Articles
In-depth research, comprehensive guides, and long-form thoughts on modern technology.
Import AI 468: RSI Ideas, PostTrainBench, and Trust in AI Racing
Import AI 468 covers 23 recursive self-improvement ideas, the PostTrainBench benchmark for autonomous LLM post-training, and the interplay of trust and transparency in AI racing.
AGI Is Not Multimodal: Why Embodiment Matters
Benjamin Spiegel argues that multimodal approaches to AGI will fail, advocating for embodiment and interaction as primary pathways to general intelligence.
After Orthogonality: Why AI Shouldn't Have Goals
This essay argues that rational AIs should not have goals but should instead align actions to eudaimonic practices, promoting excellence through excellence in a self-sustaining cycle.
Why Skepticism Haunts Mark Zuckerberg's AI Vision
Mark Zuckerberg's 6,500-word AI manifesto promises personal empowerment, but skepticism runs deep given Meta's history with social media and its current AI market position.
Open Blocks Could Standardize the Web’s Building Pieces
Joel Spolsky proposes an open protocol so that any block built once can run inside any editor that supports the standard.
Block Protocol Aims to Make Semantic Web Markup Effortless
Joel Spolsky’s Block Protocol lets developers create reusable blocks that carry structured data, starting with a free WordPress plugin.
Max Stoiber’s Path From Open Source Hits to OpenAI
Max Stoiber recounts building react-boilerplate and styled-components, the Spectrum and Stellate exits, and now shaping ChatGPT’s app platform.
DSPy Turns LLM Prompt Writing Into Programmable Code
Brett Kennedy explains how DSPy replaces brittle manual prompts with declarative signatures that compile and optimize for production LLM apps.
Using DSPy to Turn AI Evaluations into Better Chat Responses
How Dash uses DSPy to optimize LLM outputs by turning AI evaluations into a feedback loop for better responses.
Riviera: Universal Content Processing at Pinterest
How Pinterest's Riviera platform evolved to handle massive content ingestion and prepare data for AI/ML applications.
Content Ingestion and Podcast Video Incident Report
A detailed post-mortem of a podcast video processing incident, analyzing the root cause and the importance of incident reports.
Indexing the Data Lake for Online Point Queries
Exploring new strategies for indexing data lakes to enable fast, point lookups at scale, balancing storage costs with query performance.