AI DOERS
Book a Call
AI DOERS · Insights

Playbooks for performance marketing.

Sharp, practical thinking on Meta and Google ads, lead generation, measurement, and AI for local and home-service businesses.

AI Excellence

Claude's New Plugin Packs Bring Real Automation to Small Business

Claude shipped preset plugin packs for small business and legal work that bundle prompts, live data connectors, and automations into one system. Here is what they do, where they fall short, and how to set them up.

May 21, 2026 · 10 min read
AI Excellence

Why a Top Founder Dropped Claude Code for Codex

He switched because Codex gave him a tighter agentic loop and used far fewer tokens for the same task, and because he believes agents are about to absorb the software layer entirely, making context and brand the only durable edge.

May 21, 2026 · 10 min read
AI Excellence

What Is WebMCP and Why It Makes Browser Agents Reliable

WebMCP is a Chrome capability that lets a website declare agent-callable tools directly in its HTML or JavaScript, so browser agents take deterministic, schema-checked actions on your site instead of guessing from raw HTML or screenshots.

May 21, 2026 · 10 min read
AI Excellence

What the Pope's 40,000-Word AI Encyclical Actually Says

The Pope's encyclical Magnifica Humanitatis argues that AI power is concentrating in too few private hands and that simulated relationships can erode real human connection, a warning that doubles as a practical checklist for any business wiring AI into customer interactions.

May 21, 2026 · 10 min read
AI Excellence

Claude Mythos 5 and Fable 5 Explained: What Anthropic Actually Shipped

Anthropic shipped two models a tier above Opus. Fable 5 is available to everyone, while Mythos 5 is the same model with cybersecurity safeguards removed and is restricted to Glasswing partners. Here is what that means for picking a model for real work.

May 21, 2026 · 9 min read
AI Excellence

What Are Local AI Agents And How Do You Build One?

A local AI agent is an AI that runs on your own machine and takes actions on its own. You build one by assembling a home, a chat channel, a model brain, memory files, tools, a scheduling heartbeat, and computer-use eyes.

May 21, 2026 · 11 min read
AI Excellence

Why Keeping Up With AI Is Making You Fall Behind

Chasing every AI release traps you in a fear and FOMO loop that blocks action. The way out is to pick one clear goal, niche down on tools that already exist, and deliberately cut your AI news intake.

May 21, 2026 · 10 min read
AI Excellence

Should You Switch From ChatGPT or Claude to Gemini Right Now?

Switch if you want the most generous free tier, the strongest 100 dollar bundle, or real video and audio understanding, because Gemini 3.5 Flash benchmarks near the frontier models while running faster and cutting you off far less.

May 21, 2026 · 10 min read
AI Excellence

Claude Opus 4.8: Smarter, Same Price, And Dynamic Workflows

Claude Opus 4.8 raises intelligence over 4.7 while holding the same price, which is effectively a price cut, and adds dynamic workflows that fan out hundreds of parallel agents with adversarial checking. Here is what that means for a real business.

May 21, 2026 · 9 min read
AI Excellence

Don't Run Fable 5 in the Claude App, Run It Here Instead

Fable 5 is Anthropic's biggest model jump since GPT-4, but where you run it decides whether you get its power or a wall of refusals. Here is the setup I use so a tripped safeguard quietly downgrades instead of failing.

May 21, 2026 · 11 min read
AI Excellence

How to Build Claude Code Subagents Better Than 99% of People

A subagent is just a markdown file with a short header that the main session delegates work to, running it in a fresh context window. That clean context is why subagents save tokens, let you run cheaper models for grunt work, and run many independent jobs in parallel.

May 21, 2026 · 7 min read
AI Excellence

One Context Gives Coding Agents a Git-Style Memory That Actually Persists

One Context is a simple file framework that lets a coding agent branch, commit, and merge its own memory across any session, which lifted Claude Code performance by about 13 percent. Here is what persistent agent memory means for a real business.

May 21, 2026 · 10 min read
AI Excellence

Every Way to Run Open-Source AI Models, From Easiest to Hardest

There are four practical ways to run open-source AI models, from a two-minute local install to your own rented server. A small model already runs on an ordinary laptop, so any business can start today without buying hardware.

May 21, 2026 · 10 min read
AI Excellence

The White Collar Bloodbath Got Cancelled, and Here Is the Real Reason

The two loudest voices behind the AI-jobs-apocalypse walked it back because hiring is rising, not falling. The real bottleneck is not the model but that almost nobody knows how to use AI past simple chat, and most layoffs blamed on AI were overhiring corrections.

May 21, 2026 · 5 min read
AI Excellence

Google AI Mode Is Live Worldwide. Here Are the Four Things That Actually Changed

Google AI Mode, the biggest search upgrade in over 25 years, is now live everywhere and turns search into a conversation. The box expands as you type, a plus button adds files and images, follow-up questions work, and Personal Intelligence can connect Gmail and Photos, all running on Gemini 3.5 Flash. Here is how I would put it to work.

May 20, 2026 · 10 min read
AI Excellence

Anthropic Quietly Rebuilt Tool Calling for Claude. Here Is What Changed

Anthropic's advanced tool calling lets Claude write code that invokes many functions instead of emitting JSON one call at a time. That keeps bulky results out of context and can cut token use 30 to 50 percent while improving accuracy and speed.

May 20, 2026 · 9 min read
AI Excellence

Anthropic Says AI Is Starting to Build Itself. Here Is What the Numbers Actually Mean for Your Business

Anthropic's new paper argues AI is increasingly building AI, with task length doubling every four months and most of its own code now written by Claude. Strip out the fear and the real signal is simple: execution is getting cheap, so your judgment is the asset to protect.

May 20, 2026 · 10 min read
AI Excellence

Laid Off to Head of AI in One Year: What Actually Got Her Hired

One person went from a layoff with zero AI experience to running AI strategy for fifteen companies in a year, and the thing that got her hired was not a credential. It was recorded proof of what she had built. Here is the playbook any business or operator can copy.

May 20, 2026 · 10 min read
AI Excellence

Claude Cowork Fundamentals: From Messy Folders to Self-Running Workflows

Claude Cowork is a local AI agent that lives on your computer and acts on your real files, apps, and tools. You start by pointing it at a folder, save the useful steps as skills, connect your apps, then wrap it in a Project that remembers context and runs scheduled work with no code.

May 20, 2026 · 10 min read
AI Excellence

Why Claude Code Managed Agents Could Mint the Next Wave of AI Agency Owners

They move a fully loaded Claude agent into pay-per-runtime cloud infrastructure, and once you wrap it in a cognitive architecture of databases, documents, and nightly memory consolidation, you can package a continuously learning AI operating system that a client simply talks to.

May 20, 2026 · 10 min read
AI Excellence

Claude Opus 4.8: Why the Quick Course-Correct and Dynamic Workflows Matter

Claude Opus 4.8 is a fast course-correct on a divisive 4.7. It restores creativity, adds five effort levels, and introduces dynamic workflows that spawn hundreds of sub-agents to take large jobs all the way to done.

May 20, 2026 · 11 min read
AI Excellence

How to Prompt Gemini 3.1 Pro for Animations That Look Production-Ready

Gemini 3.1 Pro produces launch-grade animation only when you plan first and hand it a scene-based prompt with timing, a defined UI state, the action, and effect keywords. That single shift separates the model's spatial planning from its building and lifts quality two to three times.

May 20, 2026 · 9 min read
AI Excellence

Perplexity Computer: A Hosted AI Agent That Does What OpenClaw Does Without the Pain

Perplexity Computer is a fully hosted AI agent with a real code-running environment, hundreds of prebuilt connectors, selectable orchestrator models, and proactive scheduled tasks, so most people get serious agent capability without the setup, maintenance, or security burden.

May 20, 2026 · 10 min read
AI Excellence

How to Turn Claude Into a Second Brain and AI Operating System

Turn Claude into a second brain by defaulting to one tool and following the four C's in order: context as a routing tree, connections to live data, capabilities as reusable skills, and cadence for automations that run while you sleep.

May 20, 2026 · 11 min read
Insights | AI Doers