It’s not the first time I’ve felt replicable though. As a young postdoc, I once met with a bunch of Stanford scholars for lunch at a conference. Their entitlement aside, I’ve rarely felt less accomplished than listening to their publication records and job market strategies as I was out on the job market myself back then. Surely, I would not have a chance at a coveted professor job if that’s the competition. These people made me feel small despite my unfettered work ethic.
Little did I know back then that this fear was just infused by feeling judged by them and by my future superiors. Let’s face it. As researchers, we live in a constant state of judgment. There’s always a dean, panel chair, VP of Product, hiring committee, or worst of all, our inner reviewer passing judgment on our abilities. There are always gatekeepers that we’re facing. The fear of replaceability is never abstract.
The fear is fuelled by the abstraction though. Thinking that AI is replacing you is vague enough to not take action. The fuzziness is comfort and catastrophe. But the gatekeepers are real. Someone you respect or look up to will look at your work and possibly decide it’s interchangeable and treat you accordingly. Peer review is really everywhere. And it’s often done by a small handful of humans with budgets, reputations, and a short attention span.
Here’s a 5-step process to become irreplaceable to gatekeepers:
Repeating this process makes you better at it. And repeating the production of such artifacts in public compounds for you. Repetition is the antidote to replaceability. But you have to repeat the right process.
The crucial thing you can’t forget here is that whatever you write, should be based on your work. On the research you’ve done or vetted. You have a distinct point of view. AI does not. It represents the amalgamation of everyone’ point of view. So, keep your stance and knowledge and question outputs.
When everything tastes like bland cauliflower, people are dying to find a spicy carrot in their veggies. Be that spicy carrot. Lean into your uniqueness, but be aware of the arguments of your gatekeepers of why your work might not be good enough. Convince them otherwise. Imagine their claims and challenge them. Can you use AI in that process? Absolutely. It’s a fast research assistant and can check sources with the right prompts and plugins. But should you let it turn your carrot into a cauliflower? Absolutely not.
Don’t wait for people to certify your work. Trust your gut over automation.
You don’t need permission to sell your competence to the world.
—has taken our professional lives by storm. The more I’m using it in my day-to-day tasks, the more I feel I’m solving problems that I didn’t actually have in the first place. It is just too easy to go down a rabbit hole of “What if I just write a quick tool for this?”, and then find yourself 2-3 hours of fable prompting later when you have built a tool that you’ll likely only use once. You snort that peak productivity like cheap cocaine. Only, to feel an emptiness at the end of the day that is hard to describe.
It’s the Pareto principle all over again. I produce about 20% amazing workflows, automation, and apps for the 80% junk that I end up throwing out or not even finishing after a full afternoon of attempts to fix it, rewrite it, or shelve it after I can’t quite get it to the polishing level that I’m happy with. On top of that, the 20 different agent tasks running on my computer feel like a busy conversation at the watercooler, out of which I can really only pick up one or two strands of intelligible conversation. My feeble mind is clearly the bottleneck here. Sometimes this fills me with frustration about how all these AI influencers on YouTube and social media are automating away their solopreneur businesses (I call BS!). Other times, it gives me hope for humanity and that not all of our jobs are going to be automated away much too soon.
All of this hits an aspirational problem that I think many of you share with me. We always find something that we would like to do more of, but that we do not regularly do yet. You might find that your AI automations are competing with an existing habit instead of attaching to one. For example,, if you already read papers on Sunday mornings with your double espresso (or Latté if you’re like me), you will likely abandon a workflow that requires you to first pipe the PDFs through three different tools before you can even begin reading them. The cost is much higher than the benefits that you get from this task. Why? Because now you have an AI output review queue. And believe me, this is one of the insidious tidbits of AI work. Sure, it produces tons of output in minutes, but I still have to review it. And I’m a slow reader.
So, you face two choices: Take your time (and accept that reviewing AI work is now part of your job) or just do it from scratch without AI. For many creative tasks, it ends up taking the same time. Just with AI, you’ve added work to remove work. “Mischief, thou art afoot.”
If you feel like your AI work costs your more to maintain it returns on your time investment, it’s probably time to check your workflows. You shouldn’t have to invest 20 minutes a month to tune a workflow that saves you only 10 minutes. So, today I’m giving you a checklist of four questions that you can run before you build anything with AI to see if the balance sheet stays positive:
Any problem that you face over and over again, whether it is during a day, a week, or a month, is worth considering automating. So, the first step is to just watch for activities that you notice you are repeating regularly in your work. Those are the most valuable ones to automate with AI.
Let’s say you’re a policy researcher who briefs a committee every other week, writing those briefs is a recurring problem. A workflow that helps you structure or draft them is of true value to you. If you’re thinking, “I should probably read more broadly in adjacent fields,” that is just aspirational. Here, you’d wait to build the workflow until your reading becomes a recurring commitment or habit.
Here’s a simple reflection question that you can ask yourself at the end of a week or a month: Did this exact problem slow you down in the last week/month? If you have to think hard about when it last came up, it’s not recurring enough to automate.
I think the trap that many of us fall into is that we are immaculate at imagining valuable work, which means that we can always see an ideal version of how we’d do things. We could imagine a better knowledge base, a more optimal scanning of the recent literature, or a more elegantly organized research archive. However, we have to remember that a workflow doesn’t become useful because it supports the future person that we’d hope to become. It becomes useful only if it can support what we’re currently repeatedly doing.
So, you always have to start with the work that you’re currently doing.
As humans, we operate our daily lives on habits, repeatable structures that shape how we do things. The best way to wire a workflow into your workday is to attach it to a trigger that is part of something that you already do regularly.
Let’s say I already know that I’m opening my email at 8 AM and then pull up the same analysis environment. Or I always draft a one-page summary before I have an important client conversation. I would consider these behavioural anchors and I could turn them into triggers for my behaviour. Workflows can be triggered well if they’re linked to these existing anchors and this is much easier to do than establishing a habit or a new behaviour from scratch.
So, here’s another simple reflection question you can ask yourself: What did you do this morning that you do most mornings? The answer to that will give you an opportunity for creating a new workflow or automation based on your habit.
For example, suppose you routinely annotate articles before adding them to your reference manager. An AI workflow that turns your annotations into three reusable components—a claim, a limitation, and a possible implication for your current project—fits such a habit. You don’t need to adopt it as a separate practice because you’ll just do it where you already work on a paper.
By contrast, a weekly automated research intelligence briefing sent to a dedicated inbox may look sophisticated, but unless you read this, it won’t be useful to you. You have to ask yourself, if this is something that you’d check, or if it’s competing with everything else that’s demanding your attention currently. Just thinking you’ll read a brief because it’s created is unlikely to happen. It’s much better to work it into a ritual that you already have (e.g., reading a newspaper).
Good workflows shouldn’t take more than five minutes to clean up and put into action. The best output is output that you can act on right away. Otherwise, if you have to spend too much time re-editing everything, you end up resenting this. And then over time, this will just not naturally fit into your work, and you’ll just start skipping it. So, there’s no efficiency or productivity gain here.
For me, five minutes is really the maximum I would invest into editing it. With the ideal being closer to 30 seconds of a quick review where I just scan the output, maybe tweak a line, and then use it right away. That, to me, is an actually efficient output.
Let’s imagine you always write a one-paragraph framing note before presenting quantitative findings. Then, you should build a workflow that generates a first-draft version of that note from your results and a few keywords. Reviewing this will be manageable and feel familiar because you already know what the purpose of the paragraph is. This will also help you to understand what’s wrong in seconds.
Or just imagine that your clients always come to you with the same questions repeatedly. Things about what a specific finding means, how you would justify a method, or what they could actually do with it. In these cases, you’d want to design a workflow that turns your research report into exact responses to these question types (think of it like an FAQ). It would still be based on your research report, it would just deliver the answers that they’re looking for much quicker.
Problematic workflows will deliver formats to you that you’ve never actually used before, or they will use a voice register that you might not use in your own writing. These would need to be adjusted, and you’d have to improve the output quality just so that you don’t feel resistance to the artifacts you get.
I’d almost say that this is the most important rule because to really understand a process, you have to run it by hand first. And I would say three times is a good minimum threshold to see how the process is actually done, so that you can describe it in detail. This helps set the quality bar for the automation you’re trying to build. It lets you understand what steps are essential to this workflow and what criteria determine the quality of the output that you get.
The first manual run shows you what the task is. The second shows you what is annoying about it. The third shows you what matters most in the output.
Going through this more labour-intensive run of the process three times will allow you to not guess about the task but have a solid understanding of it and your automation won’t miss the point of the task.
The goal for any good automation is to model how you actually think through a task. And for this to succeed, you do have to go through the process of thinking through that task first. Otherwise, you might get something plausible and it might look right, but, at the end of the day, it’ll still feel wrong, and you won’t actually use it.
But even as you’re building these AI workflows (e.g., skills) and automations (e.g., routines), you have to keep an eye on the ones that actually provide benefit for you. That means running a weekly tally of workflows that you’re actively using. So you have to make three decisions here: whether you want to keep the workflow, redesign it, or retire it. This is the best way to keep your AI work manageable as it likely grows.
Keep an automation/workflow if it runs as expected, produces usable output within five minutes of review, and you would notice its absence immediately if it broke.
Redesign it if you use it occasionally but skip it more often than not, if the output usually needs heavy editing, or if you have started working around it. Redesigning means returning to the four questions and rebuilding from scratch. Don’t just tune the prompt. Rethink the workflow.
Retire it if you can’t remember the last time you ran it, if you are running it only to justify having built it, or if the maintenance cost is higher than the return.
As a simple rule of thumb, you should ask yourself after you’ve run a workflow/automation three times without producing anything that you’d end up using, now should be the time to change or delete it.
Why even trim workflows, you ask? Because in theory there’s lots of storage available and a lot of the times you have more tokens than you need for your AI work. But there’s a mental overhead when you constantly run a system that’s only half managed, so do yourself a favour and clean it up.
The next time you catch yourself mid-rabbit-hole, thinking “What if I just write a quick tool for this?”, stop and ask the four questions. Recurring problem. Existing anchor. Usable output in under five minutes. Three manual run-throughs before you automate. Any idea that makes it past this test deserves your afternoon, and the tool you build will still be useful next month, saving you time while everyone else reviews an inbox full of AI briefings that nobody asked for. The rest of your ideas can stay ideas. Close the tab, pour the Latté, and read your papers the way you always have. The high of building something you use every week beats the cheap stuff.
Build better AI workflows that help you thrive in your own work.
This week, Premium Write Insight + AI Research Stack members get the complete workflow audit kit: a Review Queue Balance Sheet spreadsheet, printable Automation and Keep, Redesign, or Retire guides, a downloadable Claude skill that interviews you about a repeated process and turns it into a reusable skill, and five paste-ready prompts for committee briefs, reviewer responses, paper annotations, recommendation letters, and quantitative framing notes. You also get 10 practical resources for deciding which systems deserve a permanent place in your week. Join Premium, run the audit, and reclaim the afternoons you now spend building workflows you abandon before your week ends.
]]>I’m as far away from hustle culture as you could get. But student requests, grant coordination, customer emails, building things, it all feels like it swallows your day whole. I still ask myself sometimes how all of these online creators are creating so much content without ever burning out. Or maybe they’ve all just burned out and they just never admitted it in public and then just hired a team? And I’ve optimized my systems with productivity setups to the teeth.
Either way, as I’m inching towards the end of my sabbatical, I am thinking deeply about how I can optimize my systems and create things that won’t burn me out as I’m going back to a full teaching load. I think a large part of this is the AI agent army that expects me at work every day. I’m typing so many different requests across Claude Codex, Hermes, Notion AI, Antigravity and others, that it feels like speeding up. However, it also means I’m juggling 10,000 tasks at once. My brain isn’t ready. I might forget a task that I started in the morning and only pick it up later in the day. And then as you’re changing computers or setting up your computer anew (like I just did), it all feels like your memory is scattered and fragmented across many different sessions. The soup can be hard to warm up. And I’m feeling the struggle to reconstruct what it is that I need in the moment more than ever.
I’m still deep in building the early version of my membership right now, and it has me thinking hard about this exact topic: which parts of my own process are worth showing in public, and which parts I’ve been treating as overhead. And also the core question: Is it even worth for me to keep building this membership or should I pursue other venues?
In 2020, OpenAI’s head of research quit his job. His next action could be interpreted as a disadvantage. Dario Amodei left the company with several senior colleagues, including his sister Daniela, who led safety and policy. They walked out over a disagreement about the shipping speed of the AI. They thought the company was shipping product faster than its safety work could keep up.
Their actions at the time looked like a disadvantage.
They incorporated Anthropic as a Public Benefit Corporation. Simply put, this is a for-profit company that writes its mission into its legal charter. So, the board can choose the mission over quarterly profits. They did not rely solely on human labels. Most founders treat this structure as a cost they want to avoid. Anthropic treated it as the product.
They trained their model, Claude, using a set of rules they named a constitution. They did this to avoid leaning only on humans labeling good and bad answers.Anthropic shipped Claude, built on Constitutional AI, in March 2023. Then, in September 2023, they published a Responsible Scaling Policy. This was a public document that outlined the conditions for refusing to train or deploy a model. The framework is based on biosafety levels used in labs for dangerous pathogens. It has graduated tiers that get stricter as the danger increases. They funded mechanistic interpretability tools — diagnostic probes and visualizations that inspect model internals so teams can spot deceptive behaviours before deployment. He framed his wager as a race to the top: set public safety standards and force competitors to meet them.
I won’t bore you with the full funding‑round details. Every founder who has been told depth doesn’t sell should be bothered by the short version. March 2025 — investors valued Anthropic at $61.5B, early 2026—$350B valuation. Amazon and Google were investing tens of billions jointly. They were also preparing to go public. In the months following the scaling policy, OpenAI and Google DeepMind released their own versions. Anthropic says the policy caused this change. I’d say, more carefully, that it stopped being the weird one.
Safety was not the only reason this worked. It helped substantially that Amazon and Google were hunting for an OpenAI alternative to fund. The timing of the language-model wave was pure luck. Plenty of rigorous labs have faded into obscurity. Depth alone is not enough.
Put rigour on the front door.
Anthropic took the most rigorous, least glamorous thing it did and put it at the front door. Rather than narrowing a niche, this approach is differentiating the wheat from the chaff. Niching asks how small you can cut your market until you are the only one left. Differentiating identifies the specific people who need a skill only you provide. Research-focused founders should apply the discipline they already use instead of overlooking it. The way you care about your method. Or your discipline to not over-claim results. All the things you consider essential for good research.
When you start, you might treat that discipline as overhead. But your market treats it as scarce.
I read AI labs’ safety claims carefully, because I study deception by generative models. So take this as a lesson in positioning and nothing more. I made the opposite mistake myself for years. I started by targeting PhD students because it seemed the natural choice. I narrowed the market to my obvious audience. Differentiating worked better. I kept the same level of rigour but focused on a smaller group who benefited most, and then I explained, plainly, why that rigour mattered.
Make sure you get it on the public record before someone with half your depth and a louder feed defines your category for you.
You win the next decade by putting your research on the public record now; customers, recruits and design partners will find you. The technical founder people already trust.
Everyone else calls it overhead, you call it rigour.
The full issue in The Write Insight includes the interactive Front Door Audit scorecard (12 statements measuring how legible your rigour is), a 4-phase legibility checklist, and 6 curated resources including the Constitutional AI paper and Anthropic's Responsible Scaling Policy.
]]>This notion relieves you of the need to tell people what you made. That's why it outlives so much contrary evidence.
It has never once happened.
Gregor Mendel laid the groundwork for genetics in 1866. He published his findings in the Proceedings of the Natural History Society of Brünn, after sharing them with the society in 1865. The work was rigorous and essential. However, the field ignored its importance to a great extent for over 30 years. It was rarely cited. One common account mentions “about three” citations in about 35 years. A more cautious view says it showed up in several publications. But people didn’t grasp its main point. In 1900, Hugo de Vries, Carl Correns, and Erich Tschermak revived Mendel’s work. However, the rediscovery was more complex than a simple independent rediscovery. The point is the research went unnoticed the entire time.
Douglas Prasher cloned the gene for green fluorescent protein (GFP) in 1992. Prasher mailed clones, for free, to anyone interested. Martin Chalfie asked. Roger Tsien asked. Chalfie used Prasher’s clone to express GFP. Later, Tsien mentioned that his early GFP work was possible thanks to DNA he got as a generous gift from Prasher. After Prasher’s grant funding dried up, he shut down the lab and left science for a time. In 2008, the Nobel Prize in Chemistry for work built on his clone went to Osamu Shimomura, Chalfie, and Tsien. Prasher was a courtesy shuttle driver at a Toyota dealership in Huntsville, Alabama. The early work made no mention of his contribution.
Peer review is often seen as a system that evaluates quality with no marketing. Even in that context, the paper needs a cover letter. Reviewers need a reason to look. They need a frame. A contribution statement. The editor's slush pile doesn't know which paper matters. Without a claim of significance, neither does the reviewer.
Outside academia, the margin widens. Builders ship tools. Talkers tweet about the tools and land podcast invites. The builder gets the citation (if even). The talker gets the platform. Prasher got a nod in the Nobel banquet speech (and at least a paid trip to Stockholm).
Unfair. Also true. I’m afraid that calling it unfair doesn’t change anything.
This is how the system works: quality determines whether attention converts.
A million people can ignore your work. Quality won't matter. But if a hundred people see it and five act, quality is why those five got inspired, why they stayed, why they recommended it, and why they paid for it.
Attention drives on a much different fuel. Repetition. Narrative. Status. A platform focusing a spotlight. Quality provides none of those things. It doesn't surface in feeds. It doesn't pitch itself. It doesn't email an editor.
I believed the opposite for years, and I'll admit the belief cost me. Not citations, but impact. I kept polishing work past diminishing returns, assuming polish was the only bottleneck. It wasn't. Too few people knew the work existed. My best-cited papers weren't my best papers. A co-author talked about them at the right workshop, and citations followed. It was timing. A little bit of luck. And distribution followed. Fiercely.
Attention and quality don't cooperate. They compete for the same hours, and attention wins the opening battle every time.
The attention domain runs on repetition, novelty, and volume. Rewards the loud. Punishes the patient. Daily lies beat one-time truths. The domain doesn't check your credentials. It doesn't value citations. It only cares about the headline.
The quality domain runs on evidence, replication, and time. Rewards the careful. Punishes the rushed. A weak daily claim will fall apart under scrutiny one day. The domain doesn't reward speed. Volume doesn't matter. The facts underneath the headline do.
Use rules from one domain to battle another. You will fail at both. An attention strategist will tell you to just post and skip the rigour. But if people find your work, they will discredit it. A quality strategist will tell you to skip the repetition. You will remain undiscovered. An actual audience will never get to appreciate your rigour.
Mendel fought the quality war and won it long before anyone knew there was a war. His pea experiments were rigorous, quantitative, and foundational. He lost the attention war. He published in a local society’s proceedings and sent out reprints. He also corresponded with a leading botanist. Still, he could not get the field to understand his findings. His paper wasn't unread. It was misread. It was rarely cited. Colleagues missed the point. Filed it away as plant-hybridization work until 1900.
Prasher fought the quality war and won it, too. His clones worked (for Chalfie and Tsien). His science was sound. The cDNA he shared was the starting point for work that made GFP a universal biological marker. Yet, the Nobel committee recognized other researchers for their work with GFP. He failed to wage the attention war. Lost control of the asset that could have kept him the hero of his story.
Neither domain forgives a deserter. Skip the attention war and stay invisible forever. No matter how sound your work. Skip the quality war and burn out fast. No matter how loud your launch.
Most people treat this as one campaign with one strategy. They are two campaigns. Two sets of rules, two different enemies, and no truce between them that saves you from fighting both.
I'll spare you the sociology-of-science literature review. The short version is that even scientists—the group most committed to meritocracy—run on framing devices. Cover letters. Conference talks. A collaborator vouching for you at the right meeting.
None of these inflate the work. A framing device doesn't claim more than the work delivers. It says this exists, here's why it matters, here's where it fits. It shines a spotlight the work cannot cast on itself.
Without framing, the work sits in a database nobody opens. A tool sits in a repository no feed surfaces. A book takes its place among the two million others published this year.
Framing doesn't make good work speak for itself. Nothing makes good work speak for itself. Framing makes good work findable.
Don't feel depressed now. This distinction should free you.
Stop waiting for quality to do a job it wasn't built for. No work is so perfect it defies invisibility. Not even the work that founded genetics.
Two tasks make up your job here. First, visibility. Share it, say it plainly, repeat past the point of comfort. Second, quality. Good enough that your audience acts, returns for more, and recommends it.
Most experts spend all their energy on the second task and none on the first. Then they complain about lack of impact as a quality problem and go polish some more. I get it. Polishing makes us feel virtuous. But it's just procrastination.
Visibility generates attention. Quality converts attention. Two practices. One outcome. Impact.
Mendel had the quality skill. Prasher had the quality skill, too, and mailed his impact away for free. Probably the better man. Probably also frustrated a little about someone else's Nobel win.
If nobody has found the best thing you’ve made yet, take my quiz to understand why.
Freedom begins when you stop waiting to be discovered.
You got this, my friend.
The Write Insight premium subscribers this week also get the Good Work Spotlight Kit: a 45-minute spotlight session, a seven-day sharing plan, printable framing and claim-check worksheets, and 4 AI prompts for turning one piece of work into a public case, auditing a draft, drafting a significance sentence and headlines, and tracking evidence and recipients. The kit includes worked examples, plus 5 curated resources on visibility, framing, self-promotion, and evidentiary rigour.
]]>The administrative swamp eats 14 hours of your week. Meetings, emails, and tasks you do not need to do. I’ll spare you the generic productivity hacks. On June 11 at 11:30 AM Eastern, I am showing the 7 habits that close those leaks on my own calendar. We will schedule the 3 habits that buy back the most time for your research and writing. Beth, Ijeoma, José, Rabeya, Frances, Dianna, Amy, and Nikita already registered by today. Register here to save your seat.
Academic training made you good at one game and left you losing another. For years it rewarded a single audience, the two or three reviewers who sign off on your work, so you learned to read every paper cover to cover and catch every gap. Your instinct earned their approval. But in the visibility economy, this is burying you. While you grind through 50 papers, the writing that would make your expertise visible never ships, and obscurity wins by default. You slowly come to understand that the real bottleneck here is how you time your selection decisions. You waste a day deciding what to read while you read it. I urge you to make the call before you open the paper. Here's the system I use to map a field before lunch, it saved my weekends and pulled our last paper's remaining literature together in under 48 hours.
I spent 20 years reading research the slow way, cover to cover, like most of the early PhD students I supervise. I used to open the first paper in a stack, read it line by line, examine every figure, and take dense notes. I believed highlighting and verbatim note-taking would lock the knowledge in. Cognitive science shows us this is an illusion. I love Zebra Midliners, but highlighting isn’t very effective. It locks your attention onto single ideas and buries the links between them. Those links are what let you make inferences and learn. Highlighting can even hurt how we make inferences.
One for the keeners in my lectures. Transcribing what you hear word for word won’t help you understand the ideas either. Copying closely isn’t enough for deeper understanding. It can hurt your performance on conceptual and application-based tasks. Mere transcription of thought relies on shallow processing. Writing key ideas in your own words, a process called *generative note-taking,* helps you learn much better. You pick key points, condense them, and rephrase in your own words. Then, you connect those ideas. Notes are useful only if you actively process them. Without active processing, you only look like you’re reading. But you won’t remember anything.
The secret to processing 50 papers lies in your selection criteria. You succeed by refusing to read all 50 in the first place.
Skimming is shallow reading with your comprehension switched off. Relying on it to get through your reading list is how you waste a full day. Trust me, I’ve done it.
The goal is to decide what matters before you open a single document. The way I do this is to start with an assessment and a prioritization.
I take my stack of 50 papers and sort them into three piles. I do this using only the titles, abstracts, and figures. I keep the process simple to build momentum and save the difficult work for later. Here are the three sorting piles:
Pile A holds my deep-read targets. These are the three to seven papers my argument actually depends on. I’ll read these with close attention. One test sorts them. A Pile A paper is one my thesis can’t survive without, because it supplies my core theoretical framework or the exact method I am replicating.
Pile B holds only my mapping targets. These are the 15 to 20 papers I need to know exist to understand the boundaries of the debate, but that I don’t need to master. I put a paper here if it contextualizes my research question. It shows me who else is playing in the space. It also gives background stats and notes the viewpoints I should weigh.
Pile C is my park or discard pile. The remaining papers go here. Make this decision in under two minutes per paper. If the paper’s main finding doesn’t connect to your research question, put it in Pile C. Don’t hold on to papers because they’re interesting or written by famous authors.
My oldest mistake was treating all 50 as equal. I spent all my time on the first few documents and never got to the three papers that really mattered. Now I run the sort during the week with my research question written on an index card in front of me. If a paper doesn’t directly answer that question, it goes to Pile C immediately.
The framework I use for Pile B and Pile A comes from the three-pass method introduced by Keshav (2007). In his classic paper, “How to Read a Paper,” he shares a clear method. This method connects your deep reading to the process of making decisions.
I’ll spare you the academic jargon and show you how it works in practice.
Each pass narrows the field. Pass one runs across both Pile B and Pile A. Only Pile A survives into pass two, and just the two or three must-read papers reach pass three.
Pass one is a quick five-minute scan. Focus on the title, abstract, introduction, section headings, and conclusion. You glance at the references to see who is being cited. At the end of the five minutes, you must answer his five C’s:
A shallow skim has no exit criteria. The five C’s give you five, and they make you read the paper’s structure. That’s how you process a reading list in a morning.
Spend about one hour on pass two, focusing solely on Pile A. Read the paper carefully, study each table, and note references to follow up later. If you hit a difficult proof or a dense mathematical block, you put it down. The common mistake here is rereading the same page five times. Keshav’s fix is to leave the hard parts for the next pass, and I love it.
Pass three takes four to five hours, and few papers in your pile warrant that much time. Reserve this pass for the two or three papers you must master or review. You rethink the study, question every assumption, and spot all the gaps.
A bibliography is a list of who said what. Build a review around authors and you get a string of summaries with no synthesis.
Jane Webster and Richard Watson made this point in their paper, Analyzing the Past to Prepare for the Future: Writing a Literature Review (Webster & Watson, 2002). They were tackling a shared issue in doctoral education and research. The issue is reviews that summarize past research in order, listing each contribution. This structure lacks synthesis and fails to expose gaps. To solve this, they introduced the concept matrix as a method to force researchers to structure literature around ideas. Their whole goal was to make a literature review concept-centric.
I built a concept matrix. Rows are papers. Columns are claims, methods, samples, and findings. The matrix turns 50 separate documents into one map you can work from.
For example, a concept matrix for research on educational gamification might look like this:
| Paper / Source | Self-Determination Theory (SDT) | Leaderboards | Extrinsic Motivation | Method | Key finding |
|---|---|---|---|---|---|
| How gamification motivates | Yes | Yes | No | Experiment | Badges, leaderboards, and performance graphs raise competence need satisfaction and perceived task meaningfulness. |
| Assessing the effects of gamification in the classroom | Yes | Yes | Yes | Longitudinal field study | Over a semester, leaderboards and badges lowered students' motivation, satisfaction, and final exam scores. |
| An empirical test of the theory of gamified learning | No | Yes | Yes | Experiment | Students with a leaderboard spent far more time on task, and that time raised their academic performance. |
Laying each paper out in a breakdown like this makes the gaps obvious. If five papers share the same experimental design and ignore motivation, start your next paper there.
First, I use semantic academic search engines like Consensus or Google Scholar Labs before collecting documents. I skip the hassle of boolean keywords and just type my research question into the search bar. This method finds important papers that regular keyword searches miss. It speeds up my discovery process.
Second, once I’ve collected these papers, NotebookLM is excellent for finding where they disagree. I prompt, “Where do these papers contradict each other?” The tool shows the passages, and I click the citations to verify the source material on the spot.
Third, Claude is useful for drafting the synthesis from a concept matrix. Instead of copy-pasting raw text, I upload the matrix either as a CSV, an Excel file, or provide a link to the Google Sheet where it lives. This lets Claude ingest the structured data cleanly, so I can ask it to draft the synthesis and argue against my interpretation.
Never trust either tool to generate a citation. The risk of fabrication is high, and grounding only lowers that risk.
I tell my graduate students and research coaching clients to run this workflow during the week. They don’t need to work weekends. They process their 50 papers in business hours because they stop trying to read every word. This system kills the need to read all 50 papers. You read the few your argument stands on and map the rest.
AI can compress a whole literature faster than any of us. The judgment of what to read yourself stays with you. That’s how our last paper’s literature came together in 48 hours, with one index card on the desk.
Your turn. Grab your next stack of 50 papers, sort them into three piles before you open a single file, run the three passes, and drop what survives into a concept matrix. Give it one morning. Then hit reply and tell me what your index card said.
This week, premium subscribers get a print-ready concept matrix workbook, the full one-morning field-mapping protocol as a checklist, three AI prompts (triage 50 abstracts into three piles, run the five C's on any paper, and turn a gap into your paper's angle), and five curated resources on academic reading and synthesis.
]]>A week is 168 hours. It's a fixed number. Simple math. When a busy senior expert feels short on time, more discipline won't help. Discipline just adds more tasks to an already packed week. It treats the symptom. Every week, 20-30 hours vanish. They slip away into admin tasks, fragmented work sessions, and distracted family time. Maybe you spend that time babysitting your AI agents. But really, you’ve never figured out where they actually go. I've been in this situation. Heck, sometimes I even still am. You could use those lost hours to put yourself in front of people and make your work visible.
This newsletter issue helps you heal that wound. A tenured professor's field-defining work never leaves the journals. An industry researcher's name stays locked inside the company, and often remains unpublished. An ex-academic explains their credentials on every sales call. They do this because no one already knows who they are. Expert founders watch hollow competitors win spaces they should own. You have the credentials but remain unseen. It's because you tell yourself you lack the time to build a public presence.
I write this while managing a research lab, running a company, and publishing a newsletter for 14,000 readers. So, this isn’t just theory from a random office. I feel your pain. Today, I'm giving you two systems. First, a quick one-week inspection saves you about 24 hours. Then, a publishing schedule turns some of that time into authority. Reclaim the time. Spend a little of it to make yourself visible to the public.
Before you hunt for lost time, sit with a finding that pulls against the whole idea of hidden hours. Recently, I watched productivity expert Laura Vanderkam on Cal Newport's podcast. She has read thousands of real time logs for her books, including 168 Hours and Big Time. She found that the time is almost never missing from people's lives. People fill it with low-quality default leisure they don't even remember. But then we tell ourselves a story of having no time at all.
Time is a limited resource. Once you realize the hours are always there, you see you might be miscounting how you spend them. This realization helps you fix an inaccurate record. And running a business when overloaded is not a matter of cutting everything back to nothing. Vanderkam draws a line that's worth keeping. Complexity and chaos are not the same thing. Chaos is the enemy of productivity. Complexity is fine.
Life is complex by design for:
You face this complexity like a conductor faces an orchestra. Calm, rehearsed, every move timed. A real orchestra is not chaos. It is a tightly organized operation, where nobody gets a solo at the wrong moment. Aspire to that. Create the symphony.
I'll spare you the full literature review. Here is my system.
Set a silent alarm on the hour. When it goes off, write down what the past 60 minutes held, in 15-minute slices. This includes email, scheduling a client call, a school pickup, focused work, and TikTok scrolling you don’t want to admit to. Be brutal. Memory tricks you. It makes you think you had more productive blocks by smoothing over the distractions and gaps. A simple spreadsheet works better than your weekly perception. So does any good time-tracking software. Seven days. No edits, no judgement.
Most people stall right here, because honest tracking is uncomfortable. Log the ugly hours first. Three days of lunchtime scrolling and 90 minutes of half-attention email is data. Treat it as data.
At the end of the week, sort each hour into a few categories that suit your life:
Researchers and entrepreneurs most commonly experience two issues. Admin friction runs 15 to 18 hours a week; two full workdays are lost to inbox, scheduling, and status pings. Constant Slack notifications create a persistent drain on productivity. The worst part is that constant interruptions force you to work in short, 30-minute chunks. These fragments are too brief for deep focus and hurt your working memory.
Leroy's research shows that when you switch tasks, some attention lingers. It stays on the previous task. A 22-minute work block runs at lower capacity. This happens because your brain is still working on the last task. This mental fragmentation explains why your 24 hours feel unproductive despite constant effort.
Now you have recovered time, and the instinct is to spread it thin across everything you're behind on. Resist that. Anders Ericsson's 30 years of research show that focused creative work lasts 3-4 hours a day. A senior expert can manage about 10 to 15 hours of focused work each week, no matter how much they try to do more. Protect one small piece of it, just 2-3 hours, for the one task that strengthens your authority for the next 10 years. Publishing what you know.
The objections are always the same, and I get it. You don't have a free afternoon to write. Guess what. You don't need one. In 1989 the behavioural psychologist Robert Boice tracked 20 academics for a year. The group that wrote in brief sessions of 10-20 minutes produced 157 revised pages. The group that waited for empty afternoons and binged produced 17. A 9x gap, same talent, same topic. Brief and frequent writing beats rare and heroic.
Three actions help weekly publishing last a whole week.
A system you renegotiate every week falls apart in a busy week. I like the concept of house rules that Newport and Vanderkam discussed. These are standing decisions you make once and never revisit. Her family eats pasta every Monday so nobody debates dinner. Yours might be "I publish Tuesday mornings," "admin happens in two windows, not all day," and "I stop at a fixed time each night." Cal Newport's A World Without Email makes the similar case for your inbox. Clearing it is not the work your business gets paid for, so stop turning inbox zero into a video game. My inbox is crowded. I search for what I need to find. Batch it. Give each message the time it deserves and not a minute more. Or let an AI agent surface important emails.
Then build the net before you fall. Pre-declare a fallback cadence. When the grant deadline or a sick kid happen, and one of them will, you ship the shorter piece instead of vanishing. I've done that a couple of times with this newsletter and I still struggle to deliver on the same day each week. From time to time the system drops the ball, too, but those are exactly the moments you built your fallback for.
Build your time-saving habits it in this order.
Track minutes from day one. Peykar and colleagues (2023) conducted a trial with 132 nurses. They found that time-management training reduced work-family conflict and distress. However, this effect faded without a refresher. So re-audit at the end of a month and compare the results.
Start the tracker tomorrow morning. Next Sunday at 10 PM, instead of feeling that guilt, you will know where all 168 hours went. One of those hours will show your published work. It will reach the audience you need.
The Write Insight subscribers with an AI Research Stack premium account this week also get this bonus: A one print-ready PDF worksheet (a one-page Invisible-Loss Reckoning Card), 3 AI prompts (sort a raw time logs, diagnose where your week leaks hours, and design a rebuilt template), a visual reckoning protocol, 5 curated resources on time tracking and attention, and a two-week audit-and-rebuild rollout.
In today's newsletter, I'm breaking down the seven strategies that turn AI from a vending machine into a strategic partner, in the order they need to be set up. Based on today's newsletter, I've also recorded this as on overview on YouTube, check out the newest video:
Quick update: I made a quiz that lets you self-assess why you aren't the obvious choice for your clients yet. If you're someone that is starting a business part-time or full-time and has an expert background, then that quiz is for you.
The Authority Lab had another open coaching session this week for member Balaji, where I helped him transition from an academic position in AI for scientific simulations toward building a personal brand and potential business. One of the interesting nuggets I touched on in the session was that side benefits compound even if your business does not take off right away. Speaking invitations, board seats, funder confidence, and conference programming flow toward visible authority regardless of whether the commercial offer scales. If you want to be part of the Authority Lab, join us here.
I've also settled on the date for our upcoming Mastermind (Thursday, June 4, 12 PM ET). Check your email to help me choose the topic.
Everyone at first treats AI like a vending machine. Put in a prompt, get out content. The output looks polished but is starting very much to sound like everyone else. MIT’s 2025 NANDA report on the GenAI Divide found that 95% of organizations got zero measurable return from generative AI pilots. Only 5% extracted real business value. The other 95% bolted AI on without defining what it was supposed to do, for whom, and in what voice. I think this is one of the most common problems of AI use in business that we’re currently seeing. People are rushing to implement it in everything without having a proper plan or even strategy for how it can improve their work or output. So I figured I might as well share my seven strategies below to try and close that gap. You can set those up in three phases (I talk about this in my YouTube video): foundation, execution and distribution.
Language models default to agreeing with you, because agreeable answers score higher in their training data. SycEval (Fanous et al., 2025) tested ChatGPT-4o, Claude-Sonnet, and Gemini-1.5-Pro and found sycophantic behaviour in 58.19% of cases. So, when you ask a frontier LLM about how you can improve a landing page. You usually just get average assumptions for what that would look like. And it probably looks very pretty at first, but once you do a couple of those, you see how they all average out. So again, it makes sense here to reframe the question and tell the model that the campaign for the landing page already failed and ask it why. A good prompt for that would be: “Assume this landing page converts at 0.3%. List the five most probable causes ranked by likelihood.” Fifteen minutes of this before every launch saves a month of polishing the wrong thing. Try it.
If you don’t have any reference documents or a specified voice document and examples, you will feel that AI just has this kind of generic flavour to its writing, which evens out any of the bumps and interesting tidbits that you usually find in your writing. The texture feels rather flat, but you should know that the texture is what your readers usually come for.
An easy fix to do that is to do a four-part file that’s quite easy to create with the help of AI or by hand. And that file should contain:
Once you have a setup like that in place, it gets much easier to get really good output from an AI prompt specifically when you’re running it with an on-device tool like Codex, Antigravity 2, or Claude Cowork/Code.
The thing that I hear most from people that I work with is that they struggle to turn one good idea or one published research paper into many different pieces of content that they can use to advertise that work. And as a result of that, they struggle with time management.
Now, one system that works for that really well is that you don’t draft every piece of content from scratch but that you record maybe a 30-minute conversation about your expertise (and you can record that right into AI by talking) or you could talk to somebody and create a transcript from that or you could write it down in a book chapter or a YouTube video and then you can create something from this original source that allows you to do short clips, another newsletter issue, maybe some LinkedIn post or an FAQ page or even sell an email course.
The thing is you want to start with something substantial because this depth can’t be reverse engineered from a single tweet or even just a statement that you have. Now the statement or tweet can definitely be something that gets you started but you want to transfer it into those different formats and AI can really help you with that to chop that down into different formats that you can then use to draw the attention to that deeper bit of content that you’ve written at the beginning. This is a technique that has been popularized by many online creators such as Dan Koe.
The thing that was hardest for me to learn when I started a side business on top of being a professor was to understand that the content that I create online and in my newsletter doesn’t really address my own problems, but it addresses the problem of the people that I want to do service for or that I want to consider potential buyers of my products or services.
We call this market segment an ideal customer persona. And this is really important for somebody that builds their own business because if you have a fuzzy category like small business owners, you’re competing on price. But if you have a very specific end-user category such as senior researchers in their seventh year who want to leave academia for a commercial venture without losing intellectual rigour, then you have what we call an ICP. Here you just compete on specificity. And a great way to arrive at an ICP is to feed AI a stack of your client conversations, your coaching transcripts, or your discovery calls.
Or if you’re not doing any of these activities yet, maybe just to have a talk with the people that seem to be interested in the work that you do, to try and tease apart what their pain points are, what their urgency triggers are, and the exact phrases that these clients use to describe what exactly they want from you. The narrower you can go there, the warmer the messages reaching out to you will be.
It can be really hard to get testimonials from the people that you work with. And most people just post one testimonial and then they move on. But the persuasive material that is already in your business is much bigger than just testimonials. You probably already have email threads, call transcripts, Slack direct messages, maybe some survey responses, something like that.
You want to mine all of that feedback that you’re getting from your target audience to get the exact phrases that clients use to describe the before and the after state of working with you. And these are the kind of snippets that you want to drop into email welcome sequence, maybe a booking page, a newsletter, a blog post and an ad copy. This kind of proof works when a buyer is at a decision point. For example, here is a video I recorded with my coaching client Jia:
Now when you create an email list, this is usually a great way to communicate directly with your audience. And the nice thing about an email list is that you can send regular updates, just as what I’m doing right now with you, where I’m giving you relevant messaging that is interesting to the majority of my email list. The interesting bits come from then segmenting your list based on the preferences. See, for example, that I asked you a question in a poll at the beginning of this email, where based on how you answer to that, I will then try and optimize my messaging for you (in addition to picking a great mastermind topic).
A 2025 case study on Klaviyo by Oguta and Eling reports that behaviourally segmented campaigns hit 42.5% open and 18.3% click-through, against 28.7% and 9.5% for unsegmented broadcasts. The interesting part here is that the writing volume didn’t change, but the engagement was much higher because the subscribers stopped getting content that they didn’t feel they signed up for.
This is one of my favourite strategies and the one that my highest-performing clients credit for every good output. Because the generic output from models is that they serve everyone very broadly, so they fit no specific business venture very cleanly. They also don’t fit a specific academic project very cleanly.
You want to close that gap by writing enough context that you could have a brief you could give to consultant that would help you day by day. This is everything based on your personal or your company background, your ICP, pricing, voice document, past campaigns and their results, SOPs, banned phrases. Sure, it might take you two or three hours to get this started and created in the first place, but every session can load it, every interaction can reference back to it, you can reference it in your CLAUDE.MD file, and having such a context stack is really what makes all the difference in output quality that you get from AI. Without it, the other six strategies underperform.
The common trap is reaching for distribution work while the foundation is still missing. Segmentation and personalization can’t compensate for an undefined voice or a vague ICP. If a draft still sounds generic after Strategy 7, the context document is missing voice or audience detail. It takes a while to get these right and provide enough context. Add it. Regenerate. Ship.
If we’re breaking this down into the three phases of foundation, execution, and distribution, I would spend two weeks on writing the AI context document and building the voice system and spending some time to define your ICP. Then I would move towards the execution phase, which would probably take you another three weeks to do properly, where you set up the recording engine, run some background research on the next campaign and mine your existing library for any proof. Of course this depends on how many documents you already have.
And then I would finish with the distribution phase, which is probably four weeks at the very end to get going where you’re setting up the segmented email and the personalized sequences. And then you just let it compound over time. But you have to build the brief first. And then once you’re in the distribution phase, you will see that the speed will follow.
Why do my ChatGPT drafts sound generic?
Because the model defaults to averaged, agreeable outputs. Sycophantic behaviour appears in 58.19% of frontier model responses [2]. Provide voice rules, ICP, and a context stack to fix it.
What is an AI brief?
An AI brief is a structured document containing voice, audience, examples, and constraints that the model loads before generating. It functions like a creative brief for a consultant.
What is the best AI tool for writing in your voice?
Any frontier model (Claude, ChatGPT, Gemini) works once you load a voice document and context stack. The bottleneck is the brief, not the model.
How long should an AI context document be?
Long enough to brief a new consultant for a day, typically two to three hours of writing across background, ICP, voice, and past campaigns.
What is an ICP in marketing?
ICP stands for ideal customer persona, a narrow definition of the buyer including pain points, urgency triggers, and language.
Do segmented email campaigns actually perform better?
Yes. Klaviyo data from 2025 shows 42.5% open and 18.3% click-through for segmented campaigns versus 28.7% and 9.5% unsegmented.
How many pieces of content can I get from one recording?
Roughly 30 across short clips, newsletter, LinkedIn, FAQ, and email course, depending on depth of the source.
Enemy-based positioning is a copywriting and brand-strategy method where an expert defines themselves by the obstacle they oppose (e.g., a habit, system, environment, or false belief) rather than by credentials or features. It earns attention because readers recognize the friction immediately, then engage with the expert's nuance.
I started last Friday’s Authority Lab hot seat with the least heroic opening possible.
“Hello everyone, or no one yet.”
Then I sat there for a short minute, waiting to see if anyone would show up.
I wondered if I should have sent an announcement. I wondered if the time was wrong. I wondered if Friday was a bad idea. Very strategic founder behaviour. Sitting alone in a video call, questioning my calendar choices.
Then many people joined. Phew. Awkward minute over.
And the session turned into one of the clearest conversations we’ve had in the Authority Lab so far.
The format was so simple. People brought their positioning statements that I helped them formulate over the last month. The classic “I help X do Y without Z” kind of thing.
I expected us to clean up at least the language, but I had a bigger plan: Confrontation.
So, we found the missing fight.
Positioning is the deliberate choice of which problem you solve, for whom, and against what alternative, a definition popularized by April Dunford in Obviously Awesome (2019).
Experts love accurate language. I do.
And this all sounds good until you watch it turn into pudding.
Most expert positioning fails because it stacks careful adjectives ("timely, relevant, inclusive") that are technically correct but emotionally invisible. Readers feel problems.
"Safe language is the pot belly of positioning: accurate, comfortable, and useless in public."
None of this language is wrong. That’s the problem.
It's too safe to be useful in public.
Your audience does not feel a pair of careful adjectives. They feel a problem. They feel the thing in their way.
Your public positioning gets 100% better when it names that friction. Even better, it names what causes the friction.
That’s the enemy.
Impact needs an enemy.
The word enemy makes thoughtful experts nervous. I get that.
Most of the people I work with do not want to bash anyone. They do not want to flatten complex problems into cheap conflict. They do not want to become LinkedIn theatre people yelling about how everyone else is stupid.
Good.
Please keep that instinct.
The trick is to aim the fire at the obstacle, not the person.
→ Your reader is the person stuck in the fight.
→ Your customer is not lazy. The enemy might be inertia.
→ Your client is not ignorant. The enemy might be the technical gap that makes a good decision feel risky.
You become useful when you name the thing they are already fighting.
It makes all the difference.
If you make your customer the villain, they will defend themselves. If you make their obstacle the villain, they will lean in.
They’ll finally feel seen.
The enemy is the obstacle blocking your reader's outcome. Donald Miller's Building a StoryBrand (2017) frames this as the villain in the customer's hero journey. Valid enemies include habits (perfectionism), systems (assessment-driven schooling), environments (algorithmic feeds), and false beliefs (nuance = trust):
These enemies work because you already feel them and your audience does, too. You’re naming the conflict that has been running in the background.
One Authority Lab member works with athletes.
His first framing was about performance pressure. Athletes train well, then score lower in competition. The obvious solution was mental training.
Useful, but still a little broad.
So we kept digging.
The enemy was the inner voice. The self-talk. The fear of losing. The fear of winning and then having to keep winning. The tiny internal commentator that turns a competition into a referendum on your entire identity.
Now the positioning had heat. Spicy. I loved it.
He could stand against that voice.
Another member teaches pre-service teachers.
Her first framing was full of strong ideas: Creativity, process-based learning, risk-taking, student engagement, meaning-making.
All good.
Still too clean.
Then she named the thing she stands against it all: Shallow learning.
The grading system. The product-based logic. The classroom habit where students learn to ask, “How do I get the A?” before they ask what the work is trying to teach them.
That is a real enemy. Boom.
Everyone who has been through school recognizes it immediately.
Even AI fits this story. AI did not create shallow learning. It accelerates the shallow learning already built into the system.
Now, she had a sticky point of view.
One member works with literary scholars.
Her challenge was different. She resisted the customer language, which made sense. In scholarship, people do not always think in terms of markets and solutions. They think in terms of knowledge, methods, traditions, and ways of seeing.
So we stayed with that.
The enemy became institutional sameness. The Eurocentric lens. The habit of talking about diversity while keeping the same old epistemological frame.
That’s strong.
It gives her public access into deep work without making the work shallow.
You know, this is the piece that many experts are afraid to admit.
An enemy does not reduce your nuance.
It earns attention long enough for your nuance to matter.
Most expert writing fails because it always starts with the answer.
Here is my
The reader has no reason to care yet.
They don’t know what fight you’re entering. They have to learn about what problem you’re refusing to tolerate. They must be convinced why your sentence should interrupt their day.
Friction creates this hook.
This is true in stories. A good story needs pressure. Something has to be at stake. A character wants something, and something blocks them.
Public authority works the same way.
You want to be known for something. Your audience wants a result. The enemy blocks the result. Your position becomes the stance you take against that enemy.
That is where attention starts.
Forget your list of credentials.
Move past your polished bio.
Don’t write another sentence that sounds like it survived three committee reviews and a light afternoon nap.
It starts with a fight.
Take your current positioning sentence. Find the outcome. Then find the obstacle.
Most people stop too early. A little less Mr. Spock, a little more Captain Kirk.
They write:
→ I help experts communicate their ideas more clearly.
Fine. Clear enough. Also forgettable.
Now add your enemy:
→ I help experts communicate their ideas clearly without cutting off the depth that made those ideas valuable in the first place.
Better.
The sentence now has friction. It names the fear many of your carry: That visibility will make you shallow. And, yes, that’s how I position myself.
Here is another:
→ I help researchers write papers faster.
Useful. Flat.
Add the enemy:
→ I help researchers write papers faster without letting perfectionism turn every draft into a swamp.
That’s what I have done for years now. Now, the reader can feel it.
They know that swamp.
They’ve lived in that swamp.
They’ve probably decorated it like Shrek.
The formula is quick and simple:
I help [specific reader] get [specific outcome] without [specific enemy].
The without part is where the positioning gets serious.
BTW: I created a quiz that lets you check your current positioning.
Take your current positioning sentence, isolate the outcome, then name what blocks that outcome for 80% of your audience. According to Nielsen Norman Group research on landing-page comprehension, users decide relevance in under 10 seconds, which means the obstacle must appear in the first sentence.
Conflict is the oldest attention mechanism in narrative. Harvard Business Review's analysis of B2B messaging found that contrast-driven value propositions outperform feature lists in recall tests. Naming an enemy front-loads contrast, so that your nuance then has somewhere to land.
"An enemy does not reduce your nuance but it earns attention long enough for your nuance to matter."
If your positioning feels vague, you probably do not need more adjectives.
You need a sharper enemy.
Name the habit, system, environment, activity, or false belief that blocks your reader from the result they want.
Enemy-based positioning focuses on obstacles. The customer is the hero and the enemy is the friction. Attack marketing names competitors or shames audiences, a tactic the FTC notes can violate truthful-advertising standards. Enemy-based positioning stays empathetic to the reader while being smart about the system.
Keep your empathy for the person.
Aim your fire at the obstacle.
That is how your writing gets clearer.
That is how your hooks get stronger.
That is how your authority starts to feel like a point of view instead of a professional summary.
I am opening the waitlist for a free Authority Lab Mastermind called How to create hooks for your writing.
We will take this exact idea and turn it into a practical hook system.
Bring the topic you keep circling. Bring the sentence that feels too soft. Bring the idea that deserves more attention than it is getting.
We will find the fight hiding inside it.
Q: What is enemy-based positioning in one sentence?
It is positioning that defines an expert by the obstacle they oppose, not the features they offer.
Q: Can my enemy be a competitor?
Generally no. Effective enemies are habits, systems, environments, or false beliefs that your audience already feels, competitors invite legal and reputational risk.
Q: How do I know if my enemy is specific enough?
A stranger in your target audience should nod within 10 seconds. If they ask what you mean, the enemy is still abstract to them.
Q: Does this work for academics or scholars?
Yes. Scholars can position against institutional sameness, methodological orthodoxy, or epistemic flattening, all without market language.
Q: Won't naming an enemy make me look polarizing?
Only if you target people. Targeting an obstacle increases recall while preserving credibility, consistent with StoryBrand research.
Q: What is the positioning formula?
I help [specific reader] get [specific outcome] without [specific enemy].
Q: How often should I revisit my positioning?
At least every 6–12 months, or whenever your audience changes. Dunford recommends re-testing whenever win rates drop.
Expert knowledge extraction is the process of pulling hidden judgement, examples, rules of thumb, and decision cues out of an expert’s work before turning them into writing. AI can help when it interviews the expert, analyses real work, structures cases, and converts raw notes into reusable knowledge cards.
Every week, I sit down to write my LinkedIn posts with 20 years of research behind me, hundreds of customer conversations in my head, and a stack of ideas worth sharing.
Then the dark screen wins.
(I like dark mode, ok?)
Not once in a while. Every week. This is how it always starts.
I will not call it writer’s block.
That phrase makes it sound romantic, like I am sitting in a French café, nibbling on my little baguette romantically in thought like a tortured soul, waiting for my muse to show up in a pretty scarf. My problem is less poetic. After 10 or 20 years of getting good at something, your ideas stop arriving as fleshed-out sentences.
The problem is that I know too much.
Sounds like a humble brag. But it’s not. It is the root of most of my writing problems, and it gets worse the better you get. Malcolm Gladwell would agree.
Expertise becomes a writing trap because the better you get, the less visible your own thinking becomes. You stop noticing the assumptions, shortcuts, and pattern recognition that make your judgment useful, so the reader gets conclusions without the path that made them credible.
When you set out (like the little Indiana Jones that you are) to become exceptional at something, no one warns you about the cost of being hypervaluable as an expert. Depth makes you useful. It can also make you miserable at explaining what you know.
You’re stuck because you’re stupid. Quite the opposite. The deeper you dig your vampire fangs into the neck of your beautiful domain, Edward, the more your assumptions fade into black. You stop seeing the system—the paradigm that you’ve slobbered up like lukewarm milk from science mama’s teats—because you have become part of the system yourself. You’re not Alex Hormozi speaking in quotable tweets at podcast interviews. You don’t have clean little sentences waiting in the idea pantry. Your knowledge lives in your timing, pattern recognition, and gut-level judgment.
That is why you cannot just write it down. Much of what you know stopped being verbal years ago.
Knowledge management scholar Ikujiro Nonaka gave this phenomenon a name: Tacit knowledge. Tacit knowledge is experience soaked in judgment. It is the surgeon who spots trouble before the monitors confirm it. The consultant who reads a room in 30 seconds. The expert who sees the flaw in a methodology while a junior researcher keeps rereading the paragraph.
You cannot squash tacit knowledge into a LinkedIn post. It doesn’t work like that.
Nonaka’s SECI model describes how knowledge moves from tacit to explicit through externalisation, meaning metaphors, analogies, and dialogue pull embodied judgment into words. Trying harder at a blank page will not do that. You need a way to drag the knowledge out before you can shape it.
Most experts solve the wrong problem. They switch note-taking apps. Buy more AI tokens. Try a new morning routine. Blame discipline. The actual problem is extraction.
The apps are innocent. You are the bottleneck. I mean that with affection. Mostly.
Your words lag behind your skill because expertise flattens knowledge into instinct. Early in your career, you explain every step because you are still learning it. Later, the steps disappear into timing, pattern recognition, and judgment that feel obvious only to you.
Early in a career, your knowledge still has grip. You learned it last week. You tested it yesterday. You explained it to anyone who would sit still long enough. Moving from novice to expert forces articulation because people keep asking you to justify your stuff.
Then you become senior.
People stop asking you to explain yourself and start asking you for advice. Simple decisions arrive like Amazon same-day delivery used to. The judgment feels instant. Your knowledge sinks into your body, and the trench between what you know and what you can write grows faster than your expertise.
This means the smarter you get, the worse the writing problem becomes. The more you deserve an audience, the harder that audience is to reach.
Expertise hides its own guts. The more you have, the less visible it becomes to you and to the people you want to reach.
Then someone asks you to write an article. Or build a course. Or share your insights on LinkedIn.
You turn into Jon Snow. Noble. Confused. Somehow still in charge of these shenanigans.
Camerer, Loewenstein, and Weber coined the term the curse of knowledge in a 1989 paper on asymmetric information. Better-informed people could not ignore what they knew, even when it cost them. Fischhoff mapped the same concept in 1975 through hindsight bias. Once people knew an outcome, they overestimated what they would have predicted before they knew it. What they knew changed what they thought could have happened.
Nickerson reviewed evidence in 1999 that people use their own knowledge as the starting point for estimating what other people know, then correct too little. Hinds gave the expert version some bite that same year in Journal of Experimental Psychology: Applied. Experts guessed how long beginners would take. They did worse than people with some experience did. Debiasing attempts did not save them.
Elizabeth Newton’s 1990 Stanford dissertation made the problem concrete. Participants tapped rhythms of well-known songs and predicted listeners would identify them 50% of the time. The actual rate was 2.5%. The tappers could hear the song in their heads. The listeners heard random fingers doing a tap dance.
You are that tapper. Sorry.
Your audience hears random knocks until you translate your knowledge for them. But this is not a skill you can just summon like a Patronus Charm. You have to construct it slowly, Hermione.
The blank page gets blamed too easily here. When, really this is more of a knockout match between tacit knowledge and the language you use. Joining that fight proves your knowledge has depth. You want to win this like Conor McGregor’s 13-second knockout of Jose Aldo did.
Writing more will not fix this when the real bottleneck is how your extract your knowledge. A habit can produce more words, but it cannot automatically recover the hidden examples, decision cues, failed cases, and assumptions that make expert writing useful.
Productivity advice loves repetition. Write every day. Build the habit. Push through resistance.
It’s true, but only works when consistency truly constrains you. Having tacit knowledge is a different beast of a problem. Telling an expert to write more is like telling Terry Fox to sprint faster with one leg. Effort is not your missing ingredient. Knowing how to extract your knowledge is.
If you’ve ever forced content onto a page without a system, you’ve likely produced one of these results:
Think of it as Eminem’s 8 Mile freeze, except you are alone at your laptop at midnight. Mom didn’t make spaghetti and the hostile crowd hasn’t even arrived yet. You still cannot start.
Change how to get things out of your brain and onto the paper before you judge what’s even on the page.
Here are eight simple methods to help you do that. You don’t need all of them. Just pick two or three that match how your ideas are stuck.
Extracting first means capturing the raw judgment before trying to shape the prose. Use AI to interview you, observe your work, analyze transcripts, compare cases, and turn messy expertise into reusable material. Editing comes after the thinking has been pulled into view.
Expert knowledge does not come out through one magic prompt. Researchers usually extract it by combining 3 families of methods. They watch the familiar task. They interview the expert. They use special tasks that make hidden judgement visible.
That sounds dry. It is also exactly what AI makes easier now.
This is the baseline because every other method needs raw material to work with.
Set a 10-minute timer and write everything down. Worries. Half-formed ideas. Client patterns. Research findings. Concepts you keep circling. Unedited. Unstructured. Messy. Capture first. Organize later. Alternatively voice record or speech-to-text this part, but I highly recommend going through the agony of writing it down by hand.
Then add the material experts usually forget to include because it feels too obvious. Slides. Emails. Meeting notes. Old proposals. Client audits. Screenshots. Loom transcripts. The documents around your work often contain the tricks your brain doesn’t even notice anymore.
Save the pile to a local folder and prompt via either Codex, Claude Co-work/Code, or Antigravity:
This borrows from Hoffman’s point about documentation analysis. Manuals, notes, and old artifacts can produce a first-pass knowledge base before any interview begins.
What often goes wrong is capturing without review. Ideas pile up and decay. A 15 to 30 minute weekly review turns the dump into an asset. You scan, promote, and discard. Without that review, your capture system becomes an archaeological dig. You will find notes from 2019 and have no memory of writing them.
The Raiders of the Lost Ark warehouse, basically. Full of important things. None of them findable when you need them.
Use this as your daily baseline. It is your foundation.
Voice capture works best when you use it during the actual task, not 3 days later when your brain has already cleaned up the crime scene.
Open a voice recorder (or even better an AI speech-to-text app like superwhisper, WisprFlow, or Monologue) while you review a draft, work through a client problem, plan a workshop, annotate a paper, or redesign an offer. Say what you notice. Say what you reject. Say which detail made you suspicious. Say the sentence you almost wrote and why you cut it.
Then upload the transcript to AI and prompt:
This adapts protocol analysis for normal people with jobs. The point is not to sound polished, but to catch the judgement while you have momentum.
Use this when the knowledge lives inside a task you perform on autopilot.
Hoffman separates familiar task analysis from interviews for a reason. Experts often explain the official version of their work. Watching the work exposes the real sequence.
Record yourself doing a small real task. Review a client page. Build an outline. Fix a broken paragraph. Make a go or no-go call on an idea. Use screen recording if the work is visual. Use audio if the work is mostly thinking.
Then ask AI:
This is real task analysis. No performance required.
Use this when you know the answer, but you can't explain how you figured it out.
Klein also points out that experts reveal more when they work through test cases, especially difficult or unusual ones. Easy cases produce slogans. Tough cases produce thinking.
Create a small bank of examples:
Then ask AI to interview you through them one at a time:
Failed cases don't warn you. They're where your argument gets tested.
Use this when your advice sounds too clean because you have removed the pain that produced it.
Most people use AI to generate content. Use it to interview you.
Unstructured interviews ramble. They can build rapport and warm up the domain, but they are weak at extracting knowledge. Structure makes the interview do work.
Prompt AI like this:
Add Feynman mode at the end:
Use this when the expertise is deep, specific, and hard to pull out alone. Anecdotally, 45-minute AI interview can surface more usable material than 10 hours of staring at a blank document.
A critical incident is a real event where something changed. For example, the client finally understood, the case broke down, the system failed, the advice worked under pressure, the audience reacted in a way you did not predict.
This works because experts remember judgement through cases. Your best reasoning often shows up in one specific incident where something went wrong, and that moment revealed the pattern you rely on.
Here's an AI prompt for it:
This adapts event recall and the Critical Decision Method for writing. The abstract topic gives you the TED Talk version. The incident gives you the usable material.
Use this when you need a story, a lesson, or a better argument.
Some knowledge only appears when you force comparison.
Take 20 to 40 raw notes, examples, claims, or stories. Ask AI to turn them into cards. Then sort them into groups. Rate each card from 1 to 5 for audience pain, originality, evidence strength, and ease of explanation.
Then ask AI:
This adapts judgement tasks, scaling tasks, and decision analysis. Sorting exposes categories. Rating exposes priority. Decision trees expose the logic your draft keeps hiding from you.
Use this when the idea feels tangled or your outline keeps turning into soup.
Tiago Forte’s progressive summarization solves the second-order problem. Capturing ideas is easy. Returning 6 months later and finding the usable signal is where the bodies are buried.
Use 4 layers.
Now add an AI layer to it.
Ask:
Most experts have piles of notes they never use because sorting them takes too much time. Progressive summarization cuts that effort before your future self is tired, rushed, and trying to convince themselves that 400 notes is a system.
Use this when you have too many notes and no fast way to turn them into writing.
Build an expert extraction system so every idea has a location to go to. Capture fast. Review weekly. Run a simple extraction session on a set schedule. The system has one job, which is to keep your expert judgment from fading before you can use it in your writing.
Ideas do not arrive as a tidy backlog. They keep coming. Research findings. Client insights. Shower thoughts. Conversations that wreck your tidy little theory before you buy lunch for the team.
You need 4 pieces in place:
You know this feeling too well. Two 30-minute extraction sessions per week are much better than 10 hours of procrastination suffering because they attack your bottleneck. Your problem is often just a missing review cycle (see also Boice, R. (1990). Professors as Writers: A Self-Help Guide to Productive Writing. New Forums Press).
AI has made conversion faster. Raw capture can become a draft in minutes. The remaining friction is getting the idea out of your head before you forget it.
This is the gap between what experts know and what they can explain on demand. The blank page looks like the enemy. Deep knowledge often loses its words before the expert tries to write.
I wrote this because I needed it myself.
I still sit down every week with ideas in my head and nothing on the page. I struggle. And I have many systems. The shower thoughts still wash off before I reach my laptop. The blank screen still has its vampire fangs deep up my neck.
My mental model changed when I named the process. Years of practice shape expertise until the words fall off. Laziness is not the culprit here. Imposter syndrome is not either, even though it loves claiming credit for the damage. This is just how expertise works.
These methods interrupt the squeeze. They force tacit knowledge back toward language because experts need different tools than beginners.
Your difficulty putting ideas on paper does not prove you have nothing to say. It proves you have been using the wrong tool. A surgeon does not use a screwdriver. An expert does not use a blank document and willpower.
You are using the wrong tool.
Experts often assume visible creators know more, think with more clarity, or have some gift for effortless output. In my experience, they have a system. Usually a boring one. The advantage is their setup.
You have 20 years of insight locked inside you. The world has not heard it because no one gave you the right extraction tools.
That ends now.
Build one inbox. Schedule one weekly review. Run one AI interview on one topic you keep avoiding.
You know more than you can say.
Build the machine that gets it out.
That is your edge.
Lennart Nacke, PhD, is Research Chair at the University of Waterloo and a full professor in human-computer interaction (HCI), gamification, and AI strategy for academics and founders. He has supervised 20 PhD students, sat on dozens of hiring panels, and published over 300 peer-reviewed papers with more than 45,000 citations. He helps serious experts build research-grade writing systems that make them known, trusted, and chosen, without the content hamster wheel, hype, or hustle.
What is expert knowledge extraction?
A structured process for converting an expert's tacit judgement into explicit, reusable writing using interviews, task observation, and case analysis.
How is this different from note-taking?
Note-taking captures what you already know in words. Knowledge extraction recovers what you know but cannot yet say.
Which AI tools work best for expert interviews?
Any frontier large language model with long context (Claude, ChatGPT, Gemini, Grok) works. The method matters more than the model.
How long does a useful AI knowledge interview take?
Plan 45 minutes for one topic with 10–12 structured questions before synthesis.
Can I skip cognitive task analysis if I just write daily?
No. Daily writing improves consistency, not recall of tacit cues. Both are needed.
Is tacit knowledge always extractable?
Not fully. SECI (Model of Knowledge Creation: Socialization, Externalization, Combination, Internalization) predicts that some tacit knowledge transfers only through shared practice. Extraction recovers the explicit-able layer.
What is the simplest first step?
A 10-minute brain dump plus one 45-minute AI interview on a topic you keep avoiding.