It’s a Bot World Now

I am human. You’ll have to take my word for it, though some people can vouch for my existence.

Why do I feel the sudden need to advertise my humanity? Well, I want you to know I’m not a bot. As it happens, reported yesterday, Cloudflare finds that 57.4% of requests to various websites that it hosts are automated bot requests, with only the remaining 42.6% of requests generated by humans. The tide is going out (or coming in, depending on one’s perspective). What's more, current evidence strongly suggests that it isn’t coming back.

As discussed in the article, the potential ramifications are profound. A debate rages as to whether the proliferation of furtive bots sounds a silent death knell for human-to-human web content. Called the “dead Internet theory,” this hypothesis asserts that the online universe will eventually become the near-exclusive preserve of agentic communication, a sterile, symbiotic — arguably parasitical — wasteland of data vampires.

Not everybody sees a bot dystopia on the horizon. Matthew Prince, co-founder and CEO of Cloudflare, contends that the autonomy of agents is allowing practically anybody, irrespective of technical proficiency, to hang out a shingle and unleash latent reserves of creativity. Maybe so, but will a sufficient number of humans still be traveling online highways and byways, stoping along the way to appreciate and enjoy what the liberated creators offer for the delectation of visitors?

Here’s a quote from Prince, excerpted from the article:

“I think a lot of people kind of have said, ‘Well, this has proved sort of the dead internet theory.’ I think that’s actually kind of wrong at a lot, a lot of levels,” he said. “You don’t need to be a web designer, you don’t need to know how to program, in order to create these things anymore. It’s given access to content creation to a much broader audience.”

Agentic Feeding Frenzy

Well, yes, but the broader audience, such as it is, might overwhelmingly comprise freeloading bots. That could put a crimp in the new age of creative commerce.

Speaking of which, Prince himself subsequently mentions challenges to the internet’s established business model. Bots, he observes, do not click on ads. Bots also do not, in any sentient way, appreciate or enjoy content. What they do is collect and mine it.

Prince believes a business-model adaptation might involve charging bots for access to digital users’ content. If — and here he puts a heavy load on that little qualifier — the bot toll charges are implemented effectively, Prince believes we could be “on the cusp of the golden age of the internet.” Heady stuff, that. In defense of Prince, I will say here that we’ll probably have to reconcile ourselves to a bot-infested internet. Resistance, as the saying goes, is futile; charge is hard but inevitably necessary.

In our headlong rush into the inescapable future, however, it is neither maudlin nor nostalgic to look back, occasionally with great fondness, on what we’ve left behind. We can’t live in the past — we reside in an ever-unfolding present — but we should remember that what came before wasn’t all bad. The technology industry optimistically fetishizes a shimmering future that is never fully realized in the eternal present. I understand the imperative of optimism as a motivational spur and a business accelerant, but I also have come to learn that willful optimism dims the lights that help us see the obstacles and potholes in our path.

The Hyperscale AI Edge

Despite the best efforts of IT infrastructure vendors, such as Dell and HPE, to proclaim the rise of private and hybrid AI — the successors to on-premises private and hybrid clouds — AI was always destined to play to the advantages of the IaaS cloud giants. Nobody else has the massive capital — for globe-girding datacenters, bottomless IT budgets, rapidly iterative product and service enhancements — needed to win a market war of attrition at prohibitive scale.

Don’t get me wrong. There are some enterprises who will deploy and operate AI in their on-premises datacenters, but for most others, the overall appeal and perceived risk mitigation of running AI at AWS, Microsoft, and Google are compelling. That’s not merely the conclusion I’ve reached, it’s also the considered view of UBS analyst Karl Keirstead, who finds that the hyperscalers might have more AI runway ahead of them than investors currently believe.

From an article at MarketWatch:

“The medium-term “upside in terms of revenue growth from this broadening-out of AI revenue sources is underappreciated by [Wall] Street and could drive [revenue] estimates and the stocks higher,” Keirstead added.
The hyperscalers’ beneficial relationships with OpenAI and Anthropic could already be seen this earnings season in their explosive cloud and backlog growth.
“It is abundantly clear to investors that the hyperscalers are getting a massive boost from compute-hungry model providers,” Keirstead wrote.

The article later notes that UBS’s market checks found that “enterprise AI budgets are shifting toward hyperscalers because customers want the flexibility of a third-party platform to try different models.”

Despite the AI hype, which now seems like it has been a constant presence since the beginning of time, we remain in the early innings of enterprise AI adoption. Most enterprises are still trying to figure out how to derive meaningful value from AI; they’re a long way from being able to assertively quantifying anything approximating ROI.

Enterprises are still experimenting, trialing, evaluating different models and methods. How (and how much) they utilize AI remains to be determined. Given the tentative and exploratory nature of these initiatives, taking their AI business to the hyperscalers, capable of providing flexibility and variability, makes sense.