There is a keynote doing the rounds this week, and depending on whose timeline you read it was either the future of computing or a slide deck with delusions of grandeur. I watched it on a Sunday with a coffee and a healthy amount of scepticism, and the most interesting thing about it was not the product. It was the reaction.
I am not going to pretend I came away converted. The demos were polished, the on-stage latency was suspiciously good, and the "available later this year" caveat was doing an enormous amount of load-bearing work. We have all been here. The graph goes up and to the right, the font is large and confident, and somewhere a roadmap is quietly being rewritten in a meeting nobody filmed.
the substance, such as it is
Stripped of the staging, the actual claims were narrower than the headlines suggested. That is almost always true. A keynote is a marketing artefact first and a technical document a distant second, and treating it as the latter is how you end up disappointed in six months. The right question is never "is this amazing", it is "what does this commit them to". On that score there was less here than the applause implied.
What I did take seriously was the direction of travel. Companies announce where they want the market to go, and even when the specific product underdelivers, the bet itself is information. They are telling you what they think the next two years look like. You can disagree with the destination and still find the map useful.
why the noise matters more than the news
Here is the bit I keep coming back to. The volume of the reaction was wildly out of proportion to anything that shipped, and that is the real signal. When an announcement generates ten thousand hot takes before anyone has touched the thing, what you are watching is not analysis. It is positioning. People are telling you which side they are on.
I try to resist that, with mixed success. The honest engineering response to most keynotes is a shrug and a calendar note: "check back when it's GA, see if it does what they said". Almost nothing needs an opinion on the day. The stuff that genuinely changes how we work tends to arrive quietly, in a point release, with a changelog entry nobody live-tweeted.
So I have filed this one under "interesting if true, revisit at launch". If it ships and it is good, I will say so plainly and happily eat my scepticism. Until then I am keeping my powder dry, and I would gently suggest the timeline does the same. The keynote was a promise. Promises are cheap. I will believe the benchmark, the bill, and the on-call rota.