<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0"><channel>
<title>Smeet Dhakecha</title><link>https://dhakecha.com</link><description>Research Engineer building agentic AI systems.</description>
<item><title>Why is your inference bill so high?</title><link>https://dhakecha.com/blog/why-is-your-inference-bill-so-high/</link><guid>https://dhakecha.com/blog/why-is-your-inference-bill-so-high/</guid><pubDate>Thu, 05 Nov 2026 00:00:00 GMT</pubDate><description>Five places where enterprise agent spend leaks, and fixes that cost far less than asking for more capacity.</description></item>
<item><title>Journey from Capital to Compute</title><link>https://dhakecha.com/blog/journey-from-capital-to-compute/</link><guid>https://dhakecha.com/blog/journey-from-capital-to-compute/</guid><pubDate>Mon, 28 Sep 2026 00:00:00 GMT</pubDate><description>The hard part of getting GPUs isn't scarcity, it's paying for them. How GPU financing works and what the power market can teach it.</description></item>
<item><title>Is your AI ready for real work?</title><link>https://dhakecha.com/blog/is-your-ai-ready-for-real-work/</link><guid>https://dhakecha.com/blog/is-your-ai-ready-for-real-work/</guid><pubDate>Fri, 25 Sep 2026 00:00:00 GMT</pubDate><description>How to find out if an AI agent makes a promise your business can't keep.</description></item>
<item><title>Quitting your job isn’t enough....</title><link>https://dhakecha.com/blog/quitting-your-job-isnt-enough/</link><guid>https://dhakecha.com/blog/quitting-your-job-isnt-enough/</guid><pubDate>Sat, 12 Sep 2026 00:00:00 GMT</pubDate><description>AI takeover warnings surface the problem, but they don't advance AI safety. What an actionable agenda for recursive self-improvement could look like.</description></item>
<item><title>What is Agent Evaluation and Why You Need It?</title><link>https://dhakecha.com/blog/what-is-agent-evaluation/</link><guid>https://dhakecha.com/blog/what-is-agent-evaluation/</guid><pubDate>Sun, 11 Jan 2026 00:00:00 GMT</pubDate><description>Most teams evaluate agents like they evaluate LLMs. Here's why that breaks, and what to measure instead.</description></item>
<item><title>What I Learned from How Anthropic's Team Actually Uses Claude Code</title><link>https://dhakecha.com/blog/how-anthropic-uses-claude-code/</link><guid>https://dhakecha.com/blog/how-anthropic-uses-claude-code/</guid><pubDate>Wed, 07 Jan 2026 00:00:00 GMT</pubDate><description>Boris's Claude Code setup, and my own takeaways from stress-testing it on real enterprise workflows.</description></item>
<item><title>Building Agentic AI Systems</title><link>https://dhakecha.com/blog/building-agentic-ai-systems/</link><guid>https://dhakecha.com/blog/building-agentic-ai-systems/</guid><pubDate>Fri, 28 Nov 2025 00:00:00 GMT</pubDate><description>Reflections on designing simulation frameworks for validating AI agent behaviors in complex environments.</description></item>
</channel></rss>
