David Eyk, Turtle-based 🐢🤠📚💻✝️
@eykd.bsky.social
Rhymes with "like"—Christian husband and father—Lord have mercy—Homeschool publishing engineer—Pythonista since 2.3—Strong opinions loosely held
This was mostly using Claude 3.5 Sonnet. I'm curious to see how Claude 3.7 performs for next month's batch.
March 3, 2025 at 10:27 AM
This was mostly using Claude 3.5 Sonnet. I'm curious to see how Claude 3.7 performs for next month's batch.
So far, I'm very happy with the initial results. I was able to develop a month's worth of LinkedIn content in a few days. (Building the tools took two months, but let's not talk about that!)
March 3, 2025 at 10:27 AM
So far, I'm very happy with the initial results. I was able to develop a month's worth of LinkedIn content in a few days. (Building the tools took two months, but let's not talk about that!)
Once I had five or six hub-and-spokes, it was simply a matter of scheduling the posts in Buffer, scattering the topics across multiple weeks so nothing got too repetitive.
March 3, 2025 at 10:27 AM
Once I had five or six hub-and-spokes, it was simply a matter of scheduling the posts in Buffer, scattering the topics across multiple weeks so nothing got too repetitive.
Back to the human in the loop. I review the results, pick the ones I like, and revise them further. So far, I've been getting the best results with the short post formulae. (I expect to be getting better long post results eventually. Prompt engineering is time-consuming!)
March 3, 2025 at 10:27 AM
Back to the human in the loop. I review the results, pick the ones I like, and revise them further. So far, I've been getting the best results with the short post formulae. (I expect to be getting better long post results eventually. Prompt engineering is time-consuming!)
Now, it's on to the spokes. I've got a panoply of prompts targeting different lengths and formulae for social media posts. I click a button and the application generates posts based off the hub content, including image prompts.
March 3, 2025 at 10:27 AM
Now, it's on to the spokes. I've got a panoply of prompts targeting different lengths and formulae for social media posts. I click a button and the application generates posts based off the hub content, including image prompts.
Now, insert the human into the loop. That's me! I check the sources and spend time revising the article. (Or, sometimes, I throw it away and start over.) When I'm happy with it, I schedule it on my mailing list.
March 3, 2025 at 10:27 AM
Now, insert the human into the loop. That's me! I check the sources and spend time revising the article. (Or, sometimes, I throw it away and start over.) When I'm happy with it, I schedule it on my mailing list.
Then, I take an idea that sounds intriguing and generate a hub from it. This takes the idea and runs it through a pipeline of prompts. First, it retrieves case studies using Perplexity. Then it generates a rough draft article with an outline prompt I've developed.
March 3, 2025 at 10:27 AM
Then, I take an idea that sounds intriguing and generate a hub from it. This takes the idea and runs it through a pipeline of prompts. First, it retrieves case studies using Perplexity. Then it generates a rough draft article with an outline prompt I've developed.
In this app, I create a "repository", specify the audience and style sub-prompts, and seed the repository with a ton of one-line topic ideas.
The application runs each topic idea through a custom prompt that expands on the idea based on the audience.
The application runs each topic idea through a custom prompt that expands on the idea based on the audience.
March 3, 2025 at 10:27 AM
In this app, I create a "repository", specify the audience and style sub-prompts, and seed the repository with a ton of one-line topic ideas.
The application runs each topic idea through a custom prompt that expands on the idea based on the audience.
The application runs each topic idea through a custom prompt that expands on the idea based on the audience.
The advantage is that you have a lot scheduled out ahead of time. But the challenge is you have to get over that first hump!
So, I built myself a custom web app. It's what I do, after all. :)
March 3, 2025 at 10:27 AM
The advantage is that you have a lot scheduled out ahead of time. But the challenge is you have to get over that first hump!
So, I built myself a custom web app. It's what I do, after all. :)
The challenge with hub-and-spokes is that you don't want to be hammering on the same topic for days on end. So you have to build out a library of these ahead of time so you can mix up the spokes.
March 3, 2025 at 10:27 AM
The challenge with hub-and-spokes is that you don't want to be hammering on the same topic for days on end. So you have to build out a library of these ahead of time so you can mix up the spokes.
The outcome I was aiming for with this project was a Justin Welsh ContentOS-style hub-and-spokes strategy, where a weekly newsletter (the "hub") gets spun off into a variety of derivative "spoke" posts.
March 3, 2025 at 10:27 AM
The outcome I was aiming for with this project was a Justin Welsh ContentOS-style hub-and-spokes strategy, where a weekly newsletter (the "hub") gets spun off into a variety of derivative "spoke" posts.
My hypothesis: it would be possible to use an AI-assisted workflow to produce strategic marketing content that *wasn't* mere AI slop.
March 3, 2025 at 10:27 AM
My hypothesis: it would be possible to use an AI-assisted workflow to produce strategic marketing content that *wasn't* mere AI slop.
Come January, I decided to put my avoidance to good use by building out a human-in-the-loop content-generation pipeline.
March 3, 2025 at 10:27 AM
Come January, I decided to put my avoidance to good use by building out a human-in-the-loop content-generation pipeline.
The main problem with my LinkedIn strategy last year was that I wasn't posting my own content. I love to write, but sitting down to write the sort of thing I needed to write just felt miserable. Cue the avoidance!
March 3, 2025 at 10:27 AM
The main problem with my LinkedIn strategy last year was that I wasn't posting my own content. I love to write, but sitting down to write the sort of thing I needed to write just felt miserable. Cue the avoidance!
But I'm going to keep up the in-person networking. It works! And I think I'm about to become very busy.
December 5, 2024 at 11:39 PM
But I'm going to keep up the in-person networking. It works! And I think I'm about to become very busy.
I'll continue working LinkedIn. I still think it's a good long-term play.
I'm in the middle of developing my content strategy which has thus far been severely lacking. I'll update once I've put this into play.
I'm in the middle of developing my content strategy which has thus far been severely lacking. I'll update once I've put this into play.
December 5, 2024 at 11:39 PM
I'll continue working LinkedIn. I still think it's a good long-term play.
I'm in the middle of developing my content strategy which has thus far been severely lacking. I'll update once I've put this into play.
I'm in the middle of developing my content strategy which has thus far been severely lacking. I'll update once I've put this into play.
By the grace of God (and plenty of coaching from my friend), I'm starting to get the hang of networking.
What's more, I'm already starting to see results. I sent out a proposal today for my first lead in months, and I'll be sending out another tomorrow!
What's more, I'm already starting to see results. I sent out a proposal today for my first lead in months, and I'll be sending out another tomorrow!
December 5, 2024 at 11:39 PM
By the grace of God (and plenty of coaching from my friend), I'm starting to get the hang of networking.
What's more, I'm already starting to see results. I sent out a proposal today for my first lead in months, and I'll be sending out another tomorrow!
What's more, I'm already starting to see results. I sent out a proposal today for my first lead in months, and I'll be sending out another tomorrow!
I've talked with him a lot about the value of doing hard things.
It's easy to talk, but I sometimes wondered if I was modeling this well enough for him.
When I got back from my first networking event, I told him that I'd found my hard thing. 😄
It's easy to talk, but I sometimes wondered if I was modeling this well enough for him.
When I got back from my first networking event, I told him that I'd found my hard thing. 😄
December 5, 2024 at 11:39 PM
I've talked with him a lot about the value of doing hard things.
It's easy to talk, but I sometimes wondered if I was modeling this well enough for him.
When I got back from my first networking event, I told him that I'd found my hard thing. 😄
It's easy to talk, but I sometimes wondered if I was modeling this well enough for him.
When I got back from my first networking event, I told him that I'd found my hard thing. 😄
My son has been taking lessons in Brazilian Jiu-jitsu for the last few years. BJJ is very much about learning to willingly subject yourself to uncomfortable situations. (And learning to get out of them!)
It was extremely hard for him at first, but I've been impressed with his perseverance.
It was extremely hard for him at first, but I've been impressed with his perseverance.
December 5, 2024 at 11:39 PM
My son has been taking lessons in Brazilian Jiu-jitsu for the last few years. BJJ is very much about learning to willingly subject yourself to uncomfortable situations. (And learning to get out of them!)
It was extremely hard for him at first, but I've been impressed with his perseverance.
It was extremely hard for him at first, but I've been impressed with his perseverance.
I'm a confirmed introvert. I spend 8+ hours a day in an 8x8' home office, and I love it.
Diving into a room of 10–40 people and striking up conversations is, shall we say, intensely uncomfortable.
Diving into a room of 10–40 people and striking up conversations is, shall we say, intensely uncomfortable.
December 5, 2024 at 11:39 PM
I'm a confirmed introvert. I spend 8+ hours a day in an 8x8' home office, and I love it.
Diving into a room of 10–40 people and striking up conversations is, shall we say, intensely uncomfortable.
Diving into a room of 10–40 people and striking up conversations is, shall we say, intensely uncomfortable.