AI's 'Thin Ice' Moment: Is Your Job Already Gone?"

Jobs are the scariest topic out there and we're going to talk about them in a different way today. But to get there, we need to be honest and we need to say something that I don't hear get said much. The first sign that your job is on thin ice is often a full calendar and no clue what's happening. And I don't say that to scare you. I say that because it's important to be honest about how we actually see our jobs and where productive work is actually happening in 2026. Because often times your calendar really is full. Your manager is happy. The work is getting done, but the most dangerous moment in a job. It's not the moment after everything's been decided and the work disappears. It's when the work still exists, but less and less of that work needs you. That's what happened to travel agents. Expedia did not erase that profession overnight. Right? Online booking changed the economics of the routine work at first and nothing really changed. Like if you look at the data, you don't see a drop in travel agents right away. The visible break came later when downturns forced the industry to admit what had already changed. That is the pattern that I think a lot of knowledge work is sitting inside right now. And I don't mean dramatic you're fired kind of changes necessarily. I mean the kind of changes that restack roles and rejig roles and change who's responsible for stuff and dramatically reorganize the organization. Maybe there's some layoffs in there, but the point is I think we have an overhang where we're overdue for reorgs and we're not talking about it enough. And boy is it good to talk about that now and early if you're in a role because you don't want to be the one who gets surprised. Being surprised is the worst feeling. I've had that happen. I've been fired like two or three times. Like it's not great. So underneath what I'm hearing from people and I'm getting hundreds of messages a week, guys. I'm hearing people talk about this sense that jobs are getting hollowed out. I see it online when people talk about what it's like with AI. AI doesn't have to replace your whole job to put you on thin ice. It only has to pick away at enough of the pieces inside the job that when the next shock comes, the rest of the story stops holding together for the role. A recession hits, a budget freeze, a reorg, got some kind of margin problem or growth problem. And suddenly the company asked the question that it's been avoiding for a long time. Why is this role bundled this way, right? Travel agents had exactly this issue. The routine booking layer was exposed as unnecessary. The travel agents that survived had to change. They had to be more human. They had to plan complex trips. corporate travel, luxury travel, handle emergencies and do problem solving that the simple booking path completely broke. This is the pattern that I think a lot of knowledge workers, we ought to be watching this right now. And why should we be paying attention? Because again, it's not an immediate apocalypse. That's the story. There's actually still a lot of travel agents. Did you know that there's a ton of travel agents still employed? You you may not know one down the street, but there's a huge number still around. But there have been downturns. There have been changes in the role and it looks completely different than it did in the 90s and it's been a bumpy 20-year road to get there marked by special downturns when the economy as a whole wasn't doing well and of course travel's sensitive to that. Think about that from a knowledge work perspective. You have essentially capability overhangs that exist in most organizations because they're just not organized around AI. At some point that bill is going to come due. If you can start to see pieces of your role where you're like, I don't know, AI can do that and that and that, but I could do this other thing, but I'm not being asked to do it. So, there's kind of a mismatch between my role and what AI can do. Super normal. I hear that a ton. I hear this sneaking quiet suspicion that maybe I'm the only one that notices from so many people in my DMs. And I'm here to tell you, you're not the only one. I get tons of these DMs. We're all noticing this. You're not alone. And you're going to be okay. This video is about helping you to be okay. But first, we have to acknowledge that this issue is not just anecdotes in my DMs. It's real data. Open AAI and University of Pennsylvania researchers estimate that about 80% of US workers could have at least 10% of their tasks affected by language models and maybe one in five could see half their tasks affected. Anthropics Economic Index says about 49% of jobs have already had at least a quarter of their tasks performed using CLA. Microsoft researchers looked at 200,000 Bing co-pilot conversations and found that the most common work people bring to AI is not some weird edge case. It's just gathering information and writing. It's the rewriting email thing, right? And the most common work that AI performs is writing, teaching, providing information, and advising. In other words, and this is really important, AI is not just coming for jobs. It's actually coming for pieces inside jobs, bits of tasks that live inside a job. Which means the useful question is not, will AI replace me? The useful question is, how much of my last two weeks still needed me? So, I want to show you a quick audit that you can run on your own work. It separates your week into four buckets. Theater, let's be honest. We're being honest with ourselves, right? Commodity, what's on the line, and what's durable. By the end, you'll know which part of your job to stop defending, which part to automate or let go of, and which part is actually worth building your career around. Because the danger right now, it's not that AI takes your whole job in one clean move. That's not how it's actually working. Don't pay attention to the Tik Toks that say that. The danger is that AI removes the parts of your job that were quietly propping up the rest of the story that you told yourself about your value. And by the time your performance review notices, it's too late. So, we have three parts to tackle here. We're going to do them quick. First, why jobs tend to change slowly and then suddenly, and why your company probably hasn't updated yet. Second, the actual audit itself. And you'll remember four letters, T, C, L, and D. Keep those in mind. And third, what to do after the audit. Because the point is not to scare yourself. The point is to move your time and identity toward the work that matters. And that last part matters a ton to me. A lot of people can feel the ground shifting. Very few people have a practical way to change their week before the organization decides to change it for them. The mistake most people make is asking, "Will AI replace my job?" And just stopping there. It's too large a question to be useful. Your job is not one thing. Your job is 50, 60, 300 small things packed into one title in a trench coat. Some of those things are real and durable. Some are real but increasingly not specific to you. Some are valuable today but moving in the wrong direction. And some were always a bit of professional theater. Let's be honest. 5 years ago, if too much of your role was routine, the risk was just it was slow. Maybe someone cheaper could do it. Maybe someone faster could do it. Maybe eventually it would move to another team or another vendor or another layer of software. But that risk moved at a pace of months and years. Your competition was mostly other people and other people had the same basic limits you had. They could only write so many updates. They could only sit in so many meetings. They could only read so many docs. They could only summarize so much context. And they could only carry so many projects at once. So even if a part of your work was becoming less special, you had time. That's what changed. Routine knowledge work is no longer being competed away only by other people. It's being compressed by tools that don't have the same throughput limits. And this creates a really weird lag. Your work can still look fine from the outside. The documents can get written. The updates go out. The meetings happen. Your manager can still say you're doing a good job. But the important question is not did the output happen. The important question is did that output really require you and your expertise. And most performance systems are really terrible at that question because they're designed to measure visible output. They measure whether people like working with you. They measure whether the work appears to be moving. They measure the old system. So there's going to be a period where the review system says you're fine while the actual economics of the role are already changing. This is a dangerous window and this is why the travel agent example matters. The tech did not have to eliminate every part of the role to change the role. It only had to make the routine booking layer less defensible. And then when the industry hit pressure, the adjustment happened very quickly. The agents who survived were not the ones who defended routine booking as a professional identity. They moved toward the part of the work that still needed trust and taste and context and judgment. That's the question this audit is really asking about your job. Which part of your week is routine booking and which part is complex travel? In the first place, it shows up is the polite fiction layer. Every large organization has one. Be honest. Okay. Status meeting where nobody's actually unblocked. The deck that exists because one senior person needs something to flip through even though nobody will read it carefully. the alignment call that produces no alignment but lets everyone say that alignment was attempted or the recurring update that gets sent because one time 18 months ago somebody asked for it. Can you tell these are real to me? I've lived these. The review process maybe it once solved a real problem, right? But the problem is gone and now we keep the review process. This is not a moral failure. It's a structural feature of large organizations. Large organizations create theater because theater is legible. It gives people something to point at. It creates the appearance of coordination. It lowers social risk. It lets everyone avoid saying the very uncomfortable truth directly. For a long time, this was sustainable because everyone paid some cost to maintain that fiction. But AI changes the economics of that fiction. Theater is the first layer that gets absorbed because theater was already operating below the threshold of real human attention. If no one was reading the deck closely, a model can write the deck, right? If no one was deeply attending to the status meeting, a bot can summarize it. If the alignment artifact was never changing anyone's mind, it can become an async dock that nobody opens. And AI does not need to make the theater great. It only needs to make the theater kind of adequate because adequate is what the theater already was. And when that layer starts to collapse, it reveals what was underneath. For some people, what was underneath is real work that was buried in very bad packaging. Those people become more invisible because the noise around them goes down. Other people, what was underneath is a lot thinner. They were not doing real work that happened to be packaged in theater. they were doing the theater. That's the line you need to know which side you're on. And the only way to know is to look at your real actual week. So, here's your audit. One sitting, open the last 10 business days of your calendar. You can put me in the corner of your of of your screen. Okay? Just open it up. You can hit pause. Now, open your sent email. Open your Slack DMs. Open the docs, the tickets, the code commits, the spreadsheets, the memos, whatever medium contains your actual real work. Then go line by line, item by item, and tag everything with one of four letters. Do not tag the role. Don't tag the project. Just tag individual items. The one meeting, the one memo, the update, the decision, the review, the conversation, the actual unit of time. The first letter is T. T is theater. This is the work that exists because the organization performs it, not because it produces examined value. If the work disappeared and the main consequence was that the organization had to admit it had been performing rather than producing, tag at T. Yes, this is private. This is just for you. No one's going to tell. The status meeting where nothing changed. The dock nobody read, that's theater. The check-in neither person needed, that's theater. The review that existed for political cover. You're getting the idea, right? The first round of feedback where the feedback couldn't change the outcome because the work was done. All of that is theater. It's te and te will be the hardest one to use honestly because tagging something tea means admitting you spent professional time on something that did not need to happen. Super uncomfortable. It can be private. It's also the point. Okay, that's the first one. You can put me on pause again for a minute until you're ready to go to the next one. And by the way, if you're wondering, is there a way AI can help me do this? Yes, but you're going to need to be patient. I would recommend using codecs with computer use in a browser to start to tackle some of these tasks. But the caveat is this. If you go through a twoe set of work across multiple UIs right now, it is a ton of objects, hundreds of objects. Maybe you have two dozen meetings. You may have 60 or 70 Slack messages, maybe more. A lot of us have more. You have hundreds of emails. It's too much to ask even a clever computer use agent like Codeex to do in one shot. So, you're going to have to chunk the work. And you may have to do it in pieces. You may have to use a separate agent that just looks at at email to look at that and a Codeex that just looks at calendar maybe for you and you may have to use uh some Slack summary tools or Slack search tools to figure out what's in Slack. There's ways to do it. There's ways to let AI help you. I've got more of that written up on the Substack as to how you can attack this with AI and get a little smarter. But I don't want to hide the fact that you actually need to do the work to put your preferences in. If you're going to tell Codeex, "Please categorize my calendar." Guess what? You have to express your preference. You have to be clear. What are each of these four letters? What do they mean? How do you know? How do you judge from the attendees? How do you judge from the title? That takes some time. You're going to have to be willing to prompt. And I can give you a step up. I can give you guidance. I can give you sample prompts, and I will. But you're still going to have to figure out how that works with your individual work. Don't want to hide the ball there. Okay, next letter. C. C is commodifying work. It's real work. It produces real value. It needs to happen. It just does not need you specifically. Summarizing, routing, applying known rules to known situations, coordinating on things that have already been decided, writing a status report someone does read, but anyone could have written it. Turning a meeting into next steps, producing the first version of a dock where the shape is already wellknown. The test is simple here. Could you write a spec and have someone else in your org produce an output that is just about as useful? If the answer is yes, it's probably C. And that's not an insult. A lot of C work is really important. Companies run on it. The question is not does the work matter. The question is whether your personal career should be built on being the one who does it. And look, everybody's got to do some seaw work. I'm not saying step away and avoid doing it because that makes you very unpopular at work in most cases. I'm just saying be aware of it and be aware how much you're doing it and don't make it the centerpiece of the career. The third letter is L. L is on the line. This is the really uncomfortable middle. It's work that does not cleanly fit into commodity or the other category we're going to capture later, durable. So pattern recognition where the patterns are structured is on the line. Relationship management that depends on history that you carry, that's on the line. Editorial calibration in an established format on the line. Routine synthesis across familiar inputs. work that used to feel hard and now feels a little too repeatable. It's it's on the line. It may not be quite commodity, but it's close. This is the work where a strong junior person can do 70% of it maybe, but the last 30% feels like yours and you can feel that judgment is involved a bit, but if someone asked you to explain exactly what judgment you applied, you might struggle. Tag those items L. Don't overthink it. The point is not to be perfect here. The point is to see where the line is moving. The fourth letter is D. D is for durable. This is work where the output depends on something you cannot fully describe in advance. You changed the question more than you answered it. You read what was going on in a room. You saw that the stated problem was not the real problem. Your presence in the conversation visibly changed the outcome in a way that goes beyond competence. If the work would have degraded without you there and the reason is not simply that you're faster or more organized or more available. Tag it D. D is not just hard work. Some hard work is commodity work. Some durable work looks almost invisible from the outside. D is the work where you are carrying context and taste and courage and calibration or or pattern recognition that can't be cleanly specified before the work starts. Now tag it pretty quickly as you go through. First instinct, if you agonize over an item, you just tag it as on the line and keep going. And when you're done, count it up. And that count is the first honest picture of your job that you've seen in a while. And yes, if you can describe this clearly enough, AI can help as I've called out. And I'll have part of that part of the guide on Substack will be about that. And based on dozens of conversations, here's what I think you will find. Number one, your T number is bigger than you want it to be. Most people underount theater on the first pass because they confuse this was professionally expected with this created value. Those are not the same thing. Something can be professionally expected and still be theater. Number two, your C number is also probably bigger than you want it to be. And this one feels unfair because seawwork often took years to get good at. You learned how to summarize messy context. You learned how to route decisions. You learn how to write the update that calms people down. You learn how to turn chaos into a dock. That was real skill. The problem is that a skill can be real and still become less scarce. Markets don't really care how hard something was to learn. Organizations don't protect a skill because it once took you a decade to build. They protect whatever is scarce now. Number three, the D number. The durable number will probably be smaller than you wanted it to be. And that hurts in a different way because in your head, your professional identity is often it's probably built around the durable part of the work. is built around what you think is unique. You think of yourself as the person who can read the room, who can diagnose the customer, who can spot the bad deal, or who can tell when the strategy sounds good, but won't survive contact with reality. And maybe you are that person, but the audit asks a colder question. How much of the last week or two weeks did you spend on that? Not how much of your self-image depends on it. How many hours? For a lot of people, the answer is less than they thought. Which means the problem is not that you have no durable work. The problem is that your week is not organized around it. And I want to call something out here. There is a real case to be made for people who make their career on great moments. And so I don't want to tell you your durable work isn't a huge part of your week and have you take away from that there's no value in it. I believe strongly that the hours in our working lives are distributed according to a power law. There will be a few moments in our working lives that are extraordinarily impactful. Maybe it's the time you said yes to your dream job, right? Or maybe it's the time you made one decision that saved the company and it was the right call in a moment of crisis. I don't know what it will be for you. There are not that many of those hours. There are not that many of those days in our working lives. Maybe only a handful across decades. But those make or break our careers. And so you can say in that vein, maybe it's not that dramatic, but you can say, "I have durable value, but really what I'm paid for is making extraordinarily good decisions twice a quarter." Okay, I can believe you, but let's make sure those are really excellent decisions. They're very legible. They're easy to see. And you actually build your your career around that. And you're not just filling time and producing theater. Now, the L bucket is where most of that career question lives. This is the work that's in motion right now. Some of it's going to end up in the C category. Some of it's going to become durable. And most of your work anxiety is probably sitting right there because on the line is where your identity and the moving line of AI are just like mashing together. And that's why the number that matters first is actually the combination of theater and commodity. theater plus commodity is the fraction of your current week that is on thin ice. It does not mean all of that work vanishes tomorrow. I'm not saying that. It does not mean you're bad at your job. I'm not saying that either. It means this is the part of your week where your personal claim on your work is the weakest. And once you see that number, you can't unsee it. But the audit is not the answer. The audit is just the input. The real question is what it reveals about the kind of work that survives. Right? And the first thing it reveals is that durable work is mostly questionholding, not questionancing. Most organizations reward people for answering questions. Someone asks for a plan, you make the plan. Someone asks for an analysis, you do the analysis. Someone asks for recommendation, you produce the recommendation. And that can be valuable, but it's also very commodifiable. Question answering is the surface area AI is best positioned to absorb because the question is already a given. The frame is set. The output can be judged against the prompt. Durable work often starts before that. It starts when someone asks a question and the right move is not to answer it. The right move is to say, "I think we're asking the wrong question." That's a very different skill. It's uncomfortable because the system wants resolution. The meeting wants to move forward. The executive wants the recommendation. I I know I've been there. Right? The team wants next steps. The customer wants the feature they asked for and they're right to ask for it. Question holding is the ability to honor those real commitments, but also keep the real question open long enough for a better answer to become possible. A customer asks for a feature, the feature is not what they need. The work is not just saying no. The work is holding the gap between what they asked for and what they need without losing the customer. A project can get framed around an objective that sounds reasonable, but is slightly wrong. Sometimes the work then is not to execute perfectly against the wrong frame. The work is to keep the framing question alive without becoming just a source of noise and frustration to everybody else. A team wants to solve the visible problem because the visible problem is easier to discuss. The real problem is often trust and incentives and decision rights and ownership and and even fear. Durable work is being able to read all of that and not get trapped inside the stated problem. And that's why Dwork is so hard to measure. The output is often the thing that didn't happen. The bad hire that didn't get made, the the product detour that did not consume 6 months, the customer escalation that did not become a crisis. Performance systems are terrible at crediting avoided damage. I get that. But avoided damage is often where senior judgment lives. So, if your durability bucket is small, one of the first questions to ask is, "How much of my week is spent answering questions other people handed me? And how much is spent holding questions that really matter?" That's not a philosophical question. I'm not trying to play gasha here. It's just a calendar question. You can find out. I'm showing you how. The second thing the audit reveals is that different kinds of work compound in different ways. Theater compounds to nothing, right? Let's just be honest. You attend the meeting, you generate the deck, you send the update, everyone performs their bit, nothing gets learned, nothing gets encoded, nobody gets sharper. Commodity work can compound, but it mostly adds value to the organization. If you build a better routing system, a better summary pattern, a better checklist, it's real work. The company is the one that benefits and the institution gets smarter and the tool absorbs the pattern and it's super useful and you're going to have to do some of it and it's good. But be honest about who owns that compounding. And don't just do that. If the thing you're improving can be captured by the system, the system will capture it. Durable work compounds very differently, it tends to compound to you. The pattern recognition you build from holding ambiguous questions for years does not live entirely in a document or a prompt or a workflow. It lives in your calibration, in the scar tissue of having been wrong before. And in the moment where you can feel a clean story becoming just a little too squeaky clean and you got to add some reality. You can teach around that. You can share parts of it. You can build institutions that benefit from it, but you can't fully transfer it. So when you look at your audit, don't just ask what work is valuable. Ask where does that value accumulate. If it compounds to nothing, if it's just theater, stop volunteering for it. I know you'll have to do some of it. If it compounds to the organization, do it. Systematize it and let it go where it's reasonable to let it go. If it compounds to you, if it's durable, protect and feed and nourish that. The third thing that this audit reveals is the legibility paradox. Durable work has to be legible enough that the system values it, but not so legible that the system can run it without you. Some people keep their best work invisible. They they read the room. They prevent the bad decision. They shape the strategy in a conversation and they quietly save the project and then nobody can explain what they did so they get underredited. The other mistake is making the durable work too explicit and people don't realize that's a mistake but it is. You take the thing that is actually judgment and you turn it into a process document. You overexlain every step. You write down the decision tree and maybe you get it right, maybe you don't. You underdescribe it and some of it is a desire to be generous. Some of it is good management. But once a piece of work is actually specified, it can be delegated and you run the risk of if it was true judgment, delegating it badly and seeing people not be able to do it well at all and then generating mistakes for the business. Or maybe you were right and you were able to specify it and now it's a commodity. And look, I'm not saying you should not write things down. Quite the opposite. I make lots of videos saying you should for agents. So, I'm not saying don't write stuff down. I'm saying there's going to be something like partial legibility here. You want to show the outcomes and you want to recognize the difference between a process that does need to be written down to scale the business and a piece of essentially magic that you do. high quality judgment in very ambiguous circumstances that if you wrote it down, it wouldn't be real anymore. You wouldn't make great decisions. A competent person couldn't pick up your process and execute it reliably because real highquality judgment, the ability to hold those questions in the room, the ability to reset the team, no one writes that down. That's why you have years of experience. It's not easy. And if you think you can write it down, you're probably fooling yourself. And if your organizations thinks that you can write it down, they're fooling themselves. So be partially legible. Sure, show the outcomes, but don't expose the mechanism where it doesn't make sense. There may not be a mechanism you can even articulate there. Like you may not be able to say, I held this framing open for three more weeks and it prevented us from solving the wrong problem because that would make everyone embarrassed. And you don't even know why you picked that moment to hold the frame open. Really, you just it was instinct and it worked. That kind of ability to articulate is something you have to develop an understanding for because there will come a moment when you're going to want to say the decision was made because I made sure that we actually asked the question properly. That's often at a performance review. Sometimes it's in conversations with your boss casually. There is a moment to bring up a piece of that, a piece of that outcome. That's the partial legible part where you notice what you did. But the rest of it kind of stays hidden. It's your secret sauce. What you're developing is meta judgment. It's the skill of knowing when a particular case is distinct, when the rule technically applies but practically fails, when a clean framework is hiding the real issue. And that that's not something you can explain very easily. The fourth thing that the audit reveals is that identity is a true obstacle. The mechanics are easy. You could run this audit this afternoon. The hard part is what the tags are going to do to the story you tell yourself about who you are at work. Theater threatens the story because it says some of my professional time was meaningless and now I know which ones. C. Commoditized work threatens the story because it says some of the skills I worked hard to build are becoming way less scarce and less valuable. On the line L, it threatens the story because it says some of the work I identify with is moving. And D for durability threatens the story because it says the part of me that I think is most valuable might be a really small fraction of my week and I don't want to admit that. It's why a lot of people I know avoid thinking about their time this way. They they don't avoid it because they lack the tools. They avoid it because the audit asks you to challenge some of your self-identity. And identity updates are psychologically expensive. You pay the cost now and the benefit arrives later. It's a terrible deal psychologically. It's also the deal that serious operators and serious entrepreneurs and serious builders make. The advantage goes to the person who can update their self-image before the organization forces the update on them. And that is what this audit gives you. It gives you the chance to move while you have agency. So, what do you actually do? You've done this audit. You've done the brave thing. Here are your next six moves. First, stop performing the theater. You can stop if you can do it without consequences. Don't start with the hardest political case. Start with the theater that exists by inertia. Right? The recurring report nobody really reads. The meeting where you're the third senior person and the second one is enough. The check-in that made sense two years ago and has become a ritual. The status update that could become three sentences in Slack. Cancel. Skip. Send a short version, whatever it is, watch what happens. Most of the time, nothing happens. That's the point. The move is not to become the person who declares war on every meeting. The move is to stop donating hours to theater that nobody requires. Second, do not pour your recovered time into more commodity work. That's the trap. AI helps you write the update faster, so you write more updates. You cut two useless meetings, so you fill the space with more routine coordination. You become twice as productive at the part of your job whose value is collapsing and it feels like progress because the system still rewards visible throughput. I think that's a bad trade long term in most cases. The better move is to put recovered time into durable skill development. Take on cases that don't fit the patterns you already know. Choose projects where the framing is unclear, not just the execution. Sit in on conversations where you're not the expert and you have to read the room. get closer to the people carrying context your normal workflow abstracts away. This is not upskilling in the normal sense. Upskilling usually means adding a tool, a credential, a workflow, a technique that you can prove, right? Instead, this is developing judgment under conditions where the right answer isn't known. Third, build a private track record of durability calls. At the end of every week, write down one call you made where the outcome depended on judgment you cannot fully reduce to rules. Write down the line. say, "What is the call I made? What was the context? What was the result? Or the date when I'm going to know the result." This is not a brag sheet. Just keep the record. After a year, you're going to have something like 50 entries if you do one a week. After 3 years, you literally have a portfolio of judgment. And when someone asks what you do or why your role matters or why the work should come to you instead of a cheaper process, you're not reconstructing your value from half-remembered memories of three months ago. You have evidence. You have evidence. Fourth, use that record to gradually refuse commodity work that doesn't fit your trajectory. Most people can't simply say, "I don't do routine work anymore." No, it doesn't work that way. First, become visibly valuable on your non-rine work. Then, use that value to renegotiate the routine load. And this often happens just through project selection before it happens through formal authority. When you have a choice, choose the project where the answer is uncertain over the project where the path is documented. Choose the conversation where you have to understand what is happening over the workstream where you just apply a known playbook. Durable judgment needs raw material to get better. It needs messy cases. It needs exposure to reality before reality has been cleaned up into a memo format. Right? If your week never gives you that mess, your durability bucket isn't going to get better. It's not going to grow. Fifth, make your durable work legible enough to be valued but not legible enough to be commodified. Like I said, talk about outcomes. I was concerned that we were solving the wrong problem and I got us to have the conversation. We changed the plan. The data pointed one way, but my judgment was that the case was different. And look, here's what happened. I was right. Those are visible claims. They help the system understand where you contribute outside commoditized work. But they don't turn your judgment into a recipe. Also, separate analysis from judgment in the way you talk. Analysis is the work that can be transferred and it's important work. Judgment is what you do with the analysis. The analysis says this, my judgment is that we should still do that. The framework says one thing, but this case is different. This is why that language teaches people where to bring you in. It tells them, give me the cases where the analysis is not enough and I need to apply some level of intuition. That's the work that you want. Sixth, be honest. If the audit says you're in the wrong role, if most of your week is theater and commoditized work, if the on the line work is mostly drifting toward commoditizing and there's no realistic path inside the role to build meaningful, durable skills, the answer may not be better time management. It might be moving. Roles are not equally rich and durable work. Some roles are veryheavy because the organization is theaterheavy. Some roles are commodityheavy because they were designed for an earlier era and no one has rebuilt them yet. Some roles naturally create durable work because they force someone to hold really ambiguous questions in real time. If you're evaluating a new role, don't just read the job descriptions. Job descriptions almost never describe the durable work because durable work is hard to put into a job description at all. So, look at the people actually doing the role. Ask what they spent time on last week. Ask where the ambiguous questions are. Ask what calls they made that could not have been made by a process. Ask what work they would keep if half the routine output disappeared. If they can't answer in specifics, be careful. A role with no visible durable work might still be prestigious. It might be well paid. It might be safe for now. It may not be a good role for the next 10 years. The deeper point is that the system you're operating inside was built around an old assumption about knowledge work. Performance reviews, promotion frameworks, quarterly goals, and skill ladders, and headcount plans, and meeting rituals. All of it assumed human output was the scarce thing. And that assumption is breaking really unevenly right now. It's not breaking all at once. It's not breaking in the same way for every role, but it is breaking enough that you cannot outsource the diagnosis to the way the system always worked. The system will update eventually. It's going to lag unpredictably. And during the lag, the people who can see their own work clearly will have an advantage over the people waiting for the review cycle to catch up with them. Seeing clearly does not mean just reading more essays about the future of work. It means looking at the actual 50 hours you spent last week and sorting them into buckets that reflect what is true, not what is comfortable. And if you're wondering, can I get this done? The answer is yes. You can feed this transcript into codeex into claude and you can say, help me make a plan to do this. I'm scared. And it will. So stop volunteering for theater. Use tools to compress commodity work, but don't reinvest all those gains into more commodity work. Spend the recovered hours on harder cases. Build a record of calls that only make sense if judgment was involved. Make the outcomes visible. Keep the deepest part of the mechanism, the secret sauce, in your head and in your hands. And if the role doesn't give you enough durable work to grow into, start moving toward one that does. That's the whole game. It will not look impressive in the first month. For a while, it may look like nothing is happening. You cut a meeting. You chose a harder project. You write three private lines at the end of the week. You ask a better question in a room where everyone wanted an answer. These are small moves. But small moves compound when they're aimed in the right direction. The choice is not whether to learn AI tools. You're going to need to learn those anyway. The tools are part of work. I talk about them a lot. But that is no longer the differentiator by itself. The real choice is what you invest in with the time and attention the tools give you back. You have basically two options. You can use the time to do more of the work where value is collapsing. More updates, more summaries, more routing, more documents in known formats, more throughput through the same old shape. And that will feel productive for a while. It will be rewarded in the short term because a lot of old systems still like that visible output. Or you can use the time to do more of the work whose value is compounding. Harder questions, stranger cases, better judgment, more calibration, more direct contact with reality, more willingness to hold the problem open when the room wants closure. Now that will feel less productive at first. It will be less legible. You may not get immediate credit, but years from now, the difference between these two choices is going to pay off. One group of people will become extraordinarily efficient at work. the system no longer needs and another group will have used this transition time to move closer to the work that still requires a person. You are not yet on the wrong side of the sorting, right? That's the important thing. You have time. The audit that you do, it's not a verdict. It's just a starting point. So, have a look at last week's calendar. Be brave. Look at the week before. Be honest. Don't panic. Just count it up. Then ask the only question that matters after the count is done. What part of my week am I going to stop defending? and what part of my week am I going to start to feed? That's what matters. And if you're wondering, yes, this is all written up on the Substack. Yes, there are guides to help you through this audit process. And yes, I cover all kinds of career advice related to AI on this channel all the time, especially as the tools keep coming out. So, look for more like this soon. I'll see you around. Cheers.