Essay
When the technology isn’t there
It’s been quite a week for technology.
On Tuesday, problems with Google Cloud disrupted banks, public services and even court systems. Then on Wednesday we had the discovery of critical vulnerabilities in PHP, a language used by an extraordinary proportion of the world’s websites, with organisations racing to patch their systems and some taking services offline.
Two very different incidents, but I think they raise some similar questions about how dependent we’ve become on technology that most of us barely think about until it stops working.
I’m not suggesting we should stop relying on cloud services or open-source software. They’ve enabled an enormous amount of progress, and I don’t think the answer is to retreat from them. But I do wonder whether we’ve become better at designing services around the assumption that everything will work than we are at thinking through what happens when it doesn’t.
And that feels particularly relevant as we start to build more AI into our organisations. We’re rightly excited about what agents might do, how much work we could automate and how services could operate differently. But every new dependency also raises questions about what happens when a model is unavailable, a supplier has a problem or something we rely on turns out to be vulnerable.
I’ve always thought resilience should be about more than restoring the technology. It’s also about whether the service can continue, perhaps in a different way, while the technology is unavailable.
I suspect this week will have prompted quite a few organisations to revisit their assumptions. It certainly seems a good moment to ask whether we really understand our dependencies, and whether our plans for when things go wrong are as developed as our plans for what technology might make possible.