Lately, I’ve been watching some videos from Matthew Berman and I’ve realy enjoyed them. They’re great deep dives into various AI topics.

Most recently, I watched his take on Meta CEO Mark Zuckerberg’s statement on Joe Rogan’s podcast that AI will soon be writing all the code for Meta’s apps, and “that will augment the people working there.” There are a million predictions floating around about how exactly AI will impact jobs, particularly those white-collar jobs that many have been chasing for the past couple of decades. It’s hard to tell which, if any, of these predictions will come true.

In my experience working with AI so far, it can definitely make me faster and better at my job. I produce content for a living and I’ve been working on creating a good local copy editor, with good results. On a daily basis, I can definitely work less for the same results, and I’ve enjoyed the process of refining my Copy Editor digital worker.

However, productivity is a tricky thing. If everyone just gets faster, what exactly does that mean? For example, in the area I live, it seems like 1 in 3 cars is a Tesla. By comparison, I drive a Jeep Gladiator.

Here’s a breakdown of those two cars:

Tesla Model 3 (2022)

  • Acceleration (0-60 mph):
    • Performance trim: 3.1 seconds
    • Long Range trim: 4.5 seconds
  • Top speed: electronically limited to 125mph (201 km/h)
  • Quarter mile time (8.04 seconds) for the Performance trim

Jeep Gladiator (2022)

  • Acceleration (0-60 mph):
    • Sport S trim: 6.1 seconds
    • Rubicon trim: 6.4 seconds
  • Top speed: electronically limited to 108 mph (174 km/h)
  • Quarter mile time (14.3 seconds) for the Sport S trim

Breaking down those numbers, a Tesla Model 3 has between 40-93% faster acceleration, a higher top speed, and a much faster quarter-mile time. However, if I’m driving next to one, and we’re both commuting from the suburbs where I live to the office park where we work, we’d get there in the same amount of time.

My point with this comparison is that I’m faster at my job, but every bottleneck that existed before generative AI took off still exists. I’m faster, but I don’t actually get work done faster. I can theoretically get more work done, but the pipelines that lead to and from what I do are still the same.

Without systemic and organizational incentives and structuring, it may not matter if individual productivity goes up. This also means that someone needs to be steering these AI initiatives to ensure that work moves at a pace that is worth the investment.