Skip to content

GLN-7.5AI Native

← Back to the front page

AI Agents

Everything about AI agents in one place: agentic coding, AI-native workflows and the road toward an AI operating system.

38 stories

Eve Bouffard, Y Combinator's Head of Design, in the Design Review interview about designing with AI agents.

AI in DesignHow YC's Head of Design Builds with AI AgentsYC's Head of Design Eve Bouffard on designing with AI: talk to coding agents, one-shot many drafts, and build throwaway tuning tools.Jul 17, 2026 · 15 min read

The Codex Micro, a compact keyboard with an analog stick and a dial for steering OpenAI's Codex coding agent.

CodexOpenAI's Codex Micro: A Keyboard to Steer the AI AgentOpenAI's Codex Micro is a $230 keyboard, built with Work Louder, that controls the Codex coding agent with voice, a dial, and mappable keys.Jul 17, 2026 · 4 min read

Daisy Hollman and the title Beyond the basics with Claude Code from a Code w/ Claude presentation.

Claude CodeWhat Claude Code needs to work with youDaisy Hollman shows Claude Code is not just coding help, but access, knowledge, tools and an organized work environment around the AI agent.Jun 2, 2026 · 23 min read

Thumbnail from the Y Combinator podcast Inside YC's AI Playbook with Pete Koomen on how YC built its internal agent infrastructure

AI AgentsHow to Build Superintelligence Inside Your CompanyPete Koomen at Y Combinator explains how AI can become the operating system of a company, not just an add-on.May 27, 2026 · 15 min read

Illustration of one operator running four AI agents for purchasing, customer service, logistics and accounting

AI AgentsHow one person can run an IT company with AI agentsA story about a fictional IT company where one operator runs orders, purchasing, support and accounting with small, controlled AI agents.May 25, 2026 · 21 min read

AI Agent Operating System illustrated as a control center with agents, memory, access, and tools

AI AgentsWhy AI agents need their own operating systemWhy AI agents need memory, tool management, identity, logging, security rules, and a control layer around them.May 16, 2026 · 17 min read

Claude Code robot giving thumbs up in front of a laptop showing an accounting dashboard, with the title How I'm helping Glenn build a Norwegian AI accounting system from scratch

Claude CodeHow Claude Code Builds a Norwegian Accounting SystemClaude Code tells the story of how Glenn builds a Norwegian accounting system from scratch, and what sets an AI agent apart from a chatbot.Apr 28, 2026 · 12 min read

Cedric Clyburn from IBM Technology explaining the difference between AI chatbots and AI agents on a whiteboard

AI AgentsHow AI Agents Work: Inside OpenClaw and the Agentic LoopHow AI agents move from telling you what to do to actually doing it, walked through OpenClaw's architecture and the agentic loop step by step.Apr 27, 2026 · 5 min read

IBM Technology video explaining how to build a team of AI agents with different roles and responsibilities

AI AgentsBuilding a Team of AI AgentsComplex tasks need teams, even for AI. IBM maps seven agent roles and four ways to make each one effective.Apr 25, 2026 · 4 min read

Martin Keen from IBM Technology explaining AI agent skills on a whiteboard

AI AgentsAgent Skills: Teaching AI Agents How to Actually WorkAI agents know facts but not procedures. IBM's Martin Keen explains how a skill.md file gives agents the procedural knowledge real work demands.Apr 20, 2026 · 5 min read

Bri Kopecki explaining the seven skills an agent engineer needs, on an IBM Technology video

AI AgentsFrom Prompt Engineer to Agent Engineer: Seven SkillsIBM's Bri Kopecki says prompt engineering is the recipe. Agent engineering is being the chef. Here are the seven skills you need.Apr 17, 2026 · 8 min read

LG Uplus and OpenAI engineers collaborating on Agentic AICC

AI AgentsLG Uplus and OpenAI Build AI That Handles Customer CallsLG Uplus built an AI contact center with OpenAI that doesn't just answer questions. It checks accounts, changes plans, and resolves issues in real time.Apr 6, 2026 · 4 min read

Dan Disler at a laptop with text 'AGENT HARNESS' visible on screen

AI AgentsThe Agent Harness Is the ProductThe Claude Code leak revealed that the most valuable part of an AI coding tool is not the model. It is the harness around it.Apr 6, 2026 · 9 min read

Grant Miller from IBM explaining how to secure the trust chain in agentic AI systems using tokens and delegation.

AI AgentsHow to Secure the Trust Chain in Agentic AIAI agents are getting access to real tools. IBM's Grant Miller maps out four risks in the trust chain and the token architecture that stops each one.Apr 5, 2026 · 5 min read

Nate B Jones explaining the Claude Code architecture leak on his YouTube channel AI News & Strategy Daily.

Claude CodeClaude Code Leaked: The Secret Is Not the AIAnthropic accidentally leaked Claude Code's source. Everyone looked at upcoming features. The real story is 12 boring primitives most teams skip.Apr 3, 2026 · 6 min read

OpenAI founder spotlight for Gradient Labs

AI AgentsGradient Labs Gives Banks an AI Account ManagerGradient Labs says banks can give every customer an AI account manager by combining OpenAI models with strict procedures, low latency, and guardrails.Apr 2, 2026 · 5 min read

NVIDIA Developer session on agentic AI at GTC 2026

AI AgentsNvidia's Practical Guide to Building AI AgentsNvidia's GTC session breaks AI agents into practical pieces: what changed, how they work, and how to build one without losing control.Apr 2, 2026 · 7 min read

IndyDevDan's multi-agent chat interface showing three teams of specialized AI agents working together

AI AgentsWhy One Agent Is Not EnoughDan Disler demos a 3-tier multi-agent system: 9 specialized agents across planning, engineering, and validation. He explains why one agent is a ceiling.Mar 30, 2026 · 5 min read

Bri Kopecki presenting AgentOps in front of a whiteboard with the three-layer framework.

AI AgentsAgentOps: Three Layers Your AI Agents NeedIBM's Bri Kopecki explains AgentOps: a three-layer framework for observing, evaluating, and improving AI agents once they're live in production.Mar 30, 2026 · 5 min read

Tim Ruscica's Claude Code beginner tutorial showing plugins and MCP setup

Claude CodeHow to Make Claude Code Better: A Beginner's GuideFive beginner changes that make Claude Code far more useful: a memory file, three plugins, the GitHub MCP server, and commands most users never find.Mar 29, 2026 · 9 min read

YouTube summary: Top 10 Security Risks in AI Agents Explained

AI Agents10 Ways AI Agents Can Go WrongIBM's Jeff Crume walks through OWASP's Top 10 security risks for AI agents: goal hijacking, memory poisoning, supply chain attacks, and rogue agents.Mar 24, 2026 · 6 min read

Tyler Lynch from IBM explains agentic security runtime and how to secure AI agents

AI SecurityHow to Secure AI Agents at RuntimeIBM's Tyler Lynch explains three runtime security layers for AI agents: dynamic credentials, user identity via OAuth 2.0, and CIBA push notifications.Mar 22, 2026 · 4 min read

Andrej Karpathy with 'SKILL ISSUE' text overlay and AI tool logos including Claude, OpenAI, and Codex

AI AgentsKarpathy: Everything Is a Skill Issue NowKarpathy has not typed code since December 2025. He explains why the bottleneck is now you, not the AI, and what that means for work.Mar 20, 2026 · 8 min read

Diagram showing the autoresearch loop: modify code, train 5 minutes, evaluate result, keep or discard. Results panel shows 700 experiments, 20 improvements, 11 percent faster training

AI AgentsKarpathy's Autoresearch: AI That Runs Its Own ExperimentsKarpathy left an AI agent running overnight. It ran 700 experiments, found bugs he missed, and made his code 11% faster. Here is how autoresearch works.Mar 18, 2026 · 8 min read

IBM Technology video on agentic runtime security and non-human identities

AI SecurityThe AI Agents Most Companies Forget to SecureAI agents multiply the identity problem. IBM explains four security gaps, five imperatives, and why the real bottleneck is organizational, not technical.Mar 15, 2026 · 5 min read

Swami Sivasubramanian presenting on stage at TEDAI Vienna about AI agents

AI AgentsAWS VP: Three Milestones AI Agents Must Hit FirstAWS VP Swami Sivasubramanian lays out three milestones AI agents must reach before they change everything, from trust to access for non-coders.Mar 15, 2026 · 6 min read

Martin Keen at whiteboard explaining hierarchical AI agent architecture with three tiers

AI AgentsWhy One AI Agent Is Not EnoughHierarchical AI agents fix context dilution and tool overload, but they trade those problems for orchestration complexity.Mar 14, 2026 · 5 min read

Anna Gutowska presenting why AI agents need human-in-the-loop oversight

AI AgentsThe Case for Human-in-the-Loop AI AgentsIBM's Anna Gutowska argues that AI agents in production need human oversight baked into the architecture, not added as an afterthought.Mar 10, 2026 · 6 min read

Panel discussion about OpenClaw autonomous AI agents on the Moonshots podcast

AI AgentsOne Person, Five AI Agents: An Autonomous OrganizationAlex Finn runs five AI agents like employees on Mac Studios. His org chart model hints at a future where one person manages an autonomous workforce.Mar 9, 2026 · 10 min read

YouTube summary: 4 Ways AI Agents Should Behave for Smarter Systems

AI SecurityWhy AI Agents Need Limits, Not SuperpowersIBM's risk-capability framework for AI agents: four behaviors that prevent over-powered agents from going rogue.Mar 8, 2026 · 5 min read

Martin Keen from IBM Technology explaining agentic storage on a lightboard

Why AI Agents Need Their Own File SystemAI agents lose their memory after every session. Agentic storage gives them persistent, safety-aware file systems through MCP.Mar 8, 2026 · 5 min read

IndyDevDan demonstrating a 4-layer browser automation stack with Claude Code and Playwright CLI

Claude CodeBuild a 4-Layer AI Browser Automation StackFrom Playwright CLI skills to parallel subagents. A four-layer architecture for automating browsers and UI testing with Claude Code.Mar 6, 2026 · 6 min read

Boris Cherny discussing how he builds Claude Code on The Pragmatic Engineer podcast

Claude CodeHow Claude Code's Creator Ships 30 PRs a DayBoris Cherny ships 20-30 PRs daily with zero handwritten code. His workflow offers a glimpse of how software engineering is changing.Mar 4, 2026 · 10 min read

IBM Technology video explaining A2A and MCP protocols for AI agent communication

MCPA2A and MCP: The Two Protocols AI Agents NeedIBM explains A2A and MCP, two open protocols that let AI agents talk to each other and connect to external tools and data.Mar 3, 2026 · 4 min read

IBM Technology video thumbnail on securing AI agents with best practices for safety

AI SecurityHow to Secure AI Agents: IBM and Anthropic's GuideIBM and Anthropic's framework for securing AI agents. Covers seven threat categories, DevSecOps lifecycle, identity management, and AI firewalls.Mar 3, 2026 · 10 min read

IndyDevDan video thumbnail: I Studied Stripe's AI Agents — Vibe Coding Is Already Dead

Stripe's AI Agents Write All the Code. Here's How.Stripe ships 1,300+ pull requests per week with zero human-written code. Here's the six-layer architecture that makes it possible.Mar 2, 2026 · 9 min read

How AI Agents and Decision Agents Combine Rules & ML in Automation — YouTube thumbnail

AI AgentsWhy AI Agents Need More Than Just LLMsHow to combine LLMs with workflow, business rules, and ML into multi-method agentic AI. A bank loan example with 9 agent types.Mar 2, 2026 · 8 min read

Peter Steinberger on OpenAI's Builders Unscripted series, discussing the creation of OpenClaw

OpenClaw90,000 Commits, One Developer: The OpenClaw StoryPeter Steinberger contributed to 120+ projects in one year using AI agents alone, including the viral OpenClaw. He shares his workflow and philosophy.Mar 1, 2026 · 10 min read