How AI is going to give us a different way to manage people

Something I think most of you have seen a lot from me is how I'm trying to think about AI and how we can use it to improve our lives - rather than make things worse and even more terrible than they already are.

Something else you might not know about me is I think a lot about good leadership, good management, and building high-trust teams, mostly because I've been on the opposite of a lot of that many times in my career. I think one of my core principles is to never be "that manager" for anybody.

So today I've been thinking a lot about this article from Lorin Hochstein - and how AI might change how we manage people and companies.

Make sure you read the article to get the background of Deming versus Drucker - but if you've ever heard the term OKRs, you're probably pretty familiar with the Drucker way of viewing management and business. The short version is that Drucker's way of viewing things through the lens of OKRs took off because:

1. It's very top-down.
2. It's pretty clear on what you need to do.

Deming's way is harder because it required a lot of statistics and mathematical overhead that required a lot of horsepower behind the scenes to integrate and run.

AI, however, is eliminating the bottleneck introduced by Deming's approach by allowing us to do statistical reasoning at scale and analyze data in a way that no single manager could track over time.

I think this approach could be incredibly valuable for early-stage startups. If you're at the seed stage, you don't even really have a company yet, much less a product. You're still discovering what things are going to be. A lot of teams set up OKR targets before they understand the connection between what they do and the outcomes they want.

Instead the approach is to build an instrumental learning loop - applying lowercase ‘a’ agility to all corporate systems.

I'm gonna expand on this a little bit more in a longer essay and map this out - how this could actually work within a digital mental health company. But right now it seems like being able to use AI agents to build internal systems at scale that allow you to directly measure and analyze systems in real time may end up being a lot better than quarterly OKR reviews - where a lot of times you are analyzing lagging indicators.