I believe the blank page approach is better suited for writing fiction or artistic creative work. When your focus is rather creative problem-solving, meaning that you aim to help people solve problems they might have, in a creative way, a more data-backed approach wins. Cause otherwise, how well can you respond to people's actual problems?
I love that you kept the creative wrestling but redirected it toward better problems. It reminds me of how good research doesn't eliminate the hard thinking, it just ensures you're thinking hard about questions worth answering.
This is great, Wyndo! Within the last month, I also moved the biggest part of my newsletter creation to a customized Claude project I use weekly. However, you gave me some great new ideas to try and add to it!
My newsletter is just over a month old… so I’m resisting automation for now. I’m a big believer in “feeling the pain first” only then automation becomes truly useful. Would you do it differently?
P.S. I’m already dreaming up some context-specific systems for later.
if u dont have any strong data yet, i'd suggest to follow your full intuition, write anything that excites u, there's nothing better than that. later on once you have a better understanding about what to write and what resonate most with readers, then combine with automation :)
WOW. This is extraordinary. Definitely going to implement this. The more we can connect AI to real data, the better we can have it perform extraordinary work for us.
This is such a great system, Wyndo! The way Claude helps you spot high-performing content is seriously impressive, I’d probably click on some of those suggested titles if I saw them from a stranger.
I’ve been sitting on a long list of ideas I want to write about, but I really like your approach of filtering by common pain points. That feels so smart and practical!
"Claude analyzes comments across my newsletters and social posts to identify what my audience actually cares about. Instead of guessing what readers want, I let their actual words guide my content direction." -> this is great! Do you upload the newsletter and comments manually, or what's your workflow here?
Yeah I do it manually every week, but since I use Claude Code, I can ask AI agent to scrape my newsletter post and put it inside my repo, but I manually update the comment by copy pasting it, i know this is hassle :)
Thanks for letting me know! I just started setting it up and love the step-by-step guide you've provided. Here's a revised prompt I used for step 1, in case helpful for you :)
#ROLE
Act as a research assistant with access to my Gmail. Your specialty is spotting high-value insights at the intersection of AI and [your focus topic - could also be the viral article themes you've identified in your dashboard!].
# TASK
Search my Gmail inbox for newsletters received in the last 10 days. Identify and summarize any newsletters that contain *relevant, insightful information* on:
- [key theme 1]
- [key theme 2]
- [key theme 3]
# CONSTRAINTS
- Search only newsletters.
- Be thorough but concise; focus on idea density.
- Ignore generic updates or news not directly tied to the themes above.
- Do not include personal or sensitive information.
# OUTPUT
For each relevant newsletter found:
- **Subject line + sender**
- **Date received**
- **Key idea(s)**: 1–3 bullets on what’s valuable or novel
-**
- **Possible angle for The Lifelong Learning Newsletter**: 1 sentence on how to dig deeper
# TARGET GROUP
Your audience: intellectually curious professionals (founders, coaches, professors, product managers) seeking practical, evidence-based tools to optimize thinking, learning, and leading. They value clear, actionable takeaways.
# GOAL
Deliver a short list of fresh, idea-rich newsletter insights to explore for future editions of The Lifelong Learning Newsletter, aligning with the mission: *help professionals leverage AI to boost intellectual capacity while strengthening critical thinking and human judgment.*
Great framework, Wyndo!
I believe the blank page approach is better suited for writing fiction or artistic creative work. When your focus is rather creative problem-solving, meaning that you aim to help people solve problems they might have, in a creative way, a more data-backed approach wins. Cause otherwise, how well can you respond to people's actual problems?
yeah i think it's true. listening to readers matter more in our line of work and AI unlocks its possibilities :)
I love that you kept the creative wrestling but redirected it toward better problems. It reminds me of how good research doesn't eliminate the hard thinking, it just ensures you're thinking hard about questions worth answering.
true, with AI, we'll have more options to choose, we will wrestle with more ideas, the hard thing becomes how to judge which options worth exploring!
This is great, Wyndo! Within the last month, I also moved the biggest part of my newsletter creation to a customized Claude project I use weekly. However, you gave me some great new ideas to try and add to it!
Looking forward to see creative ways you are going to use it Joel!
You can also run a poll to ask your community… lazy mode 😅
feeling ashamed to admit i haven't use substack's poll yet, gonna try this soon 😄
I really enjoyed seeing your approach! 🙋🏻♀️
My newsletter is just over a month old… so I’m resisting automation for now. I’m a big believer in “feeling the pain first” only then automation becomes truly useful. Would you do it differently?
P.S. I’m already dreaming up some context-specific systems for later.
if u dont have any strong data yet, i'd suggest to follow your full intuition, write anything that excites u, there's nothing better than that. later on once you have a better understanding about what to write and what resonate most with readers, then combine with automation :)
Thought so, too. Thank you!!:)
WOW. This is extraordinary. Definitely going to implement this. The more we can connect AI to real data, the better we can have it perform extraordinary work for us.
AI makes more sense if it knows our data, thats why tool integration has been a game-changer for me.
This is such a great system, Wyndo! The way Claude helps you spot high-performing content is seriously impressive, I’d probably click on some of those suggested titles if I saw them from a stranger.
I’ve been sitting on a long list of ideas I want to write about, but I really like your approach of filtering by common pain points. That feels so smart and practical!
Thanks Jenny!
I feel like with AI, we are easily surrounded with ideas to explore, now the question is which one worth pursuing?
Awesome stuff! Lots of great ideas for implementation here. Never knew Claude could integrate with so much.
Claude is wild for app integration these days
And FYI - your calendly link is not working
fixed, thanks for letting me know
Great article! Thank you
"Claude analyzes comments across my newsletters and social posts to identify what my audience actually cares about. Instead of guessing what readers want, I let their actual words guide my content direction." -> this is great! Do you upload the newsletter and comments manually, or what's your workflow here?
Yeah I do it manually every week, but since I use Claude Code, I can ask AI agent to scrape my newsletter post and put it inside my repo, but I manually update the comment by copy pasting it, i know this is hassle :)
Thanks for letting me know! I just started setting it up and love the step-by-step guide you've provided. Here's a revised prompt I used for step 1, in case helpful for you :)
#ROLE
Act as a research assistant with access to my Gmail. Your specialty is spotting high-value insights at the intersection of AI and [your focus topic - could also be the viral article themes you've identified in your dashboard!].
# TASK
Search my Gmail inbox for newsletters received in the last 10 days. Identify and summarize any newsletters that contain *relevant, insightful information* on:
- [key theme 1]
- [key theme 2]
- [key theme 3]
# CONSTRAINTS
- Search only newsletters.
- Be thorough but concise; focus on idea density.
- Ignore generic updates or news not directly tied to the themes above.
- Do not include personal or sensitive information.
# OUTPUT
For each relevant newsletter found:
- **Subject line + sender**
- **Date received**
- **Key idea(s)**: 1–3 bullets on what’s valuable or novel
-**
- **Possible angle for The Lifelong Learning Newsletter**: 1 sentence on how to dig deeper
# TARGET GROUP
Your audience: intellectually curious professionals (founders, coaches, professors, product managers) seeking practical, evidence-based tools to optimize thinking, learning, and leading. They value clear, actionable takeaways.
# GOAL
Deliver a short list of fresh, idea-rich newsletter insights to explore for future editions of The Lifelong Learning Newsletter, aligning with the mission: *help professionals leverage AI to boost intellectual capacity while strengthening critical thinking and human judgment.*
thank you! great prompt :)
Brilliant framework. I’ve found the same thing. AI becomes exponentially more useful when it’s trained on your voice and past content.
no wonder context engineering gets trendy these days!
Is there a set of prompts that could take all my saved many Google bookmarked article links and put the article in my refworks.