Is Middle Management the Real Job AI Replaces?

Brian Maucere (0:00)

Hey, everybody. What's going on? Welcome to The Daily AI Show. Today is October 8th, 2026, and I'm joined by Andy and Beth.

Brian, glad to see you all here. We're here every single weekday at 10 a.m. Eastern. So make sure you come back and hang out with us if this happens to be the first time you found us, which happens every day. So we're always happy to see our new folks.

If you want to join us live, the easiest way to do that is through YouTube.

Like I said, we're 10 a.m. Eastern. So if you happen to be around during that hour or happen to have a day off, come join us live. Be part of the community. Be part of the folks that are adding to the show while we're actually doing it. You'll often hear us talk about or bring comments up on screen. So if you want to be part of the show in that way, you can. But if you don't have time during the 10 o'clock hour, that's certainly understandable. Another great way to do that is to go through our community. So you can go to the dailyaishowcommunity.com and over there, you can just sign up. It's a free Slack community, and it's a great way for you to join the conversation. More asynchronously on that side of things. So there you go. There's all the ways that, hey, if you like what we're doing, a like or subscribe along the way goes a really, really long way for us. It just helps us reach more people. I was just sharing with the team the other day, what our September numbers were for Spotify, and yeah, we continue to grow at a really consistent pace with some speed ups along the way on there. So I can remember on Spotify being really, really happy if we bounced over 100

They've changed the metrics just like YouTube does, so it's not always the same, but it was 100, it meant like 100 views, right, or 100 listens, or whatever on Spotify, it was 100 And if I saw like, if we were at 89 and like, we had a show where we popped over like 101 or whatever, I got really excited. These days, if it's under 500, I'm like, ooh, I wonder how we missed on that show, but usually it's the title or something. So it just goes to show you can get used to anything. You can get used to any numbers, but I do try to remind myself that once upon a time, when we started this show, seeing anything over 100 was really great. And nowadays, if it's 500, it's a normal show. And sometimes it can bump to a thousand or more, just depends on what the topic was and everything else. So anyway, thanks everybody for being here. Let's dive into it. I know we have some new news coming out about some of the, the models and things like that. So I wasn't here yesterday. So with that, I'm going to kick it over to you guys, because I'm not quite sure of what you guys spoke about yesterday.

Andy Halliday (2:30)

I'm going to talk a little bit here about jobs and the up and down question of whether AI is a job displacer or not. Because yesterday, the head of AI at Microsoft, Mustafa Suleyman, the person behind Pi, the original emotional AI-tuned thing, which was acquired by Microsoft, really to acquire Mustafa Suleyman.

I think he has either his bride or a partner or somebody else was also done in that acquirer. Anyway, yesterday, he wrote on X, a post that kind of shifted his perspective on this toward the endorsement of and promotion of a recent essay that was written by Daron Acemoglu, who is an MIT economist and who is widely regarded as a smart guy in the world of economy. But the essay basically says that AI will replace only about 5% of what humans do over the next 10 years.

Now, that's very different from what it feels like when you see so many people in leadership positions inside corporations saying, okay, we want to automate things and we want to reduce our head count. And that actually happens, right? And we have plenty of numbers associated with that head count reduction, particularly in areas where AI has a serious contribution to knowledge work. And so a lot of people are worried that AI will take their jobs and there's a resistance that builds in the workforce against AI because of that. And Suleyman is shifting because just in May, he said that AI would reach human-level performance in most professional jobs within 12 to 18 months. So that just empowers the notion that, hey, all of these knowledge work positions can be replaced by AI. And increasingly, we are now treating AI agents as colleagues in the team who can actually be delegated to. And I have another note that I want to talk about, which has to do with Goldman Sachs. And I'll come to that in a moment. But anyway, Daron Acemoglu, the economist, says that AI developers should stop building AI automations and agents to replace people, but rather shift their focus instead on making humans more effective using AI. And that makes sense, of course.