thinking flesh & thinking machines

The Mula-Bhashya Problem: Teaching AI About Textual Authority
APR 13, 2026 // TECHNOLOGY

The Mula-Bhashya Problem: Teaching AI About Textual Authority

Every knowledge tradition has a hierarchy of texts. How to teach AI to distinguish between primary and secondary texts.

When AI Sounds Dharmic But Doesn't Think Dharmic
APR 13, 2026 // TECHNOLOGY

When AI Sounds Dharmic But Doesn't Think Dharmic

Can Small Language Models Exhibit Dharmic Coherence

APR 13, 2026 // TECHNOLOGY

Why Sanskrit Breaks Your Tokenizer (And What That Tells Us About Multilingual AI)

How to work with languages that have an inner structure.

SanskritKatha: 43 Reviewers and Counting
APR 5, 2026 // TECHNOLOGY

SanskritKatha: 43 Reviewers and Counting

Updates on the SanskritKatha project. 43 reviewers signed up. 45 reviews done.

Claw Claw Clawdius — Same Model, Different Harness, Different Personality
APR 1, 2026 // TECHNOLOGY

Claw Claw Clawdius — Same Model, Different Harness, Different Personality

The harness - the UI, the tools available, the interaction pattern - shapes the AI's effectiveness.

tweet-scheduler-mcp: Why I Built My Own Tweet Scheduler
MAR 31, 2026 // TECHNOLOGY

tweet-scheduler-mcp: Why I Built My Own Tweet Scheduler

Tweet scheduler MCP/CLI, integrates with ClaudeCode/Cursor and more. Don't waste money on Buffer-like tools.

Generating 50,000 Sanskrit Stories with AI: Building SanskritKatha in Public
MAR 31, 2026 // TECHNOLOGY

Generating 50,000 Sanskrit Stories with AI: Building SanskritKatha in Public

Generating original Sanskrit narratives in Devanagari, with age-appropriate vocabulary, coherent plots & ontology

Sanskrit Aid v0.2 - Sandhi Splitting, Pronunciation, and Vidyut-Powered Declensions
MAR 19, 2026 // TECHNOLOGY

Sanskrit Aid v0.2 - Sandhi Splitting, Pronunciation, and Vidyut-Powered Declensions

Three capabilities - Declension tables powered by Vidyut, two-tier sandhi/compound splitting, and audio pronunciation

I Let an AI Run 14 Experiments on My Sri Aurobindo Language Model — Here's What It Found
MAR 13, 2026 // TECHNOLOGY

I Let an AI Run 14 Experiments on My Sri Aurobindo Language Model — Here's What It Found

Counterintuitive ideas explored by Autoresearch method on my Sri Aurobindo Small Language Model

MAR 13, 2026 // TECHNOLOGY

First User Ran a Real Campaign with the Outbound Playbook MCP — Here's What Happened

Outbound Playbook MCP proves effective in real-world usage