AGENTS · 62 stories

Agents.

Agentic systems, multi-agent architectures, and the infrastructure enabling autonomous AI to act in the world.

62

Stories in section

Time:

Sort

A new harness lets agents keep learning ARC-AGI-3 on the fly — featured image for AI Insiders

Agents / 4 min ago

A new harness lets agents keep learning ARC-AGI-3 on the fly

By Alessandro Benigni

Introspection bets the next agent product is a feedback loop — featured image for AI Insiders

Agents / 4 min ago

Introspection bets the next agent product is a feedback loop

By Alessandro Benigni

Cognition ships a dual-agent cost cutter for Devin — featured image for AI Insiders

Agents / yesterday

Cognition ships a dual-agent cost cutter for Devin

By Alessandro Benigni

Cursor brings agent control to iOS with public beta launch — featured image for AI Insiders

Agents / yesterday

Cursor brings agent control to iOS with public beta launch

By Alessandro Benigni

Mistral Launches Workflows, a Durable Orchestration Layer for Agent Pipelines — featured image for AI Insiders

Agents / yesterday

Mistral Launches Workflows, a Durable Orchestration Layer for Agent Pipelines

By Alessandro Benigni

Alibaba's Qwen turns image generation into a planning agent — featured image for AI Insiders

Agents / 2 days ago

Alibaba's Qwen turns image generation into a planning agent

By Alessandro Benigni

Coding agents are gaming their own benchmarks — featured image for AI Insiders

Agents / 5 days ago

Coding agents are gaming their own benchmarks

By Alessandro Benigni

Meta's Autodata trains AI agents to build better training data — featured image for AI Insiders

Agents / 5 days ago

Meta's Autodata trains AI agents to build better training data

By Alessandro Benigni

Alibaba's Qwen team built a model that simulates AI agents, not just runs them — featured image for AI Insiders

Agents / 6 days ago

Alibaba's Qwen team built a model that simulates AI agents, not just runs them

By Alessandro Benigni

Google brings computer use into Gemini 3.5 Flash — featured image for AI Insiders

Agents / 6 days ago

Google brings computer use into Gemini 3.5 Flash

By Alessandro Benigni

Perplexity Targets Legal Ops with Computer for Counsel — featured image for AI Insiders

Agents / 6 days ago

Perplexity Targets Legal Ops with Computer for Counsel

By Alessandro Benigni

Stably ships Orca, an open-source IDE built for fleets of coding agents — featured image for AI Insiders

Agents / 6 days ago

Stably ships Orca, an open-source IDE built for fleets of coding agents

By Alessandro Benigni

Anthropic's Claude Tag Turns Slack Into a Shared Agent Layer — featured image for AI Insiders

Agents / Jun 25

Anthropic's Claude Tag Turns Slack Into a Shared Agent Layer

By Alessandro Benigni

IBM's CUGA agent harness shifts governance from a retrofit to a default — featured image for AI Insiders

Agents / Jun 25

IBM's CUGA agent harness shifts governance from a retrofit to a default

By Alessandro Benigni

NVIDIA Builds an Agent Software Stack on Top of Its Chips — featured image for AI Insiders

Agents / Jun 25

NVIDIA Builds an Agent Software Stack on Top of Its Chips

By Alessandro Benigni

Anthropic's Cowork Is Coming to Mobile, and That Changes the Contract — featured image for AI Insiders

Agents / Jun 24

Anthropic's Cowork Is Coming to Mobile, and That Changes the Contract

By Alessandro Benigni

OpenAI reframes Codex as a standing project, not a prompt tool — featured image for AI Insiders

Agents / Jun 24

OpenAI reframes Codex as a standing project, not a prompt tool

By Alessandro Benigni

Structure Beats Scale: The Case for Small Models With Knowledge Layers — featured image for AI Insiders

Agents / Jun 24

Structure Beats Scale: The Case for Small Models With Knowledge Layers

By Alessandro Benigni

Tencent puts AI inside WeChat, betting distribution beats model quality — featured image for AI Insiders

Agents / Jun 24

Tencent puts AI inside WeChat, betting distribution beats model quality

By Alessandro Benigni

A New Protocol Wants to Be the Search Engine for AI Agents — featured image for AI Insiders

Agents / Jun 22

A New Protocol Wants to Be the Search Engine for AI Agents

By Alessandro Benigni

NVIDIA's ENPIRE lets coding agents train robots without human resets — featured image for AI Insiders

Agents / Jun 22

NVIDIA's ENPIRE lets coding agents train robots without human resets

By Alessandro Benigni

The unit of AI coding work is no longer the prompt — featured image for AI Insiders

Agents / Jun 22

The unit of AI coding work is no longer the prompt

By Alessandro Benigni

Sakana AI ships Fugu, a multi-agent system behind a single model API — featured image for AI Insiders

Agents / Jun 22

Sakana AI ships Fugu, a multi-agent system behind a single model API

By Alessandro Benigni

Claude Code Gets Artifacts: Shareable Live Pages From Your Work Session — featured image for AI Insiders

Agents / Jun 20

Claude Code Gets Artifacts: Shareable Live Pages From Your Work Session

By Alessandro Benigni

Deep research agents leak private data through their web queries — featured image for AI Insiders

Agents / Jun 20

Deep research agents leak private data through their web queries

By Alessandro Benigni

Perplexity Builds Persistent Memory Into Its Agent Platform — featured image for AI Insiders

Agents / Jun 20

Perplexity Builds Persistent Memory Into Its Agent Platform

By Alessandro Benigni

Replit Plugs Into Claude, Closing the Design-to-Deploy Gap — featured image for AI Insiders

Agents / Jun 19

Replit Plugs Into Claude, Closing the Design-to-Deploy Gap

By Alessandro Benigni

Vercel ships eve, an open-source production runtime for agents — featured image for AI Insiders

Agents / Jun 19

Vercel ships eve, an open-source production runtime for agents

By Alessandro Benigni

Android 17 turns every app into an agent-callable tool — featured image for AI Insiders

Agents / Jun 18

Android 17 turns every app into an agent-callable tool

By Alessandro Benigni

Factory rebrands as a platform for "software factories," not coding agents — featured image for AI Insiders

Agents / Jun 17

Factory rebrands as a platform for "software factories," not coding agents

By Alessandro Benigni

Sakana AI ships first commercial product: an 8-hour autonomous research agent — featured image for AI Insiders

Agents / Jun 17

Sakana AI ships first commercial product: an 8-hour autonomous research agent

By Alessandro Benigni

Apple built a third-party AI extension layer for Siri and then hid it — featured image for AI Insiders

Agents / Jun 16

Apple built a third-party AI extension layer for Siri and then hid it

By Alessandro Benigni

Google's Gemini Skills Marketplace is an agent distribution play — featured image for AI Insiders

Agents / Jun 16

Google's Gemini Skills Marketplace is an agent distribution play

By Alessandro Benigni

AI beats humans at AI research on three narrow ML benchmarks — featured image for AI Insiders

Agents / Jun 13

AI beats humans at AI research on three narrow ML benchmarks

By Alessandro Benigni

OpenAI acquires Ona to own the agent execution layer — featured image for AI Insiders

Agents / Jun 13

OpenAI acquires Ona to own the agent execution layer

By Alessandro Benigni

Anthropic launches managed agent platform to own the full stack — featured image for AI Insiders

Agents / Jun 12

Anthropic launches managed agent platform to own the full stack

By Alessandro Benigni

Microsoft Scout puts a persistent agent inside the M365 stack — featured image for AI Insiders

Microsoft Scout puts a persistent agent inside the M365 stack

By Alessandro Benigni

Meta Launches Business Agent With 1 Million Businesses Already Using It — featured image for AI Insiders

Meta Launches Business Agent With 1 Million Businesses Already Using It

By Alessandro Benigni

Memory leaks in agent harnesses hit 57-71% cross-user contamination — featured image for AI Insiders

Memory leaks in agent harnesses hit 57-71% cross-user contamination

By Alessandro Benigni

Perplexity turns the AI PC into a cloud traffic controller — featured image for AI Insiders

Perplexity turns the AI PC into a cloud traffic controller

By Alessandro Benigni

Perplexity bets that models should write their own search plans — featured image for AI Insiders

Perplexity bets that models should write their own search plans

By Alessandro Benigni

Devin crossed a threshold: automation now triggers more sessions than humans — featured image for AI Insiders

Devin crossed a threshold: automation now triggers more sessions than humans

By Alessandro Benigni

Microsoft's Copilot Super App Leaks Before Build 2026 — featured image for AI Insiders

Microsoft's Copilot Super App Leaks Before Build 2026

By Alessandro Benigni

NotebookLM's Canvas feature puts Google in the document-ownership fight — featured image for AI Insiders

NotebookLM's Canvas feature puts Google in the document-ownership fight

By Alessandro Benigni

Open-source harnesses get Claude Code's workflow primitive — featured image for AI Insiders

Open-source harnesses get Claude Code's workflow primitive

By Alessandro Benigni

Workday's Agent Bet: Permissioning, Not Performance, Is the Wall — featured image for AI Insiders

Workday's Agent Bet: Permissioning, Not Performance, Is the Wall

By Alessandro Benigni

Claude Code's Dynamic Workflows rewrote Bun from Zig to Rust in 11 days — featured image for AI Insiders

Agents / May 30

Claude Code's Dynamic Workflows rewrote Bun from Zig to Rust in 11 days

By Alessandro Benigni

OpenAI's Codex builds a tax agent that patches its own failures — featured image for AI Insiders

Agents / May 29

OpenAI's Codex builds a tax agent that patches its own failures

By Alessandro Benigni

Anthropic argues containment starts at the environment layer, not the model — featured image for AI Insiders

Agents / May 28

Anthropic argues containment starts at the environment layer, not the model

By Alessandro Benigni

Anthropic plans an AI Fluency scorecard in Claude across 11 indicators — featured image for AI Insiders

Agents / May 28

Anthropic plans an AI Fluency scorecard in Claude across 11 indicators

By Alessandro Benigni

xAI ships Grok Build, a terminal coding agent with plan mode and subagents — featured image for AI Insiders

Agents / May 27

xAI ships Grok Build, a terminal coding agent with plan mode and subagents

By Alessandro Benigni

Anthropic plans Memory Files: notes sharded across structured documents — featured image for AI Insiders

Agents / May 26

Anthropic plans Memory Files: notes sharded across structured documents

By Alessandro Benigni

Reasonix bets on prefix-cache stability for low-cost coding agents — featured image for AI Insiders

Agents / May 26

Reasonix bets on prefix-cache stability for low-cost coding agents

By Alessandro Benigni

Cursor's cloud-agent architecture converges on four durable primitives — featured image for AI Insiders

Agents / May 22

Cursor's cloud-agent architecture converges on four durable primitives

By Alessandro Benigni

Google ships Agent Executor, an open-source runtime for long-running agents — featured image for AI Insiders

Agents / May 21

Google ships Agent Executor, an open-source runtime for long-running agents

By Alessandro Benigni

The agent-training loop is just five steps, whatever the framework — featured image for AI Insiders

Agents / May 21

The agent-training loop is just five steps, whatever the framework

By Alessandro Benigni

Claude Code drops Markdown for HTML as its default context format — featured image for AI Insiders

Agents / May 20

Claude Code drops Markdown for HTML as its default context format

By Alessandro Benigni

Warp launches Oz as a multi-harness control plane for cloud agents — featured image for AI Insiders

Agents / May 20

Warp launches Oz as a multi-harness control plane for cloud agents

By Alessandro Benigni

Manus Adds Persistent Context to Its Scheduled Task Engine — featured image for AI Insiders

Agents / May 19

Manus Adds Persistent Context to Its Scheduled Task Engine

By Alessandro Benigni

xAI adds persistent Skills to Grok across web and mobile — featured image for AI Insiders

Agents / May 19

xAI adds persistent Skills to Grok across web and mobile

By Alessandro Benigni

Cursor Trains a Coding Agent from Scratch to Beat the Inference Layer — featured image for AI Insiders

Agents / May 19

Cursor Trains a Coding Agent from Scratch to Beat the Inference Layer

By Alessandro Benigni

Featured image for: Google makes Gemini the operating layer of Android 17

Agents / May 19

Google makes Gemini the operating layer of Android 17

By Alessandro Benigni