09 / Insights

Essays that practice
what they teach.

Applied AI, learning science and marketing performance. Each essay adapts to what you already know, explains its statistics instead of just citing them, checks your understanding as you read, and ends with an exercise you can use at work the same day.

◆ Adapts to your prior knowledge ◆ Statistics shown, not asserted ◆ Retrieval checks throughout ◆ One workplace exercise per essay
0.79σ
What tutoring systems achieve · drag it yourself inside
Learning Science × AI07 Jul 202612 min + exerciseINTERACTIVE

Adaptive Learning Is Not Personalisation

Netflix personalises. It does not teach. Take the 30-second diagnostic, read on the path it recommends, drag the effect-size evidence yourself, then audit a vendor demo before you ever sit through a real one.

Start the essay →
Learning Science × AIINTERACTIVE

How Do You Know an AI Tutor Taught Anyone Anything?

Usage, test scores and changed behaviour are three claims with three instruments. Separate them, then build the strongest sentence your own evidence can defend.

14 Jul 202612 min
Learning Science × AIINTERACTIVE

The Job Moved from Prompting to Delegating

A quarter of Codex users delegated a task of more than a full working day in one request. Translate the shares to your team, then audit your AI training against the task that is actually changing.

08 Jul 202612 min
Performance MarketingINTERACTIVE

The Day Your Target Starts to Bind

On 17 August, Google Ads makes your target the binding constraint. Put your own numbers in to see your exposure, then take a three-question account audit to the deadline.

08 Jul 202611 min
Learning Science × AIFORTHCOMING

The $0 ROI Problem: Three Technology Booms, One Structural Error

$650B of AI capex meets MIT’s finding that 95% of enterprise GenAI pilots show no measurable P&L impact. With a workflow-redesign audit exercise.

In production10 min
Learning Science × AIFORTHCOMING

Two Companies, Two Theories of Learning

Google and Anthropic take opposite public positions on how people learn. Includes a classroom-philosophy classifier you can apply to any tool.

In production9 min
AssessmentFORTHCOMING

The Assessment Validity Problem

When the same model designs and completes the assessment, validity collapses. With a redesign worksheet for your own assessments.

In production10 min
Performance MarketingFORTHCOMING

The CMO Competency Trap

98% of CMOs use AI, 30% can scale it. Self-locate your organisation on the three-stage maturity model and get the next move for your stage.

In production8 min
AssessmentFORTHCOMING

How Do You Know an AI Tutor Taught Anyone Anything?

Usage is not learning. Kirkpatrick’s forgotten levels applied to the AI tutoring wave, with a measurement-design template.

In production11 min
Performance MarketingFORTHCOMING

When AI Handles Execution, Judgment Becomes the Constraint

Meta automates the campaign, Google automates the bid. Assess your team against the four judgment competencies that remain.

In production9 min
(function(){ ‘use strict’; var shell = document.getElementById(‘m4i-insights-index’); if (!shell) return; var filters = shell.querySelector(‘#m4i-filters’); if (!filters) return; filters.addEventListener(‘click’, function(e){ var b = e.target.closest(‘.m4i-chip’); if (!b) return; filters.querySelectorAll(‘.m4i-chip’).forEach(function(c){ c.classList.remove(‘m4i-on’); }); b.classList.add(‘m4i-on’); var f = b.getAttribute(‘data-f’); shell.querySelectorAll(‘#m4i-cards .m4i-card, .m4i-featured’).forEach(function(c){ c.style.display = (f === ‘all’ || c.getAttribute(‘data-p’) === f) ? ” : ‘none’; }); }); })();