Tag
A discussion on where AI agents fail in real workflows, highlighting issues with coordination, reliability under messy inputs, and the challenge of reducing human intervention in production.
A detailed guide on using OpenAI's Codex as an operating system for knowledge work, including setup, workflows, and a seven-day starter plan, written using Codex itself.
OpenAI released a library of ready-to-use prompts for Codex, featuring project workflows and automations that can be adapted for other AI coding agents.
The article argues that while many are building and selling AI agents, the real value lies in the workflows and training that make them useful, not the underlying technology.
A personal take that AI agents feel genuinely futuristic because they can autonomously plan steps, use tools, and recover from errors, marking a shift from AI that only answers questions to AI that actually does things.
Claude Code is releasing a /workflows feature that turns standard business work into pseudo-deterministic workflows based on SOPs, which the author sees as a significant step for Enterprise AI.
Boris Cherny, creator of Claude Code, argues that single-agent workflows are obsolete and explains the future is teams of specialized agents working together.
The author describes Briefing Fox, a prompt engineering tool that now includes an Agentic AI & Workflow feature, allowing even first-time AI users to create custom AI agents in minutes. It is free and requires no signup.
Matt Pocock published his daily-used Claude Code skills for real engineering, with over 90,000 stars and MIT license.
Anthropic quietly shipped /workflows in Claude Code, replacing LLM orchestrators with code-based control flow to avoid token tax and context sloppiness.
Boris Cherny, creator of Claude Code at Anthropic, explains on a podcast why some users save 10 hours a week while others save none, highlighting hidden features like CLAUDE.md, task chaining, batch operations, and automations that most users overlook.
A security research reveals a technique named Megalodon for mass backdooring of GitHub repositories by exploiting CI workflows.
A tweet announces an Agent Mode version of a report for Every subscribers, containing the full essay, sources, references, and workflows, designed to be used with Codex or Claude Codex.
Pao announced the launch of Handinger, a managed cloud agent for automating business tasks such as email workflows, reporting, and data analysis.
The author reflects on the current limitations of AI agents for complex, long-running tasks, citing reliability issues and suggesting that agents are better suited for narrow, supervised tasks rather than full autonomy.
The framing that AI is only for coders has become outdated; many of the most interesting AI workflows are now run by non-programmers.
An exploration of how using multiple AI models for agent workflows reveals hidden uncertainties and reasoning gaps, suggesting that future systems may rely on cross-model consensus rather than single-model chains.
SaaS-Bench is a new benchmark built on 23 deployable SaaS systems across six professional domains, containing 106 long-horizon tasks for evaluating computer-using agents. Experiments show that even the strongest models complete fewer than 4% of tasks end-to-end, highlighting significant limitations in current agent capabilities.
A founder shares 19 reusable skill instructions for AI agents (Claude/Claude Code) to automate early-stage startup functions like positioning, pricing, prospecting, and copywriting, based on their own SOPs.
Mintlify introduces self-updating knowledge bases through workflows, automating documentation maintenance.