One of my core skills used to be making PowerPoint decks. It was a skill I was proud of, one I had deliberately honed. In a corporate setting, slides are a powerful vehicle for sharing ideas, building awareness, and driving decisions. I learned to make them fast, pretty, and with the exact level of detail to make them digestible.Now, I have stopped making my own decks, for the most part. Claude does it for me. It also creates my Word documents and handles Excel like a champion. The entire core Microsoft suite — the backbone of white-collar labour — fully automated. All this led to an existential crisis: If I dropped it so easily, what was I actually doing all those years?AdvertisementThe answer, I think, is doing labour that seemed like thinking. The real value was never the slides themselves. It was knowing what to communicate, to whom, at what level of detail, and with what framing. But what was true in the old world was that thought and production were interspersed in the same workflow. This led me to confuse the two for years. I think most knowledge workers are stumbling upon this same confusion and feel a rising sense of panic.Speed is Not the PointThe internet is rife with influencers boasting about their incredible AI setups that let them do months’ work in hours. But doing more stuff is not inherently valuable. You have to be doing the right things. Case in point — AI slop has clogged every platform. Technically impressive but substantively empty.The true insight that I think very few people have internalised is that AI speed should not collapse the process. It should let you run the process many more times.AdvertisementI used to make a deck in four hours. In the first hour, I would tighten and refine the story. In hour two, the build would start, and then halfway into hour three, I would realise the framing was wrong and restructure the whole thing. Finally, in hour four, I would refine, tweak, polish. That friction was annoying, but it forced me to do something essential — I had to iterate. The deck got better not because I spent more time on it, but because I constantly went through micro cycles of building, evaluating, and rebuilding.So now I can produce the same deck in five minutes. The question becomes what to do with the other three hours and fifty-five minutes. The wrong answer is to move on to the next thing. That’s how you get the content flood, the slop, the fast garbage. The right answer is to iterate. Run it again, challenge the framing. Test a different structure. Bounce it off someone. Come back and run it once more.The Thinking PartnerThis article didn’t arrive fully formed. It started as a series of loosely structured thoughts I had been carrying around my head. Something incoherent. I wrote down what I had and then did something that would have been impossible two years ago: I sat down with AI and argued about it.Claude, my buddy, pushed back hard on my framing. Brought forward tensions I had been skating over. Proposed structures I hadn’t considered. It sparked a reaction in me: Suddenly, I had new insights, rejected previously held precious beliefs, and added new threads. Looked over it the next morning and realised it was not useful enough. So I went back and argued a bunch more. Through conversation, the final piece emerged.There’s this narrative out there that AI is “replacing human thought and cognition”. Trust me, you don’t want to use AI as a replacement for thinking. Think of it as a sparring partner instead. AI cannot drive this process. You have to take the wheel. In some sense, arguing with AI is like arguing with an amplified version of yourself. Reality is always the final feedback loop. That’s the real unlock. Don’t speed to completion. Speed to learning.The Abstraction Layer We Can’t ImagineBut if the lower layers of cognitive work are being automated, how will future generations develop the judgement and taste that sits above them? Don’t you need to go through the labour to build the intuition?I don’t have a complete answer. But I think people massively underestimate what it means for an entire generation to grow up with these tools.Consider a simple comparison. A 50-year-old and a 21-year-old both use a smartphone. But they don’t use it the same way. The 50-year-old has to think about how to use the phone; it was invented roughly midway through a life where it wasn’t the assumed reality. It’s a tool they consciously deploy. The 21-year-old quite literally thinks through the phone. It’s not a tool. It’s an extension of cognition. The distinction isn’t skill. It’s the abstraction level.The future is extremely hard to imagine, no matter how close to it you are. Two years ago, I was making slides. Today, I build web apps and write scripts. Not because I became a software engineer, but because AI handles the layers below me, allowing me to operate at a higher abstraction level. Now scale that forward a generation. The people who develop taste in that environment will be operating at an abstraction layer we literally cannot imagine. Several jobs, unimaginable to us today, could be a reality soon. That’s the nature of these transitions — they create entire categories of human endeavour that didn’t exist before.you may likeThis is Not A DrillThis is happening. It’s happening faster than any other technological revolution. The tech CEOs are not being hyperbolic when they make alarming predictions about the nature of work changing. I believe they are, through the imperfect medium of alarming headlines, making a point about the scale of AI’s impact.Most people have not actively changed their workflows, broken apart their jobs, or asked themselves enough hard questions about what it is they actually do. The barrier to capturing AI’s value isn’t intelligence. It’s engagement. Engagement means doing the hard things — separate thinking from labour, and iterate with speed. Pay attention and take it seriously. I’m watching the smartest, most locked-in people pick this up and run with it at a staggering pace. Anthropic has spent the last few weeks shipping features at a terrifying pace. Entire market categories are being violently re-priced as we speak. People who are awake to what’s happening and choosing to engage have an outsized advantage.Something is lost when cognitive labour disappears. Something enormous is gained. Both of these things can be true — the tension between them doesn’t resolve neatly. But the people who will thrive are the ones paying attention and being thoughtful about how to use the newfound speed. The truth is that we can all move at the speed of thought now. The question is whether you’re willing to think more than once.The writer is HR leader at Lennar