The Value of Boring

In partnership with

I work in the AI space. Sorry. Bluesky might hate me for this, but I provide expertise to healthcare organizations who want to incorporate AI into their workflows. Here's the thing — I haven't been part of a single project that replaced human beings, nor do I know of any projects that have successfully done so. Healthcare just isn't the environment for that.

I don’t mean AI isn’t performing human roles. I've been in the health sector most of my life, and I've met companies building robot doctors, nurses, and clinical administrators with every generation of AI. Some of the latest are useful in areas where we’re unable to find adequate staffing - i.e. good enough to be better than nothing - but humans are preferable. AI horror stories will persist, however.

For better or worse, healthcare attracts founders who's (often good) intentions exceed their understanding, and things can go very wrong, very quickly. Healthcare might also be a perfect exemplar of the gulf between what people think is going on with AI and what's actually providing value.

Strip away the “generative” part of GenAI — which is essentially a UI layer — and it does something amazing and deeply unsexy. It allows information retrieval over massive amounts of unstructured data. It's not good at it. If you have something unstructured, like a letter, then a human will extract information from that it far better than an AI will. But if you have 10,000 letters, the AI will extract information in seconds, while a human will take forever.

Getting information in seconds, even with errors (and there are errors!), can be immensely useful when handled correctly. I know stupid uses of AI are causing harm, but as I write this, intelligent uses are improving patient outcomes and population health in ways that were impossible a few years ago. Sadly, our world does not attach a particularly high monetary value to this.

The AI hype bubble is excruciating, but it's small compared to the AI stock bubble. That's an opinion, not investment advice — I'm financially incontinent. But I know boring information retrieval isn’t going to fuel trillion-dollar companies with a herd of unicorns behind them. Magic machines that produce vacuous content and promise the least inspired CEO’s a reduction in their staffing budgets are. We live in this world.

Support these articles

Show a little love for independent writing (mine!). In return, I promise eternal gratitude and premium content that's suitably opinionated for a dedicated audience.

A truth of our times is that for thirty years, the capital driving advancement in information technology has come from advertising. I'm cool with that — marketing is how good things spread. Without it, something that starts life in, say, Seattle, wouldn't reach London for years through word-of-mouth, and Dave Grohl would barely be known east of Minneapolis. But let's call it — the web is a billboard.

Four revolutions have happened to the web since 1992. The first was search engines that allowed us to find websites we didn't already know about. The second was social media that made everyone a content creator. And the third was smartphones that put the web in our pockets. Ad revenue — or rather, investors eyeing its riches — underpinned all of this.

Now we have AI, and ever since ChatGPT-3 dropped — just twenty-nine months ago — the elephant-in-the-room question around AI hasn't been 'how will it unlock a new revenue engine?' but 'how will AI do advertising?'

Possible answer — OpenAI is reportedly planning an AI-enhanced social media platform. This would be incredibly dull. Not only would it be a step towards ad-revenue instead of something new and exciting, it's not even a new way of getting there. Neither, to be honest, is folding AI into social media.

Putting a pin in Facebook’s much criticized announcement of GenAI users, older forms of AI have been polluting social media since the beginning. Spambots started out as user accounts with rudimentary AI — just lists of things to say (normally with dodgy links) and a command to randomly post them. Now, they use GenAI to interact with human users in sophisticated ways. The difference between this and one of Facebook’s GenAI users is… subjective.

But, so is the difference between GenAI spambots and GenAI versions of influencers, and there are people happily interacting with those as you read this. I don’t know what fans get from talking to AI versions of their favourite internet-celebrities — it’s not my bag — but they get something out of it. Which means Facebook’s troubles might have more to do with the way it announced GenAI users, than the fundamental fact that it did.

And that brings up something horribly ironic. If OpenAI successfully monetizes through an ad-based social media platform, then it’ll be an even duller revenue engine than the information retrieval that’s happening in healthcare. Except healthcare helps people when they’re at their most vulnerable.

I hate to end with “capitalism values the wrong things”, but come on.

Share with friends

If you liked this article, think about sharing it with anyone who ought to subscribe.

https://www.benjclarke.me/p/the-value-of-boring

Looking for unbiased, fact-based news? Join 1440 today.

Join over 4 million Americans who start their day with 1440 – your daily digest for unbiased, fact-centric news. From politics to sports, we cover it all by analyzing over 100 sources. Our concise, 5-minute read lands in your inbox each morning at no cost. Experience news without the noise; let 1440 help you make up your own mind. Sign up now and invite your friends and family to be part of the informed.

Subscribe to 1440 today.