I had been covering the Internet for 9 years by late 1994. I like to say that’s the year the Web was spun.
My first Web job was for Interactive Age. This was a short-lived magazine from CMP Media meant to cover infrastructure. They were focused on the two-way cable network Time Warner was building in Orlando, figuring that would be a model for the future.
They were wrong. As I reported on the Interactive Age Daily, it was the Web itself that was the revolution. The money came into services faster than it went out for infrastructure. The magazine failed within 9 months of launch.
Everyone knew, right away, how to get value from the Web. Well, almost everyone. I wound up in arguments with the art director, who wanted to force people into visiting her beautiful home page instead of just bookmarking my text-based news feed.
Then, the money came rolling in from every side. The “killer app” was obviously selling stuff. It was the infrastructure behind the sale that lagged and turned the dot boom into what I called the dot bomb. Companies like Webvan were selling fresh products that people weren’t yet ready to buy sight unseen. Many companies bought ahead of their ability to adapt to the new tools.
While the resulting crash hurt, it really did turn out to be for the best. It gave things a chance to shake out. It gave companies a chance to see where the value was, to improve their productivity, and to change how they did business. Eventually it created a new boom called database computing.
AI is the product of that boom. AI is different.
The AI Error
With AI, investment is front-loaded in infrastructure. If there were still magazines, an AI Age with ads from Dell and Super Micro would be doing quite well.
It’s enormously expensive to upgrade data centers and build valuable AI tools. Fortunately, the Cloud Czars have the cash, earned from their previous booms, to handle that load. But at some point, that money must come back from software and services that improve productivity.
There are positive knock-on effects from the Czars’ spending spree. We need better networks, we need more green energy, and we’ll all need new clients. This should keep things going, only with different winners, for a few more years.
But at some point, money must come rolling back in to justify today’s investment. The “AI Web” isn’t the original Web, where customers were spending, getting returns, and spending more. It’s driven by suppliers instead of customers. Until we see value in AI applications and buy them with both hands, a bust may be just around the corner.
So far, the returns are mixed. How many companies are still buying Co-Pilot? How many people will pay for ChatGPT or Gemini services? What is the growth rate at ServiceNow, at Salesforce, at Adobe, and at C3.AI?
It’s clear that, so far, AIs provide first-level assistants. We’re not replacing analysts or even programmers. We’re giving them a skills upgrade, so they can do more, and hopefully do it better and faster. That’s an evolution. Where’s the revolution?
Maybe Apple will tell us next month or Amazon will this fall. Maybe Meta will tell us in their next earnings release. Their bet on open source AI is a lot more consumer-oriented and global than its peers, and you can only make so much money from ads.
One More Thing…
Governments are more wary of AI than they ever were of the early Internet. We know AI could sway elections, and we know governments now have an instinct toward regulation. While the early Internet had a bias toward freedom, AI so far has a bias toward control. Crooks and liars are finding value in AI tools well ahead of consumers and salesmen.
Conservatives may blame an AI bust on government, but without ordered liberty all you have is anarchy, which evolves toward fascism. This ain’t your daddy’s Internet. It’s a completely different beast. One that’s red in tooth and claw.
