The Robert Moses paradox, part 2

The tl;dr version of yesterday’s post is that by adding capacity to the roadways, traffic gets worse. We’re seeing a similar dynamic play out in the workplace now. AI was supposed to make us more efficient. Arguably, these tools are allowing people to create more output — extremely dubious whether this equates to value — but are they also saving workers time? The answer clearly seems to be no.

In Ed Zitron’s latest newsletter he pulls a quote from Bloomberg about a guy who believes himself to be insanely productive, at the negligible cost of spending every waking moment fretting about his agents.

“Matt Van Horn, a serial entrepreneur and father of four, never turns his laptop off anymore. He has more than a half-dozen artificial intelligence agents running at all times in Anthropic's Claude Code.

”Every 10 minutes or so, they ask him what to do next.

”He keeps his laptop running at his kids' soccer practice, while dropping them off at school and in the hotel during vacations.”

This is deranged behavior. Obviously. But it tracks with what seems to happen with almost any use of AI. The promise is hyper-productivity thanks to PhD-level intelligences at your command. The reality is an ever-growing stack of processes that require studious micromanagement, check-ins, and corrections, all to create stuff that is arguably worse than what more traditional methods might have yielded. The only real argument is that, yes, these things do seem to enable workers to “do more”: produce more lines of code, generate longer PRDs, flood the zone with prototypes and artifacts that make them look like they’re helping.

The more wrappers and harnesses and whatever else that people have to graft onto the base technology in order to overcome its deficiencies only seems to result in workers having to spend more time working. This shouldn’t be a surprise. It’s been the story of every technological advance. I’m not the first to observe that futurists of 100 years ago thought by now we’d all be doing 3 hours of work a week. Instead, 9-9-6 is a thing.

The other unsurprising dynamic to those who pay attention is how quickly the dynamic changes. The machines are supposed to work for us; almost immediately we find ourselves working for the machines. Like the factory worker who must stand in exactly the same place and perform exactly the same action for hours on end so that the machine may run optimally, office workers are now contorting themselves to keep their Rube Goldbergian agent setups running. They’re so fragile. You must be on call 24/7 to keep the line running. This is not progress.

Homer Simpson said that alcohol is “the cause of, and solution to, all of our problems.” That’s a great definition for addiction, and it applies to how businesses are treating AI. We’re so convinced that it’s the solution to our problems that when new problems arise, the only prescription is more AI. We have to go cold turkey.