The K-Shaped Future of Work

this is the K-shaped future of software engineering. But I don't think it's just software engineering, Neil. I think we're talking about just many, many white-collar roles in general. So this graph over here, or this chart, I should say, in the blue, this is senior software engineering hiring. Okay, are you following?

Yeah.

And then red is junior engineering hiring. Okay, so juniors are going down, seniors are going up. And so when someone says the K-shaped future of this, the way this is, this this article starts off, it's developers are cooked. If machines can code this well right now, what's even the point of engineering? But we're seeing the same things for marketing across the board right now, if all this can be done. So my whole thing right now is I'm just gonna, I think, to motivate people to get off their butts, I'm just gonna be making videos of me doing these tasks in 20 minutes or so. And if I can get it 7 or 8, like, wake up, wake up, guys.

But the hard part is, is They just have to understand that 7 or 8 isn't good enough for customer work or even your own work because everyone's using AI and everyone's creating 7 or 8s. You have to figure out how to get to 10. And what this does, and Eric's not saying you should settle on the 7 or 8, it's you take the 7 or 8 and then you go spend your time to make it that 10.

That's what we call taste. Yes.

And that 7 or 8 that AI did helped you save a lot of hours, so you should be able to get to 10 much faster.

Yep. And here's the kicker here. So he says, look, consider 2 teams, right? So engineering is not being— it's being displaced but not replaced, okay? So you have Team A and you've worked with engineers like this before, okay? So Team A, Team A cares about impact, not activity. They handle ambiguous problems without paralysis. They understand the product, business, and the data, not just the code base. They design high-leverage systems and actively reduces complexity, that highly reduces complexity. They adopt new tools quickly and apply them with taste. They obsess over the user experience and they find creative solutions to hard problems versus Team B, Neil for engineering, okay? Team B, they build before understanding the problem. They fixate on performative code quality. They bike shed details instead of making users happy. They complain about headcount and stretches timelines. They add process when things feel chaotic, and they debate libraries and design patterns as a form of procrastination. And so Team B is going to have trouble in the future, whereas Team A, the ones that have the high agency piece that we're talking about, the taste, and just the ability to get things done quickly, those are the ones that are going to win.

I agree with you. Yep.

All right, you pick one. We have about 9 minutes left.

Let's talk about what to do about the AI.

Oh wait, wait, Neil, you'll like this one. Before we do this, slopocalypse, slopocalypse. Okay, thank you. So Procter Gamble, I was talking to someone recently. I don't know if we talked about this, but I think he, oh, he's going through a process right now. So he's selling his e-com company, okay? So Procter Gamble is trying to buy his e-com company. And so he's talking to Procter Gamble, he's like, You know, he's talking to his founder. He's like, you know, at Procter Gamble, they own what? What brands does Procter Gamble have? Remember?

Dude, they have so many brands. Diaper brands like Pampers, Gillette. Okay. Toothbrush, Oral-B is Procter Gamble. I'm blanking on brands, but let's see. Okay. Olay, Pantene, Tide, Downy, Charmin.

Great.

Braun, Vicks, Crest.

Okay, so a ton of brands, right? So guys, the owners of Crest, Tide, Downy, Procter Gamble, they told a founder who they're in the process of buying their company right now that, look, what we've learned over the years is we don't buy remote companies. And the founder's like, why don't you buy remote companies? They're like, we've tried buying so many of these remote companies, meaning that the culture is remote, it's remote work, and it doesn't work because When we buy them, there's no real culture, right? Versus when we buy the in-office companies, those are the only ones that work for us. So they're saying, hey, from an M&A standpoint, that's— this is not for everyone necessarily, but at least from what we've seen, the remote ones never work. And I think that's an interesting discussion point because you've bought, you know, a handful of companies these past years, right? And so, but you most in-office, like what's the percentage of in-office versus remote and what are the differences that you've seen?

So the biggest thing that I learned with in-office versus remote, it's easier to bring people up to speed when it's in-person versus remote. By far, I would say that's the biggest advantage because there's way more knowledge transfer that happens when people are next to you.

Which is interesting. So you say that, watch, Neil. So when you say there's way more knowledge transfer, doesn't that just mean the knowledge transfer in general is a lot faster?

Yes. And you're seeing that the thing that's valuable about in-person is knowledge transfer from seasoned execs who have been doing this for a long time. It transfers to younger people way quicker than ever. Yep.

I'll tell you, we just had a new senior brand manager for my YouTube start, right? And the way I'm like, I'm talking to him every day, but man, if we're in the office, we'd be probably accelerating like 4 times faster or so. But I'm just like, I'm like not letting my grip off. I'm like, no, you're going to learn all these things. I'm bringing you with me. I'm going to make you great with AI as well. I'm going to make you great on content side. But I just think it's interesting that that was an insight from Procter Gamble because they buy a lot of companies.

They buy a lot of companies. So they like everything that's in-person versus non-in-person.

Yeah. They've learned to no longer buy remote companies. Just doesn't work for them. Yeah.

Yeah. And I wouldn't say that you can't build a remote company because there's a lot of successful remote companies that are worth a lot of money. It's more so you have to figure out the culture of your own company and what works for you guys. And sometimes when you bring in something that's very different than what you guys are used to. It's just a disaster.

Here's the problem with most marketing agencies. They don't understand what is happening with the world of AI. With my agency, Single Grain, we think about agentic workflows. We think about creating agents for the future. When we think about SEO, we think about creating different products such as internal linking products or content consolidation products. And on the other side of things, we think about building products such as Care. We just think about making marketing easier overall. If you want to stay at the cutting edge of marketing and at the same time leverage the power of AI, go to singlegrain.com and we'll see you on the other side. Yep. Before we move to the AI slotpocalypse, so this one's a quick one, but 85% of AI search results come from third-party sites. So originally everyone's talking about making these listicles and publishing on their sites, right? But it's not hard for these LLMs to quickly figure out, oh, they're just publishing on their own site. So this is published from AirOps and he's like, we've analyzed thousands of AI search results across hundreds of brands. The content on your website accounts for roughly 15% of why you get mentioned or cited in AI responses. The other 85% comes from everywhere else. So partner sites, news articles, third-party roundups, Reddit threads, whatever. A lot of this, guys, is guest blog posting. A good chunk of it's guest blog posting back in the days.

Uh-huh. And Eric and I called this, remember we talked about this like 6 months ago being like the problem with ChatGPT's algorithm is they, you can pull a listicle on your own website and get cited and eventually this is going to go away because it's like the old way of spamming Google and stuff like that. I get it's different, but it's similar to how easy it was and we're like, this isn't going to last.

Yeah, it's too easy to game. And when it's too easy, then it is what it is. So this AI Sloppocalypse piece, so this is from Allie Miller. So this was a disagreement between Andrej Karpathy, so this strong AI scientist, and also this guy Boris, who actually is the creator of Claude code. Okay, so they actually were disagreeing about this. So they were just saying, look, when the cost of generating anything at any quality level plummets and there is a— there's still market leverage for attention, generators will take advantage of it. It's why one company is cranking out hundreds of Reels/TikToks per day. It's why another is cranking out 3,000 podcasts per week. Code might improve, but the content and social and research worlds are rougher waters. I too predict a sloppocalypse this year, one that Mosseri, I think that's the CEO of Instagram, I'm not sure if I'm pronouncing that right, and Presser and Ek, Daniel Ek from Spotify, will be thinking about daily. The smart generators are already thinking about second-order impacts. Examples for content creators, if everything is AI-generated, maybe I should only post human things. If everything is AI-generated, maybe I should have a separate community that's off socials. If everything is AI-generated, maybe I should host more in-person events. And then if everything is AI-generated, maybe I should make a social media app that is only for AI bots, right? So there's a lot of these AI if situations. The big question for me is less on quality because it will get there, right? So we talked about like, What was it? It's already a 7 or 8 out of 10 for these like advanced copywriting. And I think you and I have fairly high standards. So a lot of people, it's probably 8 or 9, if I'm being honest. I'm thinking—

Some people, it's probably even 10s. They're like, well, this is great because they couldn't write their own copy.

Yeah, so I think we're particularly hard when it comes to copy, right? And psychology. But the big question for Allie at least is less on quality because it will get there. I'm thinking about whether people continue to engage, whether people will continue to engage with AI-generated content. Whether they build affinity for AI creators and at what level compared to humans and whether they rebel against it and at what speed. So that's the coming slowpocalypse. And I think you and I can both agree that there's marketers especially, we're going to take advantage of slowpocalypse.

We are, but I believe this world is changing where people are fed up with AI content. A lot of the younger generation are stopping to use TikTok and platforms like that because they're overwhelmed with the AI content and we're seeing it and it's just becoming really ridiculous. And one, a lot of what's created is just fake. Two, people are just fed up with it. So much so, if you look at Sora, no one talks about Sora and it's not doing that well anymore.

Dude, you know who uses Sora? My entire building, the people who work in the building, right? They're like, I was talking to one of the guys who's one of the food runners and I was like, what are you looking at on your phone? Because I hear all these noises, right? He's like, check this out, dude. And he's like doing the Sora and he's like, the picture is like of another employee that works in the building and they're like doing Soras, like making fun of each other. I'm like, So you actually use Sora too? He's like, oh my God, it's so big, blah, blah, blah. So I think you and I are just out of touch.

No, have you seen the recent articles on Sora?

It's probably the usage is down.

Usage is down a lot.

I used it for like a day and that was it.

Yeah, there's a big financial site that just wrote about Sora. Sora usage down. Where is it? Okay, as of Jan 20— this is from AI Overviews. As of Jan 2026, the usage for OpenAI Sora app has experienced a significant decline following a strong initial launch, with reports of a 45% drop in monthly active installs in January 2026 compared to the previous month. That's a massive drop.

Did you see the graph where it shows Anthropic is clearly overtaking just OpenAI? Oh, Ramp published something. Okay, so Ramp published, they showed the spend per company.

Yeah, let's go back. Ramp is a credit card company that shows you where— Ramp is a credit card company that where corporations can get a card just like they get a Visa card or Amex card and allows them to spend money, you know, for their corporations.

Yeah. And I'm sure— and thank you for that. Okay, so this guy Ara, so this is actually really good. You'll like this. So Ramp, he launched this vendor section where he shows vendor spend, right? Check this out. I'm going to pull up— let's pull up SEO and AEO intelligence. So this— because they can see spend for companies, right? So you can see over time Semrush, they were up here like Q1 of 2024 and then they're kind of going down on SERP API. Because all these people are building on top of it, is going up and then Profound's up here and they're doing a good chunk, right? And you can see market share like SERP API is going up 10% adoption rate. Semrush was 73% on these SEO tools and then Ahrefs is 23%. I actually haven't— this is pretty cool.

This is pretty cool.

Right? For sure. I'll send it to you after. I knew you wanted to take a look at it, but they actually showed the difference between— check this out. So when you look at— oh my God, stop trying to sell me stuff.

Okay.

So when you look at generative AI, I think that's the one we'll click. Generative AI. Okay, so, oh my God, look at this. Who was this before? This is OpenAI. Look at this. Q1 2024, it's like 90%, right?

And also keep in mind the markets increase as well. So you can't just look at the bar meaning someone dropped in revenue.

Fair point.

Eric's not trying to say that either.

Yeah, fair point. So we're just, I mean, you know, market, this is just market share over time. You can see adoption year over year as well. But then you see who comes in here. you see Anthropic starts to really jump in here and it's gonna continue to increase. And so our point of saying all this is that I think when we're talking about OpenAI, I find myself using the Claude app and Claude Code a lot more, and I find myself using ChatGPT a lot less now. I don't know about you. I don't know how you're using Gemini quite a bit, it seems like.

I use Gemini quite a bit. I use ChatGPT for health probably like for, 3 minutes, no more, 10 minutes a day.

For your health?

For my health. I think it's amazing. It tracks all my food to eat. My kids have this new thing. They started it in late December. I don't know if I ever told you or did I tell you? No, I know I told Mike. So my kids in public, I'm a pretty skinny person overall, will start pulling up my shirt and be like, look at this belly, so cute. And then they put little fuzzies in my belly button.

That's funny. What's a fuzzy?

Like a fuzzy, like, Cotton or like lint or whatever it is. And they started, they went from doing it at home. I'm like, what the heck? So my daughter started doing it first. I'm like, where the heck did you learn this from? So then my son starts doing it because he's copying her. She's older. So then, you know, it started snowballing from there and I keep telling him to stop. And I'm not one to like hit my kids or anything like that. Like I'm not going to put them in a timeout for saying, look at your belly and like poking at it and stuff. And I'm pretty skinny. Like if I wore a shirt, you would see nothing. I can pinch an inch. So it's not really So then they started doing it in public and I got fed up with it. I'm like, okay, I'm going to lose this pinch of inch. So I started keeping track of Claude, not through ChatGPT, my daily calorie deficit.

Okay.

Because that's the easiest way to burn fat. Noah would know this more than anyone else. Is that correct? So then from December 29th to what is today, Jan 30th, because I was so fed up with it, I've dropped I went from 131 to 126.4.

Isn't it funny? All the time you spent losing weight instead of disciplining your kids.

I tried trying to convince 'em. I couldn't do it. So it was just easier. I'm like, you know what, forget this. I'm just gonna get rid of my pinch inch. I probably have like another 2 pounds. So another 2 weeks and I'm done.

Yeah.

Okay. And then after, I will not use ChatGPT at the same time.

And then you'll find something else to get you. But okay, the last thing we'll end off with here on the slopocalypse piece is, you know, I think this is, We're going to see a lot more of this. I think people like these conversations, right? And I get a lot of good comments on this podcast, but I think in-person events too. I think people making content saying this is 100% human-generated, I think that's going to do well. I think people making content in different languages. So I think this lapocalypse, I think a lot of people are going to take advantage of it. But I do think to Neil's point, a lot of people are fed up with it. And that's a good thing. I think it's good for people to be— and go outside, right? Go touch grass or something.

Yes. Yes. People need to live normal lives and interact with other humans. That's the way to do well in today's world in business or anything. I think people are like forgetting this interaction. I saw something from Marc Andreessen, which was a really good interview. It was published somewhere on the internet.

Lenny's podcast?

Yeah. He was talking about how developers believe they can do product and—

Oh yeah. The Mexican standoff.

Yeah. Mexican standoff and how product and designers think they can do development and the other features as well. And I think he's very spot on, but you still need in an organization the people who are exceptional at development or the people who are exceptional at design because AI isn't there fully yet. And even if you think, okay, AI is going to be there and it's going to write everyone's copy or AI is going to do everyone's development. Well, if you can't figure out how to do it better than what everyone else is using and stand out, you're going to fail. And I'm better at giving an example for marketing than I am for development because I'm not a developer. But in marketing, if AI is writing everyone's website copy and boosting conversion rates, that'll be the new norm. It won't convert as well because everyone sees highly optimized copy. You then have to go and figure out how to pivot and make your conversions better than everyone else. An example that everyone can relate to is if you look at ads before, ads before where purely, or majority of companies would strive for Super Bowl quality ads. Would you agree with that? That's what people really wanted.

Yep.

Now, if you look at today's advertising on the social web and even sometimes on TV, you're starting to see a lot more lo-fi stuff. Why? Because it converts better and people are tired of the over-polished stuff. And I think that's what's going to happen in AI in which you're going to have all these people using it for the same thing. You're still going to need humans. that are the best in class at what they do, you combine them with AI, you're going to crush your competitors.

You just have to find them. All I'm going to say to end it is as long as you have a bias to action and you're curious, you will crush everyone around you. That's all I'm going to say. And we would like to hire you. So both of our companies are hiring people like you. So let us know and show us your bias to action by showing us some examples first. Just don't give us a random reach out, please. Yeah.

Don't send us an email being like, you're doing this wrong. We want to know how we can do it better and show me a real example.

Or, or like, just like if you want editing work, like do like one test edit. Okay. It makes you stand out faster.

But before we go, dude, does it irritate you? Because for me, I get so many editing emails on a weekly basis and none of them ever show any examples.

None.

And if they do show example, it's an example of some random other account. And it's like, you could have used AI and showed me an editing example of my own stuff.

If you just go the extra mile, you're going to run circles around everyone. That's all we'll say. And we would love to work with you. All right, bye.

Take care.