Too Much Stuff

Many organizations that struggle are not failing from a lack of stuff. They do not fail because they can’t generate analysis or draft documents. They are struggling because the analysis does not drive a decision, or a decision does not have a link to execution and it is hard to get a clear picture of what is actually happening across functions.

AI accelerates the generation of stuff, such as reports, code, content, contracts, analysis. And we should clarify what we mean when we say ‘AI’: The kind of AI that we are talking about when we say AI today, that is to say some sort of Generative AI. So of course it excels at generating stuff.

But AI doesn’t solve the coordination problem between humans who have different incentives, different mental models, and different definitions of success.

So at the risk of degrading the Signal To Noise Ratio around the topic of what Microsoft’s AI CEO Mustafa Suleyman said in the FT interview on Wednesday, here are three ways we might think about headlines like “Virtually All White Collar Tasks Will Be Automated Within a Year and a Half”:

It is a conversation initially about Microsoft Stock Prices.

That’s a backdrop to the whole conversation. In fact, it would be helpful to imagine the interview being conducted in front of a stock price line chart for the last 5 days (at time of writing) on a screen behind them.

Keyword Task.

Jobs are bundles of tasks. So it is easy to read the headlines that the interview generated and think ‘Oh no, how will I pay my mortgage?’. If it bleeds, it leads, right? But remember, we are talking about tasks being automated, not jobs. This is where the notions above come in - yes, 100%, tasks that make some kinds of Knowledge Work stuff are being automated. Go get a subscription, try it and see. You can tell (insert favorite model) to read your notes and write a report. It can and will make lots and lots of stuff for you.

There is a genuine pressure from investors to, basically, have companies replace people with robots.

But that does not mean the market will get what it is asking for.

Folks like Daniel Miessler have written about this pressure. It is a real ‘ask’ from the investment community that we can hear. It’s a strange ‘You know we’re standing right here, don’t you?’ situation and it is worth looking at.

Start with the investor side. The fundamental problem facing Big Tech right now is that they’ve committed roughly $300 billion in combined capex to AI infrastructure over the last two years, and the revenue story hasn’t caught up. Microsoft, Google, Meta, Amazon are all building data centers at a pace that makes the fiber optic buildout of the late ’90s look restrained. The market gave them a long leash through 2024 and into 2025 on the strength of the narrative alone, but as Suleyman’s own interviewer noted, investors are getting nervous. They want to see where the returns are.

And here’s where the labor elimination narrative becomes structurally necessary for the investment thesis. Because the honest productivity story — AI helps knowledge workers do certain tasks faster, maybe 10-30% efficiency gains on specific workflows — doesn’t justify the scale of investment. If you’re Microsoft and you’ve spent tens of billions building AI infrastructure, you need the addressable market to be enormous. And the largest addressable cost in every enterprise on earth is labor. US white-collar payroll alone runs into the trillions annually.

So what am I supposed to do with this?

Note we say ‘Labor elimination narrative’. It is a story that some folks in the market tell themselves and each other. I honestly to not know if the story is going to be true or not. But the headlines rarely help us make sense of things. So if the headlines are scary, here are three things you can do to (hopefully) address some of you concerns.

1, Go to the source. Ignore the spin off headlines that say someone said ‘X’. Go see what they said.

2, Pressure test words. What I mean by this is see what phrases are carrying water. In this case Tasks does not mean Jobs.

3, Understand that a conversation about stock prices tends to go in only one direction.