Sam Rogers

Archive

202646
202565
202413
202231
202135
202015
201917
20185
20172
20164
20151

Blog

1 min read

Everything I’ve written here, newest first. Browse by year below, or jump in via tags or the RSS feed.

The Holiday Weekend Scam Call

· 2 min read

The after-hours call over a holiday weekend, demanding you verify yourself right now, is a scam. It can wait until Monday.

Audit Thyself?

· 5 min read

AI models can't grade their own blind spot very well. People don't either.

Vocabulary Debt

· 5 min read

One word naming four systems used to just confuse people. Now it makes your AI confident, precise, and dead wrong.

Teamwork Makes The Dream Work

· 5 min read

Two models and a judge beat the best single frontier model. The lift lives in the reconciling, not the genius. Same law runs your org.

Intelligence: Rent or Own?

· 5 min read

Open-weight is two quarters behind the frontier and closing. For anything you'll run indefinitely, the math has tipped, and the buying window looks a lot like 2020.

Solution Shapes

· 4 min read

Organizations buy the shape they already know how to operate, then wonder why it never fits the problem.

On Time or Ready?

· 4 min read

The clock says 7:58 PM. But the curtain doesn't move until everyone is ready.

Your AI Agent Reads Setup Guides You Can't Fully See

· 5 min read

AI agents follow setup guides through HTML, PDFs, and terminal output. Those surfaces can hide instructions from you but not from the model. GuideCheck is an open standard that makes the whole instruction surface reviewable before an agent acts.

Permanent = Expired

· 4 min read

All information is perishable. AI info ages like milk.

Canadian Ancestry

· 3 min read

Turns out I was born Canadian, like my father before me.

50 Arrows

· 5 min read

Nearly one year in, I still don't know what I have.

Turnstile or "On The List"?

· 4 min read

Stripe shipped agent payments. The language for what agents may do was already open and free.

Last to Know

· 5 min read

The job changed from doing tasks to keeping queues coherent. The question has to change with it to stay useful.

Behind the Wall

· 4 min read

David Soria Parra, MCP co-creator, at AI Engineer Europe this week: "2026 is the year agents go to production." Not a prediction.

Too potent to ship

· 5 min read

This is the team that built the capability looking at the test results and concluding: uh-oh, this is way too potent for broad deployment right now.

The Gap Nobody Owns

· 5 min read

Right now, somewhere in your organization, four different teams are managing four different slices of AI risk.

Shared Operational Language

· 5 min read

Something quietly shifted in the last year. The format used to describe what a person can do and the format used to describe what an AI agent can do are converging. Same structure. Same questions:

Skills: Have < Use

· 5 min read

Two weeks ago we flagged the vocabulary fork. Last week we showed that skills were never guarantees to begin with.

Skills Are Bets, Not Gates

· 4 min read

Last week we flagged the vocabulary fork: "skills" now means two different things depending on which room you're in.

A New Layer of Trust

· 13 min read

How the blend of systems measures and human behavior measures compliment each other to form the new trust needed in the age of AI.

The Verb is the Tell

· 4 min read

Listen closely to how people at your org talk about AI. The verb they reach for tells you everything.

NearCon 2026

· 1 min read

Love the protocol. Just back from the SF event.

The Rulebook Doesn't Exist Yet

· 4 min read

Last week I taught 27 L&D professionals in a 3-day AI certificate program. The very same week, I entered a hackathon and landed at NEARCON, a conference packed with crypto-native builders racing towar

The Understanding Trap

· 6 min read

Last week I wrote about the dangers of saying "not my job" about AI. This week, I promised we'd talk about what to build.

"Not My Job" is Predictive

· 4 min read

When I present about AI or teach skills to people who are warm to the idea, I keep hearing the same pushback: "That's not my job."

Your AI Agent Skills Have a Version Control Problem

· 5 min read

A meta-skill for tracking agent skill versions across stateless sessions, surfaces, and platforms. Works with Claude, Codex, Gemini CLI, and any agentskills.io-compatible system.

The Yes Problem

· 9 min read

How organizations can move from AI pilots to production by engineering safe approval paths into workflows, not policies.

The Weight of Avoidance

· 5 min read

You know that feeling when you've been avoiding the bathroom scale?

The Handoff Gap

· 5 min read

AI generates fast. Ownership doesn't scale.

The Copilot Problem

· 6 min read

A [recent story](https://archive.ph/UjhW0) about a physician and an AI tool has sparked debate about whether AI belongs in clinical settings. But the incident pattern it describes isn't new, and it is

The Great AI Misallocation

· 17 min read

A clear-eyed look at how organizations are misallocating AI spend and eroding trust through tooling-first decisions.

The One-Hour Year

· 5 min read

I'm not big on Twitter/X, but these two tweets with combined 22 million views tell a compelling story about where we are as 2026 kicks off.

Managers over Models

· 6 min read

Most companies think their AI policy lives in a document.

Friction Creates Shape

· 4 min read

Here in the US, it’s Thanksgiving week. That means an overload of gratitude posts for all the usual accelerators: breakthroughs, mentors, tools, breakthroughs pretending to be tools. There’s no wrong way to be thankful, but let’s take a different angle.

Synthetic Trust

· 2 min read

Modern teams trust tone more than truth. Confident output passes as credible long before its reasoning is verified.

Human Interface Fatigue

· 3 min read

_The more “human-friendly” our tools become, the more mental overhead they create._

Confidence & Calibration

· 2 min read

Confidence used to be a proxy for competence. AI changed that. The differentiator now is calibration.

Halloweens

· 1 min read

Halloween highlights from the last dozen years

Immature AI Maturity

· 2 min read

Most AI maturity frameworks measure self-assessment and inflated usage data. Real maturity is the gap between what you believe you understand and what you can actually explain.

Success at the PAICE of Work

· 2 min read

PAICE is live. It measures what adoption metrics don't — how humans and AI actually perform together.

The Trust Gap

· 3 min read

AI adoption isn't waiting for anyone. Employees are already ahead. The Trust Gap is upon us — and Integrators hold the keys.

Substacking

· 1 min read

Sharing there now too

5% Success

· 3 min read

AI transformation is like baking bread.

95% failure

· 4 min read

MIT says most orgs are getting zero return on GenAI

GPT-5 Changes the Game (Mostly Doesn't)

· 2 min read

The long-awaited iteration of the model that started the public love/hate affair with AI landed last week. And (predictably) the reaction is still love/hate.

Issue 010

· 3 min read

Your sales funnel is losing leads faster than you think, and the loss starts long before a salesperson speaks to anyone.

My Music YES!

· 3 min read

If you want me on your project, here's what I need

Automated State

· 1 min read

Quietly revolutionary ideas for building a better world

Align to the Spec, Not to the Prompt

· 4 min read

A quiet but meaningful shift is underway: developers, product teams, and AI leaders are realizing that _alignment to the spec_ is what matters. And that starts with crafting the spec with care.

Open, auditable AI stacks just shipped!

· 3 min read

No, this wasn’t in [America’s AI Action Plan](https://www.ai.gov/action-plan) that grabbed headlines this week. Though it probably should have been.

When AI Outpaces Our Senses & Expectations

· 3 min read

When machines operate in contexts we can't perceive, trust breaks down. Not because the tool is wrong — because the interface to human confidence is missing.

Signals & Subtractions Issue 005

· 3 min read

Created by Sam Rogers** · Published by [Snap Synapse](https://snapsynapse.com) for people who prefer clarity over clutter

Signals & Subtractions – Issue 002

· 3 min read

AI adoption is accelerating. No surprise there. But while AI *tools* multiply, the **decision** **rights**, **governance structures**, and **coordination mechanisms** around them lag behind.

The Grand Plan

· 4 min read

So...what if there isn't one at all? Anywhere?

Wedding

· 1 min read

Wedding pics! Getcher wedding pictures here!

Living to 123

· 8 min read

This changes how you think about life.

From HR to L&D

· 4 min read

Navigating the waters of a role transition

Just dumb enough

· 2 min read

I was asked for 3-5 bullet introduction at work. Here’s what I gave them.

5 Bullet Intro

· 2 min read

I was asked for 3-5 bullet introduction at work. Here’s what I gave them.

Later is Now

· 1 min read

It's one name I answer to.

The work is free

· 2 min read

Someone asked me recently if I’d ever done any work for free, and if I enjoyed it.

Finally backed up

· 2 min read

It's been a while since I had this feeling, I don't quite trust it yet.

Faster Braces?

· 2 min read

What happens if you change Invisalign trays every 6 days instead of 7?

Colemak

· 2 min read

Your hands will thank you, eventually.

The Tytler Cycle

· 3 min read

Didn't know this was a thing. Looks like a thing, alright.

Data Te Ching, ch 4

· 1 min read

The cost of a confirmed assumption is the highest cost of all.

Data Te Ching, ch 3

· 1 min read

Data can sway choices only when choices are swayed by data.

Best of 2020

· 5 min read

Even in the worst of times, there's still a best.

You support what you eat

· 2 min read

Someone asked me recently if I’d ever done any work for free, and if I enjoyed it.

Weirdo

· 1 min read

It's one name I answer to.

The End

· 1 min read

As another year comes to a close

Why I Do Social Media Fasts

· 4 min read

Once a year I take a month off social media to remember what the rest of life is like. This year I timed it with the altMBA. Way better.

You made a course!

· 2 min read

Everything must be all better for everyone now

Learning Asset Framework -- A Proposal in Process

· 2 min read

What is the minimum viable spec for corporate training? A framework proposal for an industry that talks about learning as if it's a thing without ever agreeing on what kind of thing.

Podcasting for Learning & Development

· 2 min read

I wanted to hear how others in L&D were solving the problems they faced. Podcasts turned out to be the next best thing to hallway conversations at conferences.

Not Even Wrong

· 2 min read

Adam Savage called it at DevLearn: sometimes we don't know enough to even be wrong. What do you do when someone's fundamental understanding is so absent they can't yet ask a useful question?

ATD 2016 as a New Speaker

· 3 min read

What it's actually like to speak at the biggest L&D conference in the world -- and why the hallways are worth more than the sessions.

How to Produce 90 eLearning Videos in 8 Weeks

· 4 min read

Google had a problem. We solved it by producing the YouTube Certified online training program -- 90 videos in 8 weeks, 10x more professionals trained in the following year.

Graph View

Backlinks