<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>A Tinkerer&apos;s Canvas</title><description>Anuran&apos;s journal on the internet - capturing his late-night breakdowns to his Eureka moments.</description><link>https://anuran.dev/</link><item><title>Semantic Consensus and Semantic Drift</title><link>https://anuran.dev/blog/semantic-consensus/</link><guid isPermaLink="true">https://anuran.dev/blog/semantic-consensus/</guid><description>Contextual AI Agents are the present, transforming every workplace. But an AI agent is only as correct as the context as it references.</description><pubDate>Sat, 06 Jun 2026 00:00:00 GMT</pubDate><category>blog</category><category>ontology</category><category>semantic consensus</category><category>ai memory</category><category>context engineering</category></item><item><title>How to ACTUALLY build a &quot;Company Brain&quot;</title><link>https://anuran.dev/blog/build-a-company-brain/</link><guid isPermaLink="true">https://anuran.dev/blog/build-a-company-brain/</guid><description>Contextual AI Agents are the present, transforming every workplace. But an AI agent is only as correct as the context as it references.</description><pubDate>Fri, 15 May 2026 00:00:00 GMT</pubDate><category>blog</category><category>ontology</category><category>company brain</category><category>ai memory</category><category>context engineering</category></item><item><title>To self-host or not to self-host?</title><link>https://anuran.dev/blog/arise-self-hosting-for-technical-founders/</link><guid isPermaLink="true">https://anuran.dev/blog/arise-self-hosting-for-technical-founders/</guid><description>Startup credits for infrastructure are a lifesaver, but only if it does not entail endless maintenance. How do you decide, as an early-stage technical founder, what to host and what not to host?</description><pubDate>Mon, 19 Jan 2026 00:00:00 GMT</pubDate><category>blog</category><category>self-hosting</category><category>cost management</category><category>budgeting</category><category>startups</category></item><item><title>Flow classes in Pymetrix</title><link>https://anuran.dev/blog/pymetrix/pymetrix-flow/</link><guid isPermaLink="true">https://anuran.dev/blog/pymetrix/pymetrix-flow/</guid><description>A look at the Flow Class objects - FlowNode, FlowLayer, and FlowGraph - that form the building blocks of Pymetrix.</description><pubDate>Sat, 20 Nov 2021 00:00:00 GMT</pubDate><category>docs</category><category>python</category><category>analytics</category><category>pymetrix</category><category>foss</category></item><item><title>Pymetrix: The Open Source Plug and Play Python Analytics Library</title><link>https://anuran.dev/blog/pymetrix/pymetrix-home/</link><guid isPermaLink="true">https://anuran.dev/blog/pymetrix/pymetrix-home/</guid><description>An introduction to Pymetrix, a plug-and-play analytics library written in Python for monitoring application performance metrics.</description><pubDate>Sat, 20 Nov 2021 00:00:00 GMT</pubDate><category>docs</category><category>python</category><category>analytics</category><category>pymetrix</category><category>foss</category></item><item><title>Pymetrix: An Overview</title><link>https://anuran.dev/blog/pymetrix/pymetrix-overview/</link><guid isPermaLink="true">https://anuran.dev/blog/pymetrix/pymetrix-overview/</guid><description>An architectural overview of Pymetrix - the Flow Graph, the Database, and the extensible Plugins system.</description><pubDate>Sat, 20 Nov 2021 00:00:00 GMT</pubDate><category>docs</category><category>python</category><category>analytics</category><category>pymetrix</category><category>foss</category></item><item><title>Machine learning and the Pythonic buzz</title><link>https://anuran.dev/blog/machine-learning/</link><guid isPermaLink="true">https://anuran.dev/blog/machine-learning/</guid><description>A beginner-friendly introduction to machine learning with Python and scikit-learn, building a first model on the Boston housing dataset.</description><pubDate>Sun, 29 Aug 2021 00:00:00 GMT</pubDate><category>blog</category><category>python</category><category>machine-learning</category><category>scikit-learn</category><category>beginner</category><category>ai</category></item><item><title>Offline Speech Recognition with Vosk</title><link>https://anuran.dev/blog/offline-speech-with-vosk/</link><guid isPermaLink="true">https://anuran.dev/blog/offline-speech-with-vosk/</guid><description>Build an offline speech-to-command system in Python by combining Vosk for speech recognition with NLTK for keyword extraction.</description><pubDate>Sat, 24 Jul 2021 00:00:00 GMT</pubDate><category>blog</category><category>python</category><category>nlp</category><category>vosk</category><category>ai</category><category>nltk</category><category>foss</category></item><item><title>Is a Large Context Window all you need?</title><link>https://anuran.dev/papers/are-large-context-windows-all-you-need/</link><guid isPermaLink="true">https://anuran.dev/papers/are-large-context-windows-all-you-need/</guid><description>Exploring Time To First Token (TTFT)-context size tradeoff for Autoregressive LLMs</description><category>Deep Learning</category><category>Context Windows</category><category>Large Language Models</category><category>Context Engineering</category></item><item><title>Mitigating Hallucinations in LLMs</title><link>https://anuran.dev/papers/mitigating-hallucinations-in-llms/</link><guid isPermaLink="true">https://anuran.dev/papers/mitigating-hallucinations-in-llms/</guid><description>Mitigating Hallucinations in LLMs Using Sieve of Fallacies and Truths: A Game Theoretic Perspective</description><category>Deep Learning</category><category>Hallucinations</category><category>Game Theory</category><category>Large Language Models</category><category>Context Engineering</category></item><item><title>Oscillating Activation Functions</title><link>https://anuran.dev/papers/oscillating-activation-functions/</link><guid isPermaLink="true">https://anuran.dev/papers/oscillating-activation-functions/</guid><description>Oscillating activation functions and other recent progress in activation functions: A survey</description><category>Deep Learning</category><category>Activation Functions</category></item><item><title>SerpyTor</title><link>https://anuran.dev/papers/serpytor/</link><guid isPermaLink="true">https://anuran.dev/papers/serpytor/</guid><description>A distributed context-aware computational graph execution framework for durable execution</description><category>Distributed Computing</category><category>Durable Execution</category><category>Computational Graphs</category></item></channel></rss>