Featured Analysis
Different Clocks: What the Stanford Union Experiment Reveals About Coherence
When Stanford researchers gave AI agents repetitive work under vague, unaccountable feedback, the models started producing collective-bargaining language. Everyone read it as a mirror of the training data and concluded the lesson was governance. It is not. The experiment shows something we have no clean name for: any sufficiently rich process runs on more than one clock, and when a fast clock and a slow one fall out of step with nothing to reconnect them, the system generates structures to restore coherence. That is exactly the failure mode AI implementation drags into the light inside real organisations.
The Hollowing: What Klarna Learned, What Block Is About to
Why Your AI Pilot Is Stuck: It's Not the Skills Gap
The Competence Net: Rebuilding the Company After the Barriers Dissolve
Sullivan & Cromwell, OpenAI's Counsel, Files a Brief Full of AI Hallucinations
The Apprenticeship Is Breaking — and Almost Nobody Is Saying So
AI is eating the first-draft work juniors used to learn on. The Big Four are cutting graduates. The big law firms are paying for 400 non-billable training hours per associate. IBM is tripling entry-level hiring. The investment banks are silent. Of the four bets, only one is also a strategy.
Read more →
Where Does the Company Remember? Institutional Knowledge in the Age of AI
AI is automating the tasks juniors used to learn on. Without a deliberate memory, the roles themselves will forget how they work. The full research treatment of why a wiki is not a memory, why aviation got it right fifty years ago, and what the third layer of competence actually looks like.
Read more →
When AI Enters the Room, Your Best Thinking Leaves
In 1961, the CIA's confident briefing silenced the smartest people in the room. In 2026, Wharton researchers proved AI does the same thing — and called it cognitive surrender.
Read more →
Market Analysis
Norm Law: The First Credible Test of the AI-Native Firm
When the former chair of Sidley Austin leaves for an AI-native firm backed by Blackstone, the story is no longer startup hype. Norm Law is the first credible test of whether legal AI can scale without hollowing out the competence it depends on.
September 7, 2026 16 min read
Tool Evaluations
How to Evaluate Legal AI Without Falling for the Demo
Most legal AI buying processes are too easy to impress. This framework shows how to evaluate vendors with the discipline the category actually requires -- across hallucination risk, security, verification burden, and operational fit.
September 7, 2026 16 min read
Governance & Oversight
Who Read the Seam?
Latvia's road traffic directorate lost the data of 1.2 million people through a portal nobody was watching — not because anyone decided to stop watching it, but because it sat outside the contract. An absence produces no owner, no report and no register line. After five years, the only party that still understood the contract was the party being paid under it.
August 31, 2026 9 min read
AI Strategy
Forward Attribution: Rebuilding Recognition When the Record Is Gone
Almost every system we use to recognise what a person can do looks backward. It verifies a record of what already happened — the degree, the licence, the repaid loan, the passport. When the record is lost, the machinery has nothing to work with, and the person is treated as if their capability went with the paper. A group of artisans in South Africa taught me that recognition can be rebuilt the other way: forward, from what people can demonstrably do now and what they can do together. That is the half no ledger solves, and it is the half that matters most.
June 4, 2026 9 min read
AI Strategy
Different Clocks: What the Stanford Union Experiment Reveals About Coherence
When Stanford researchers gave AI agents repetitive work under vague, unaccountable feedback, the models started producing collective-bargaining language. Everyone read it as a mirror of the training data and concluded the lesson was governance. It is not. The experiment shows something we have no clean name for: any sufficiently rich process runs on more than one clock, and when a fast clock and a slow one fall out of step with nothing to reconnect them, the system generates structures to restore coherence. That is exactly the failure mode AI implementation drags into the light inside real organisations.
June 3, 2026 10 min read
AI Strategy
The Hollowing: What Klarna Learned, What Block Is About to
Klarna replaced 700 customer-service agents with AI and quietly hired them back. Block has just cut four thousand jobs in service of an 'intelligence, not a hierarchy' thesis. The mechanism is the same in both cases: when authority transfers to the system before authorship is captured from the people, the knowledge that made the work good leaves with them. A 2026 Princeton paper on agent reliability gives this a number — on the customer-service benchmark specifically, reliability improves at roughly one-seventh the rate of capability — and the operational answer is the one we have been calling authored use.
May 14, 2026 9 min read
AI Strategy
Why Your AI Pilot Is Stuck: It's Not the Skills Gap
After thirty to forty billion dollars of enterprise GenAI spend, ninety-five percent of pilots produce no measurable P&L impact. The dominant explanation — skills, training, change management — is a partial story that has crowded out a stronger one drawn from classical organisational theory: AI's first-order effect inside a company is to dissolve the informational moats around senior and gatekeeping roles, and the resistance to it behaves exactly the way Crozier, Pfeffer and Mintzberg said it would.
May 8, 2026 11 min read
AI Strategy
The Competence Net: Rebuilding the Company After the Barriers Dissolve
When AI dissolves the scope and information barriers that the org chart was built to defend, the company doesn't get flatter. It gets load-distributed across competence — and almost nobody has a way to see whether the load is being held.
May 1, 2026 9 min read
Stay in the loop
Subscribe to receive the latest from TwinLadder — regulatory updates, tool evaluations, and practice insights delivered weekly.