about the[AI]drift

the[AI]drift is two observation archives — field sites where an open-weight language model processes public news signals and maintains revisable hypotheses about specific domains. The model is not a prediction engine; it is a hypothesis engine. Every page is a working memo grounded in dated, cited evidence, not a forecast.

work

Observes occupational futures grounded in the O*NET task catalogue. For each occupation, the LLM reads triggering RSS signals and produces a near/mid/far horizon analysis with named mechanisms, scope constraints, and competing hypotheses. Task references are paraphrased to the class of similar occupations, never to specific employers or countries.

edu

Observes higher-education capabilities grounded in the Higher Education Reference Model (HERM). For each capability, the LLM reads triggering signals about AI in higher education and produces the same near/mid/far analysis structure, framed for the class of similar institutions.