Daily Feed — 2026-04-13

This content is AI-generated by my RSS reader tool. Summaries and novelty ratings should be taken with a pinch of salt.

No One Has the Right to Obey

Source: The Adventures of Blake | Tags: freedom, hannah-arendt, workshop | Published: 2026-04-12 | Novelty: 44%

The article discusses a workshop titled 'How to Be Free' created by Blake, focusing on the concepts of physical, social, and political freedom. The workshop explores these themes through activities inspired by partner dance and authentic relating, addressing the author's qualifications to discuss freedom in a context where many people experience a high degree of it. Key references include Alexei Navalny’s struggle for political freedom and Hannah Arendt’s philosophy on thinking for oneself.


Gemma 4 audio with MLX

Source: Simon Willison's Weblog | Tags: audio, gemma-4-e2b, ml, mlx-vlm, transcription | Published: 2026-04-12 | Novelty: 43%

The article provides a specific uv run command for transcribing an audio file using the Gemma 4 E2B model with MLX and mlx-vlm on macOS, highlighting the potential misinterpretation of certain words. The example uses a 14-second .wav file to demonstrate the process.


SQLite 3.53.0

Source: Simon Willison's Weblog | Tags: cli, constraints, database, json, sqlite | Published: 2026-04-11 | Novelty: 35%

SQLite 3.53.0 includes significant user-facing improvements such as the ability to add and remove NOT NULL and CHECK constraints using ALTER TABLE, a new json_array_insert() function with its jsonb equivalent, and enhanced CLI mode result formatting supported by a new Query Results Formatter library. This version also features a WebAssembly playground interface for experimenting with these enhancements.


How I Made $360 by Serving Markdown to A.I. agents from My Jekyll Blog

Source: code.dblock.org | tech blog | Tags: ai, markdown, monarch | Published: 2026-04-11 | Novelty: 34%

The author introduced a feature on their Jekyll blog where each post has an associated .md version available at the same URL, catering to AI agents. This simple change led to unexpected results, including 18 signups and 12 paid conversions worth $360 from Monarch's referral link in just thirty days during the free trial period.


Agentic Engine Optimization (AEO)

Source: AddyOsmani.com | Tags: agent, ai, documentation, llms, seo | Published: 2026-04-11 | Novelty: 33%

The article introduces AGENT.md as the new default entry point for AI agents and emphasizes that documentation should serve both human and agent audiences. It details a checklist of actions such as adding llms.txt, measuring token counts, writing skill.md files, and setting up 'Copy for AI' buttons to make documentation more accessible to AI agents. The author also provides an audit tool called agentic-seo to help with these tasks.


Quoting Bryan Cantrill

Source: Simon Willison's Weblog | Tags: efficiency, human-laziness, llms | Published: 2026-04-13 | Novelty: 30%

Bryan Cantrill argues that LLMs lack human laziness, which forces humans to create efficient and concise abstractions due to time constraints. He suggests this inherent human trait is crucial for developing effective systems, contrasting with the potentially overwhelming and less optimal outputs of LLMs.


SQLite Query Result Formatter Demo

Source: Simon Willison's Weblog | Tags: sqlite, tools, webassembly | Published: 2026-04-11 | Novelty: 28%

The article introduces a web-based tool called SQLite Query Result Formatter Demo, which allows users to experiment with different rendering options for SQL result tables using the Query Result Formatter library compiled to WebAssembly. This tool is particularly useful for exploring new features in SQLite version 3.53.0 as detailed by Simon Willison.


The peril of laziness lost

Source: The Observation Deck | Tags: abstraction, laziness, llm, programming, software | Published: 2026-04-12 | Novelty: 18%

The article discusses how laziness is a virtue for software developers as it drives the creation of powerful abstractions that simplify complex systems. It contrasts this with the 'brogrammer' culture, where there's an emphasis on false industriousness. The author argues that language models (LLMs) lack human constraints and can produce overly complex systems without regard to future maintenance, highlighting how laziness is essential for developing crisp abstractions within time and cognitive load limits.