You already know the paper should come before email. The hard part is acting like you know this deeply, intricately, at 8:12 AM. when the inbox feels like duty to serve like your sticky habit demands it, and the draft feels optional at most.

Bad (🤷‍♂️): Another day, another second brain in AI. Better: Any writing system that gets your real sentence onto the page before lunch time.

The scarce resource in a lab is not intelligence. It is uninterrupted attention aimed at one hard question.

Write the short public note that explains why the work matters. Write the conference talk in plain language. Write the grant one-pager. Write the paper abstract. Write the lab update. If an idea only works in one format, it is not clear yet. Clear research thinking gets translated well.

If your lab’s AI use still feels clumsy, good. That means you are early enough to build shared workflows before bad habits harden: • One methods consistency check • One figure/table explanation pass • One literature deep research prompt • One pass to figure out what Reviewer 2 would say

My default AI workflow before I trust a draft section: 1. Ask AI to attack it like a skeptical reviewer. To find the weak claims, missing baselines, fuzzy contributions, overfit conclusions 2. Only after that, ask for a tighter rewrite that preserves my argument and citations

Underrated writing tip for researchers: In any sentence, make the last word carry the heavy meaning. Weak: “The main contribution of this paper is a new evaluation method.” Stronger: “This paper’s main contribution is a new evaluation method.” Even stronger:

Most senior researchers don’t lack expertise. They lack a weekly distribution system for it. The Write Insight is my weekly newsletter for researchers and research-origin experts who want: • Clearer public writing • AI workflows that protect rigour

How to use the PASTOR framework in academic writing: 1. PROBLEM: Identify the research gap 2. AMPLIFY: Turn up the volume on that gap 3. STORY: Share a mini-story 4. TRANSFORM: Use testimonials academia style 5. OFFER: Make your contribution 6. RESPONSE: Call for future action Done.

Harsh truth for researchers using AI: If your prompt is just "make this sound better," you are asking the model to make its own judgment. Better workflow: 1. Paste the claim, evidence, and constraint 2. Ask for counterarguments, missing citations, and weak inferences

If you are just starting to write about your research in public and almost no one is reading it, this is your beginner advantage. It’s like running studies in a private research lab before a big site visit.

For whoever needs to hear this: Your first 50 public posts will be uneven. Almost no one will see them. The few who do will not remember the awkward ones. That is your beginner advantage. You get to build the skill while the stakes are still low. Embrace the easy reps.

The easiest way to get ahead as a researcher in the AI era is to start producing in public. Every day. 45 focused minutes. Take one idea from your research, teaching, or lab conversations and put a clear version of it into the world. Hit publish before it feels finished.

You should use AI on two kinds of research work: 1. What is now routine for you, but used to take hours 2. What is still hard for you, where better questions would help Use AI to automate the routine work. Use AI to challenge the hard work. If you only do the first, you get speed.

It has never been easier for researchers to build a public body of work: • Most 'thought leadership' in your domain is soft, recycled BS • Most good papers get read by a tiny fraction of the field • Most people scroll for hours and still feel behind

My deep conviction is that writing your thoughts in public is one of the highest-leverage habits a researcher can build. • You get faster feedback • You become a better teacher • Your arguments become clearer • You build stronger collaborations • Your expertise becomes more visible

Most people want the keynote. That sexy recognition. They don't see the work you've put in to get it. They see the citations, the title, the chair position. Many assume public expertise just ships with completing a PhD. Far from true. What they don't see:

5 steps to get paid for research expertise without becoming a marketing geek: 1. Find a painful problem your field keeps creating for labs, companies, or institutions 2. Solve it in your actual work 3. Publish short pieces that make you understand why the problem is impossible to ignore

The best AI habit I’ve found for research clarity: Before bed, write down one open research problem in one sentence. Heck, talk it into an app that transcribes it.

If the paper is the priority, protect the first hours of the day for it. If the grant is the priority, protect the first hours of the day for it. If the experiment design is the priority, protect the first hours of the day for it.

In 2010, my work lived almost entirely inside papers, grants, and closed spaces. After years of publishing, I still kept running into the same problem: People who should have known my work never read it.

Time freedom = spending your best hours on work only you can do If your week is packed with work below your pay grade, choose between these two: 1. Change what you are doing, until your calendar matches your expertise 2. Change what you want, until committee minutes feel like the point

One business fear I wish more senior researchers admitted: You worry that writing online will make you look less serious. Yet funders, clients, collaborators, students, and conference organizers can only act on work they encounter. Your credibility already exists. Give it a public trail.

If I were starting a public writing practice as a senior researcher, I would: 1. Capture the questions that recur in my research, teaching, and service 2. Publish short answers before they become polished papers 3. Expand the strongest ideas into one useful newsletter each week

It has literally never been easier to build a network: • Most experts in your field publish nothing online • One useful idea a week puts you ahead of 95% of them • Every reply you answer is a warm introduction • The idea does the introducing while you sleep You just need a tiny bit of consistency.

Research happens in the literature. Research happens in lab meetings. Research happens in failed studies. Research happens in reviewer comments. Research happens in conversations with students. Kill the idea that research only happens when you sit down to write a paper.

My confidence in public got a lot better when I started viewing confidence as evidence I get rather than a mood. Here's how I built that system: • Keep a running document of every completed project, published paper, invited talk, and peer validation moment

You don't get ready by waiting. You get ready by shipping at B+. Define the bar in writing: 1. Clear claim 2. One concrete example 3. No factual errors. 4. Borrow an external deadline. 5. Run a 10-minute pre-flight. 6. Keep a 48-hour fix window.

The 5 best applications of Hermes Agent: Owning persistent knowledge networks Owning automated research synthesis Owning scheduled pipeline execution Owning parallel software assembly Owning custom skill generation

Someone disagrees with your work online. Your brain wants to fight or flee. Neither helps. Sort your critics into one of three bins instead: