" rel="self" type="application/rss+xml"/><item><title>Context Is Code: A Tour of APM and AgentRC</title><link>https://foojayio.github.io/website/today/context-is-code-a-tour-of-apm-and-agentrc/</link><pubDate>Fri, 05 Jun 2026 08:35:05 +0000</pubDate><guid>https://foojayio.github.io/website/today/context-is-code-a-tour-of-apm-and-agentrc/</guid><description>&lt;p&gt;If you&amp;rsquo;ve shipped an AI agent into a real codebase in the last twelve months, you&amp;rsquo;ve felt this: every agent, every developer, every machine &amp;mdash; different setup. A README that says &amp;ldquo;install these extensions.&amp;rdquo; A &lt;code&gt;copilot-instructions.md&lt;/code&gt; somebody copy-pasted from another repo. MCP server configs in three different files. The same skills duplicated for Copilot, Claude, Cursor, and Codex.&lt;/p&gt;
&lt;p&gt;No version pinning.&lt;/p&gt;
&lt;p&gt;No integrity.&lt;/p&gt;
&lt;p&gt;No reproducibility.&lt;/p&gt;
&lt;p&gt;That&amp;rsquo;s the problem &lt;a href="https://github.com/microsoft/apm" target="_blank" rel="noopener noreferrer"&gt;&lt;strong&gt;APM&lt;/strong&gt;&lt;/a&gt;
, the Agent Package Manager fixes. And it pairs with &lt;a href="https://github.com/microsoft/agentrc" target="_blank" rel="noopener noreferrer"&gt;&lt;strong&gt;AgentRC&lt;/strong&gt;&lt;/a&gt;
 to close the loop on &lt;em&gt;generating&lt;/em&gt; and &lt;em&gt;evaluating&lt;/em&gt; that context in the first place.&lt;/p&gt;</description></item><item><title>Context Is a Budget — Reducing Token Usage in AI-Assisted Development</title><link>https://foojayio.github.io/website/today/context-is-a-budget-eight-levers-and-three-workflow-patterns/</link><pubDate>Fri, 22 May 2026 12:52:06 +0000</pubDate><guid>https://foojayio.github.io/website/today/context-is-a-budget-eight-levers-and-three-workflow-patterns/</guid><description>&lt;p&gt;&lt;em&gt;Eight levers and three workflow patterns that pay for themselves in a week.&lt;/em&gt;&lt;/p&gt;
&lt;p&gt;A team of fifty developers can quietly burn $30,000 a month on AI coding assistants without anyone noticing. Premium-request quotas vanish by the third week. The bill arrives. Nobody has a story for where it went.&lt;/p&gt;
&lt;p&gt;The cost is the obvious pain. The other two are sneakier:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;&lt;strong&gt;Latency.&lt;/strong&gt; Bigger contexts take longer. The model thinks more, but you also wait more.&lt;/li&gt;
&lt;li&gt;&lt;strong&gt;Context rot.&lt;/strong&gt; This is the surprising one. Anthropic and Chroma have both shown that as the context window fills up, model recall and reasoning &lt;em&gt;degrade&lt;/em&gt; &amp;mdash; even well inside the advertised window. The 200K-token model is genuinely worse at the 150K mark than at the 20K mark. &lt;strong&gt;More context is not free; past a point, it&amp;rsquo;s actively harmful.&lt;/strong&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;The mental model that fixes all three: stop treating context as a free buffet. Treat it as a budget you spend on every turn.&lt;/p&gt;</description></item><item><title>Why is my Talk selected? Reflections from a Program Committee Reviewer</title><link>https://foojayio.github.io/website/today/why-is-my-talk-selected-reflections-from-a-program-committee-reviewer/</link><pubDate>Wed, 14 Jan 2026 15:14:56 +0000</pubDate><guid>https://foojayio.github.io/website/today/why-is-my-talk-selected-reflections-from-a-program-committee-reviewer/</guid><description>&lt;p&gt;If you are like me, get the adrenaline rush of getting up on stage and can&amp;rsquo;t wait to share things that you experienced and learned, then submitting talks to conferences and local meetups excites you.&lt;/p&gt;
&lt;p&gt;But as much as you are proud and happy when you are accepted to present, normally there is a lot of work before that for the reviewers and organizers to find the right presentation that resonates with the attendees and fits the whole program.&lt;/p&gt;</description></item><item><title>Foojay Podcast #49: JCON Report, Part 1 - JUGs, Communities, Generative AI, and more!</title><link>https://foojayio.github.io/website/today/foojay-podcast-49/</link><pubDate>Tue, 21 May 2024 06:05:39 +0000</pubDate><guid>https://foojayio.github.io/website/today/foojay-podcast-49/</guid><description>&lt;p&gt;On Tuesday, May 14th, the Foojay Podcast went live at the JCON conference in Cologne, Germany, to talk with speakers and visitors about all things Java.&lt;/p&gt;
&lt;p&gt;We had so many amazing talks that we will combine them into several podcast episodes in the next weeks.&lt;/p&gt;
&lt;p&gt;This is part 1 about JUGs, Communities, Open Source, Generative AI, LangChain4j, Machine Learning!&lt;/p&gt;
&lt;h2 id="h2-0-video"&gt;Video&lt;/h2&gt;
&lt;div style="position: relative; padding-bottom: 56.25%; height: 0; overflow: hidden;"&gt;
			&lt;iframe allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share; fullscreen" loading="eager" referrerpolicy="strict-origin-when-cross-origin" src="https://www.youtube.com/embed/-wVCYj8oQUY?autoplay=0&amp;amp;controls=1&amp;amp;end=0&amp;amp;loop=0&amp;amp;mute=0&amp;amp;start=0" style="position: absolute; top: 0; left: 0; width: 100%; height: 100%; border:0;" title="YouTube video"&gt;&lt;/iframe&gt;
		&lt;/div&gt;

&lt;h2 id="h2-1-podcast-apps"&gt;Podcast Apps&lt;/h2&gt;
&lt;p&gt;You can listen and subscribe to the Foojay Podcast on:&lt;/p&gt;</description></item><item><title>State of Open (Source?!) and Free AI</title><link>https://foojayio.github.io/website/today/state-of-open-source-and-free-ai-a-fosdem-recap/</link><pubDate>Thu, 15 Feb 2024 13:04:44 +0000</pubDate><guid>https://foojayio.github.io/website/today/state-of-open-source-and-free-ai-a-fosdem-recap/</guid><description>&lt;p&gt;&lt;em&gt;Disclaimer: This article is on the things I learned/observed spending the day in AI and Machine Learning Developer Room at FOSDEM 24. Opinions and statements are mine and have nothing to do with my employer. This article might raise more questions than answers, but in my opinion, we all need more awareness on this topic and get familiar with the (right) questions that are to be answered.&lt;/em&gt;&lt;/p&gt;
&lt;h3 id="h3-0-fosdem"&gt;FOSDEM&lt;/h3&gt;
&lt;p&gt;&lt;a href="https://fosdem.org/2024/" title="FOSDEM" target="_blank" rel="noopener noreferrer"&gt;FOSDEM&lt;/a&gt;
 (Free Open-Source Developers&amp;rsquo;European Meeting) is a community-organised event that is free and non-commercial. The aim is to provide a venue for free and open-source software developers and communities to:&lt;/p&gt;</description></item></channel></rss>