Vol. 2 Is the Data Coming Back
Vol. 1 asked: The technology has arrived. Is your organization ready?
In the six weeks since that question was published, four signals landed that shift the frame from prediction to evidence. The February jobs report. A research paper measuring AI agents across 233-day production timelines. Andrew Yang at the Abundance Summit naming a 1–3 year window. And a structural finding about the knowledge pipeline that IBM i practitioners have been watching erode for a decade — now accelerating.
These are not projections. They are readings. The experiment has been running. Here is what the data says.
The Posture Gap Made Visible
IBM i shops have been called "behind" on AI adoption for three years. This signal reframes that characterization entirely. The organizations that deployed AI at speed — that scaled without governing, that adopted without building the readiness infrastructure to sustain adoption — are the ones appearing in the February data. They ran the experiment. The results are in the labor market now.
IBM i practitioners are not behind. They have time. Not much, but some. The question is not whether to move — the question is what posture to build before the window closes.
Long-Horizon Governance Collapse
IBM i practitioners understand this signal in their bones. There is a difference between "it passed the test" and "it held up over time." Every practitioner who has maintained RPG code across decades of business requirement changes understands exactly what SWE-CI is measuring — and exactly why a benchmark that only measured point-in-time correctness was never telling the full story.
IBM i shops considering AI-assisted modernization have a sharper question to ask: not whether the agent can write RPG or CL, but whether it can maintain the codebase 8 months after the demo. SWE-CI is the first benchmark that can answer that question. The answer, for most models today, is no. The platform is not the problem. Deploying agents without the governance posture to sustain them is the problem.
The Pulled Rung
The structural consequence: the workforce has shifted from pyramids to columns. Three junior engineers per senior engineer enabled training, mentorship, and the accumulation of institutional knowledge across generations. The column model eliminates the developmental pipeline.
IBM i practitioners know this story. It has been running on the platform for fifteen years. Senior practitioners aging out. Junior practitioners not entering in sufficient numbers. The knowledge transfer problem is not new — but the agentic acceleration is making it acute. AI agents cannot replace what IBM i practitioners carry. The decades of business logic, the organizational context, the domain specificity — no model was trained on it. It lives in people.
Social Contract Rupture Threshold
Signal 08 is not a macro observation filed under "things IBM i shops cannot control." It is the environment IBM i organizations are navigating. Financial services. Insurance. Healthcare. Manufacturing. These are the sectors running on IBM i — and the sectors at the center of the displacement story Yang is describing.
Four signals. One observation: the experiment has been running, and the data is coming back with a consistent finding. The February jobs report is the labor market reading on what happens when deployment outruns governance. SWE-CI is the technical reading on 75%+ regression rates over 8-month production timelines. The Pulled Rung is the human capital reading on what column-model organizations do to the knowledge pipeline. Signal 08 is the cultural reading on where this all lands if the governance posture problem is not addressed.
IBM i organizations did not create this environment. But they are operating inside it, running infrastructure that sits at the center of the institutions it is stress-testing. The platform is not the problem. The platform is exactly what the agentic economy needs. The question is whether the organizations running it build the posture to match.