Extensible Software in the age of LLMskept by eddie • aug 19Extensible web software can combine a stable core, LLM-driven extensions, and capability-based sandboxes to let users safely create and share personalized features.AboutCloudflare · DeepSeek · Salesforce · Kenton Varda · pi · ObsidianFiled#ai-agents#ai-coding#cloud-computing#software-engineering#web-developmentRelatedCodex-maxxing - Jason LiuThe author details how Codex's durable threads, voice input, steering, memory vaults, heartbeats, and goals transform ephemeral AI chats into persistent, self-running workstreams for coding, knowledge work, and even customer service.also on Obsidian, #ai-agents, #ai-codingDesigning for Agents (via @teddy_riker)Designing for AI agents requires abandoning traditional deterministic UI patterns in favor of interfaces that expose intent, state, and error-handling to autonomous systems.also on Obsidian, #ai-agents, #software-engineeringAn Update on HerokuHeroku is moving to a sustaining engineering model, prioritizing stability over new features, and ending new enterprise contracts.also on Salesforce, #cloud-computingThe Death of Software 2.0 (A Better Analogy!)AI agents like Claude Code will make human-oriented software obsolete, shifting value to persistent data infrastructure and APIs, analogous to a memory hierarchy.also on Salesforce, #ai-agents, #ai-coding, #software-engineeringThe expected value of showing upThe author experiments with one‑prompt AI agents to create three video games, learning prompt‑engineering lessons and showing that entering low‑participation challenges yields high expected value.also on pi, #ai-agents, #ai-coding, #software-engineering
Extensible Software in the age of LLMs