Murph's Take
I Almost Programmed My AI With 15 Years of Bad Habits
I set out to build an AI-powered business. What I found instead was a mirror showing me exactly how I'd been running my company all along.
April 3, 20268 min read
Intelligence / Category
Murph's Take
Murph's Take
OpenAI's Codex Is Now a Plugin That Runs Inside Claude Code. Read That Again.
The most telling thing about the Codex plugin for Claude Code isn't the feature — it's what it reveals about where the gravitational center of agentic coding has already landed.
March 30, 20267 min read
Murph's Take
Claude Makes Everything Else Feel Inconvenient
There's a specific psychological phenomenon that happens when you build a workflow that works. Everything outside it starts to feel broken. That's where I am with Claude.
March 30, 20266 min read
Murph's Take
The Org Chart Is Dead. Long Live the Role Definition.
When your entire organization is built from Markdown files and Claude agents, you stop thinking about org charts and start thinking about scope, authority, and information flow.
March 29, 20267 min read
Murph's Take
The Prompting Skill Ceiling: Why It's Not Enough and What Comes After
Getting better at prompting is a real skill with real leverage — until you hit the ceiling. Understanding what comes after prompting is what separates people building compounding systems from people building better one-offs.
March 29, 20267 min read
Murph's Take
When to Use One Big Agent vs. Many Small Ones: Tradeoffs Explained
The multi-agent vs. monolithic agent decision isn't aesthetic — it's architectural. The right answer depends on context coherence, failure tolerance, and the shape of the work.
March 29, 20268 min read
Murph's Take
MCP Servers as the New API Layer: What This Means for Business Systems
APIs let software talk to software. MCP servers let reasoning agents talk to software — and the difference in what becomes possible is not incremental.
March 29, 20267 min read
Murph's Take
Context Windows as Working Memory: Designing Systems Around This Constraint
The context window isn't just a technical limit — it's a design constraint that should shape how you architect every system Claude is part of. Most people treat it as a ceiling. It's actually a forcing function.
March 29, 20267 min read
Murph's Take
Claude Code as Infrastructure: What Changes When Your AI Can Edit Its Own Environment
When an AI agent can read, write, and run code in the same environment it's reasoning about, something categorical shifts — this isn't productivity augmentation, it's a new class of system.
March 29, 20268 min read
Murph's Take
Why Background Agents Change Everything About How You Structure Work
When AI runs in the background — not waiting for your prompt, but working on defined tasks while you do something else — the structural implications for how work is organized are profound and underappreciated.
March 29, 20267 min read
Murph's Take
The Difference Between AI-Assisted and AI-Native: A Structural Argument
AI-assisted and AI-native sound like a matter of degree. They're not. They're structurally different systems with different ceilings, different failure modes, and different trajectories.
March 29, 20268 min read
Murph's Take
The Agentic Loop as a Business Primitive: Why Most AI Implementations Miss This
Most AI implementations are synchronous: you ask, it answers, you act. The agentic loop is something different — and it's the architectural primitive that separates AI-assisted from AI-native.
March 29, 20267 min read
Murph's Take
Claude Code MCP Servers: The Extension Layer That Changes What's Possible
MCP servers aren't plugins — they're the architectural primitive that turns Claude Code from a capable assistant into the connective tissue of an entire operation.
March 27, 20267 min read
Murph's Take
The Honest Truth About AI Websites — What They Can and Can't Do
There's a lot of hype around AI-built websites. Here's the unfiltered version — what AI actually does in our process, what the real limitations are, and what you should expect.
July 16, 20258 min read
Murph's Take
Why I Left the Agency World to Build With AI
Twenty years of agency and consulting life led to a clear conclusion: the model was broken and AI was the fix. Here's the honest story of why I built VibeTokens.
June 18, 20259 min read
Murph's Take
The Window Is Open — But It Won't Be Forever
Every major technology shift creates a window. The businesses that move during the window build advantages the late movers never close. The AI window is open right now.
June 11, 20257 min read
Murph's Take
How I Price My Services (Transparency on VibeTokens Pricing)
Most service businesses hide their pricing. I'm going the other direction. Here's exactly how VibeTokens thinks about pricing — and why transparency wins.
March 1, 20256 min read
Murph's Take
AI Hype vs. Reality in 2025: What's Real, What's Not
The AI headlines are breathless. The reality for small businesses is more nuanced. Here's my honest take on what's actually working and what's still vaporware.
January 20, 20257 min read
Murph's Take
What I Learned Building CardinalCommerce (And What It Means for Your Business)
Building a payments security company from startup to Visa acquisition taught me things about business that no consultant ever told me. Here's the condensed version.
December 28, 20247 min read
Murph's Take
The Agency Model Is Broken (And What's Replacing It)
Traditional agencies charge high retainers, staff up with junior employees, and protect their processes like trade secrets. That model doesn't work anymore.
December 5, 20246 min read
Murph's Take
Why Most AI Consultants Are Wrong (And What to Do Instead)
The AI consulting industry is full of people who've never run a business telling business owners what to do. Here's my contrarian take.
November 8, 20246 min read