The Death of the Spicy Take

TLDR: AI is making our exec’s spicy takes feel mild. In this article we dive into why this is happening, what to do about it, and where the future of attention comes from if all spicy takes sound the same.

For those interested in reading the articles referenced in this blog, you can read them here, here, here, and here.

Was this newsletter forwarded to you? Sign up to get it in your inbox.

If you're responsible for crafting and managing executive messaging, it might be time to revisit your spicy take strategy.

Kara Redman recently described spicy takes as a glass of warm whole milk — meaning, once-bold opinions have curdled into a flat, predictable sameness across AI-drafted exec content. Once I was done LOLing at the comparison, I realized that she's right. Because when AI starts producing the same edgy contrarianism for every exec in every industry, the spice becomes dinner table seasoning and our appetite disappears.

So here we are. The spicy take, once a reliable lever for executive visibility, has lost its bite. Which leads me to the questions that comms Carrie Bradshaw would be asking herself right about now: are spicy takes dead? And if so, what's actually working now?

I'd argue something better is taking their place. But to get there, we have to talk about why this old playbook stopped working.

At surface level, the spicy take worked because it was rare. 2-3 years ago, when a CEO publicly disagreed with the conventional wisdom of their industry, everyone noticed because the CEOs, company structures, technology, and the markets around us all kinda looked the same. If they didn’t look the same, they were at least predictable in some capacity. Today, it’s farking crazy out there, you guys. We’ve got 24-year-old CEOs running 2-person billion-dollar startups, with AI technology changing the rules of how we operate daily … add a looming war, a jobs crisis, and a recession no one is ready to admit, and not only are things now completely unpredictable, but this crazy world has transformed our flavor palate to either be immune to spice, or simply craving bland as way to feel safe.

Now, as comms people, we’re sorta stuck in the middle of this tension between trying to find out how we can cut through the noise and how we can protect the brands we work for. 

We saw an opportunity with AI. But where AI could have acted like a sparring partner in creating these spicy takes, the truth is that it really made things worse. Sure, AI meant we could produce ‘spicy takes’ faster. But the problem is how AI operates to create that spicy content. It’s exceptional at generating content that sounds contrarian without actually being contrarian, because LLMs are really just trying to please you by telling you what the most average response is. Research from Stanford has shown that LLMs tend to produce responses that cluster around normative patterns, even when prompted for novelty. Another study in Science found AI leads to idea convergence within groups, narrowing variance in thinking.

It’s the same pattern, every time: AI expands access to ideas while quietly compressing their range. So it’s not scaling intelligence, rather it’s scaling the average. That's what makes them useful and it's also what limits them because they generate from aggregated experience (not lived), drawing on what has already been said, written, and validated. So they naturally gravitate toward the statistically probable, the structurally familiar, and the safe middle.

And we all know that ‘average’ doesn't get clicks, doesn’t get press, and it definitely doesn't move pipeline.

A reframe I'd offer for any comms person still unsure how LLMs work (because it’s SUPER important we all get this): an LLM is predictive text. The model is selecting the next likely word based on patterns in its training, not based on judgment, lived experience, or knowledge. To be fair, the models are getting extremely good at this. But that mechanical reality doesn't change, and our job is to know how the tool works so we can stop expecting it to do the part of the work it was never built for.

OK, let’s move on to what we can be doing instead, shall we?

So instead of the spicy take, what I'm seeing land with my clients right now is the ‘hard question’ approach. The goal of the ‘hard question’ approach is to be seen as a truth seeker. To do this, you have your execs publicly ask the questions their industry is avoiding, and be ready with their own POV on why they're asking. They don't need to have all the answers, they just need to invite the conversation and be willing to defend their thinking if the conversation moves off of X or LinkedIn to engage in real debate.

IMO, that last part is the only piece that really matters (the debate part). Thought leadership requires actual original thought and a willingness to do second-order thinking out loud. Whether the ideas come directly from the exec or from the comms person teeing it up, the whole thing collapses the second the "thought leader" can't articulate or defend their POV in a conversation — whether that’s broadcast, on stage, or in a customer meeting.

Even worse is that even if their truth seeking idea isn't from AI, it's going to feel like it is when they can't debate it.

I’ve found that strong truth seeker execs also tend to be high-agency execs. You can spot one in a few ways: 

If you have one of these execs, I’m jealous. It means your job is likely really fun and you get to feed them friction they can run with, and use you as a brainstorm buddy. 

If you don't have one of these execs, you’re not alone. My own spicy take here is that manufactured spicy is worse than silence. A reluctant exec pushing forced-takes reads as performative and audiences can smell it. It’s like the McDonald’s CEO’s famous bite of the burger. Just thinking about it makes me cringe.

But this doesn’t mean you’re empty-handed. 

Thanks for reading this far! Here’s how to activate a truth seeker strategy. :)

I’ll admit that the replacement of creativity is not something I’m worried about with AI. To me, the risk is that AI compresses creativity into predictable forms, and that compression has a downstream cost we're going to feel for years.

When leaders (and their comms people) start outsourcing not just execution but thinking itself, the internal struggle that sharpens ideas starts to disappear. The slow formation of insight that only happens when you sit with something hard for long enough is where original POVs are born. If we let that part atrophy, we'll have a sameness problem in our content that spreads into a sameness problem in our companies.

Truth seeking is a posture. It's the willingness to ask the question before you have the answer, to do that work in public and to keep showing up when the conversation gets harder. It's the opposite of what AI rewards by default, which is exactly why it cuts through.

So to me, spicy takes are out. Truth seeking is in. The ones who can wrestle with ambiguity and sit with discomfort to get this right will shape the future of our thought leadership. The rest will keep producing warm whole milk and wondering why no one's drinking.

Doing comms on your own? That’s Comms People is built for. Join us here.