I write The Standing Wave from Sydney.
It is a weekly read on the AI build-out, published every Tuesday. Each issue tests the week’s news against a standing set of hypotheses across four layers: compute and capital spending, the physical layer of power and data centres, frontier model capability, and how AI is actually put to work. The Insight and Headline Call arrive by email, and the full Signal lives here.
A standing wave holds its shape while energy moves through it. The news is the energy. The hypotheses are the shape.
Since 2007 I have measured how news coverage changes what businesses and governments do next. Software has been part of that work from the start. In 2007 it meant an algorithm guessing whether a story was positive or negative. Today models read tone and context well enough to do work that used to need a team of analysts.
I have watched two shifts in how people get their information. The first was print losing the agenda. When publishers moved from selling subscriptions to selling clicks, the stories that travelled furthest were the ones that provoked, not the ones that informed. The second was social. I joined Facebook in late 2005 to keep up with friends overseas, then spent the next decade measuring what it became: feeds ranked by algorithm, attention sold by the impression, and creators who could outdraw a newspaper.
That work took me to Cape Town, Dubai and Abu Dhabi, Shanghai from 2012 to 2014, and Southeast Asia from 2014. In China it meant finding ways to bring WeChat and Weibo data into the same tools the rest of the world used.
The third shift is under way now. People ask a model and get an answer instead of a list of links. When the model recommends one brand, cites one source and leaves out another, it decides what the reader sees before they have clicked anything. I expect this shift to be the largest of the three, because the model no longer points to the information. It gives it. The Standing Wave tracks how fast the systems behind that answer are being built, and tests each claim against hypotheses written down before the news arrived.
By day I lead Meltwater in Australia and New Zealand. Meltwater sells media intelligence, including tools that track how AI models describe brands, and some of the companies this publication covers are its customers or partners. My employer knows about The Standing Wave. The calls here are mine, and they are made for readers, not customers.
I use AI tools to help gather and cross-check sources. The hypotheses, the calls and the final words are mine. Each week I read about 15 hours of reporting, research and releases before I write.
The investment disclosure, why the set is fixed and how a reading moves are on Why hypotheses?. The current set is on The hypotheses. When a call is wrong, the Signal says so, and the archive keeps the record.
The Signal is free. Subscribe from the homepage. It lands Tuesday.