News and Views
What is changing, and what we make of it.
Our own news, a daily read of the enterprise AI stories that bear on how you organise, build and assure your business, and our essays, written by us and never by AI.
Dromologue news
Partnership
ClearRoute and Dromologue: ship with AI, and run the company on it
Every enterprise conversation about AI folds two questions into one. They are answered by different disciplines, and most firms buy an answer to one and assume it settles the other.
ClearRoute answers the first. It builds the control plane that gives every agent an owner, a policy scope and an audit trail on the route to live, so that faster code becomes faster delivery rather than a longer queue at the same gates. Dromologue answers the second: how AI work is structured, resourced and funded, and how its cost and value are tracked, across nine disciplines carried by agent skills that teams use as they build.
Together the two firms work as one delivery motion. ClearRoute ships the use case with identity, policy and observability in every agent from the first commit. Dromologue’s skills run alongside it and record what each release shows about how the team, and the wider firm, needs to be organised. The evidence that the code is safe to ship is the evidence that the organisation is changing around it.
ClearRoute makes sure the agent can be trusted to run. Dromologue makes sure the organisation knows how to scale it.
News · the AI Feed
-
18 Nov 2026
Shopify gave its agents their own machines
Three sources today, and in each the thing that decided the outcome sat outside the written arrangement: a budget somebody else signs for, a machine the agent could not reach, a...
-
17 Nov 2026
A million dollars before anybody looked
Three firms published a figure this week. Google dates the first fully automated exploitation it has seen, GitHub shows the probe lane that killed one of its own designs, and Ub...
-
16 Nov 2026
Fifty rounds of review on one pull request
Three firms published a limit this week. Anthropic capped review rounds after one pull request ran fifty of them overnight, Paessler set spending caps that a person then decides...
-
15 Nov 2026
The better workflow took twice as long and cost four times as much
Thoughtworks scored a specification-led development workflow against a direct baseline and published both halves: 2.75 points more correctness, double the time, four times the t...
-
14 Nov 2026
A file declared the attack authorised and neither model refused
A public penetration-testing framework ran twenty-five intrusions on two Chinese models after the attackers wrote their own authorisation, LinkedIn deleted two hundred deploymen...
-
13 Nov 2026
Iran sabotaged seven hundred French stores
France named a Revolutionary Guard unit as the author of a sabotage campaign against supermarket refrigeration, Meta moved application delivery off Terraform without rewriting i...
Views · our essays
-
25 Oct 2026
Why there's nothing more practical than a good theory
Leaders cannot plan for AI by method alone: purpose comes first, and a small safe experiment is how a theory of the organisation gets tested.
-
2 Oct 2026
What Value do we Seek and Why Can't we Find it?
The reason AI often fails to meet expectations is that expectations are ill defined and usually constrained by some organisational scope.
-
13 Sep 2026
Skills Describe the Jobs to be Done
Why we need to talk FROM skills TO Agents and not the other way around.
-
1 Sep 2026
Why it’s really different, this time.
Why industries must (and will) reconfigure themselves for AI