Does AI Need Revenue Already

I saw a post once — I can't remember who wrote it — that defined a difference between Britain and America. In the latter, if you drive for an hour, you'll end up in basically the same place; in Britain, however, the scenery will be new, the accents have changed, and bread rolls are called something different.

I haven't spent enough time in America to attest to this, but the British part is spot on. Accents and dialects change quickly across our modest landmass, and besides a small strip of flat farmland in the east, our geology changes rapidly as well. Car travel can be an absolute dream — watching the hills roll by, the trees and shades of green changing, the various birds of prey hovering in the sky. Obviously, this necessitates being a passenger.

There is nothing pleasurable about actually driving in Britain. Or anywhere, in my opinion. Of all the normal things that adults are expected to do in life, sitting behind a steering wheel is by far my least favourite. I fully intend to buy an autonomous vehicle at the earliest opportunity. In fact, I'll push women, children, and the elderly out of the queue in front of me.

What I won't do is pay for AI that I don't want, and didn't ask for. This might seem like an awkward segue, but the themes combined recently in the shape of Google — or Alphabet, or whatever the tax-efficient structure is at the moment. We've been watching the company smoulder for years, and seeing Google’s Waymo cars burn was a fitting symbol for many of us in the tech sector.

It's difficult to articulate how unassailably dominant Google once appeared. By publicly asserting itself as the only good search engine, and quietly embedding itself as the tech behind online advertising, Google had won the data harvesting battle. Nobody had as much consumer data as Google did, and nobody was leveraging it as effectively. Peerless was the only fitting word.

But the years since have been a mixed bag. Social media, online shopping and smartphones became new data sources, and Google only competes in the latter. Its adventures in consumer hardware are faltering, and its mammoth push for cloud tech has yielded just 12% market share. AI is the obvious new battleground, and the Waymo vehicles are Google’s high-end showpiece.

At the consumer end, I got an email from Google saying that the price of my Workspace subscription — business email, video calling, calendar, office suite, etc. — is changing in light of the added value of Google's AI. Bollocks!

I have never used Google’s AI for anything. And when I wrote an AI article a few weeks ago, I almost forgot to mention it. This is despite it being tacked on to every Google service like a cross between the Microsoft paperclip and a particularly virulent herpes virus.

I just don't want AI in my Workspace. That Google is intending to charge me for this unwanted AI isn't merely annoying (and terminal — I'm switching providers), but also quite telling.

I've long maintained that there are two AI sectors. One concerns “techy” use cases, like autonomous vehicles, coding, and analytics. The other comprises “consumer” use cases. The problem is that although the underlying technology is the same, suitability isn't.

Consider two uses of the new LLMs. The first is me searching several thousand medical letters for information. This is a task that LLMs perform better than any other tool, but like all tools they make significant errors. Fine, I can handle those errors because I'm a professional with a deep knowledge of healthcare data — i.e. I know what I'm doing.

If, however, I send an LLM to search my email inbox — a typical consumer task — and it fails to surface a credit card reminder, the bank won't accept “AI error” as an excuse. AI is a fantastic boon in a wide range of professional contexts, but it isn't proving to be a particularly good addition to many consumer services.

A question that has gone unanswered for the last eighteen months, or so, is 'why did Big Tech snowball so much money into the consumer side without any evidence of returns?' We don't know, yet, but market realities seem to be forcing some kind of answer.

Meanwhile, a question: does Google's ham-fisted attempt to extract AI revenue say more about the industry at large, or just Google's stagnant empire?