Build a newsletter and podcast site with Replit
Another action packed week! Task: Another action packed week! This week was a roller coaster, but ultimately very productive. As you may recall, I am heading on a journey to unpack how AI is changing how we live and work, and specifically, how startups get built and funded (while I am in parallel figuring out where I will fit). As I have been making the rounds, it is humorous how many other middle-aged founders like me have been thinking similarly :) I had already determined that, while I will lead with a podcast and newsletter again, from day zero I want to be a data company and be thoughtful about how to dogfood AI tools internally to mine insights from our content and seek insights externally that we can bring back into the fold along the way. What I hadn’t thought about is that there are opportunities to get my hands dirty in the interim that are far more basic than that. I went to set up a squarespace website to house the pod and newsletter, and realized I should be “coding” one with AI tools, whether I need to or not. I started with replit, had my mind blown in the first 10 minutes when it spit out a clickable prototype, and spent the next 10+ hours getting spun around in circles as I tried to get it to loaded on my custom url. The kind folks at Val Town swept in and saved the day, offering to pair program with me using their tools to get the website up and running quickly (which we did). You can find my shiny new (MVP version of a) website here. I then turned my attention to the podcast, and got a (starter) logo made using canva, recorded a trailer using riverside, and edited it using Descript. These may seem like little things to you, but I assure you I am already getting my hands dirtier this time around than I ever have in the past :) I booked the first show guest (which is a great one!), and I will soon be off and running with the podcast content engine. I can’t wait! My schedule is getting loaded up with hammers from the space to meet with. My hands are getting dirtier every day with these emerging tools. And the pod is getting closer to lift off, meaning I can get the content machine really humming soon. Seems like a lot of progress to me! I do still think long-term I will want a junior engineer who LOVES working with AI tools to partner with. But no rush on that, and there is great value to me getting my hands dirty in these early innings, however painful and “inefficient” it may feel. Focus over the next week: getting the pod fully ready (recording trailer, stock intro, stock outro, getting the first few guests scheduled, getting the hang of editing using Descript) getting the pod fully ready (recording trailer, stock intro, stock outro, getting the first few guests scheduled, getting the hang of editing using Descript) Figuring out how to populate my new website with episodes as they come (Simplecast may do it automagically, but I haven’ t tested yet) Figuring out how to populate my new website with episodes as they come (Simplecast may do it automagically, but I haven’ t tested yet) Starting to map out a gameplan to start to incorporate the tools themselves more into the journey Starting to map out a gameplan to start to incorporate the tools themselves more into the journey Maybe starting to put some feelers out for some junior engineering help, but again, probably premature Maybe starting to put some feelers out for some junior engineering help, but again, probably premature Asks: guest ideas of people you want to hear/learn from guest ideas of people you want to hear/learn from A peer group of other founder-types building software using these tools A peer group of other founder-types building software using these tools Smart podcast people that may have feedback around all aspects of show production Smart podcast people that may have feedback around all aspects of show production Also, I am going to go ahead and call my shot here for one potential way this could all play out. I could start down the path of podcast/newsletter. Could dogfood internal tools that help us make this flywheel hum better over time. When we build ones that add value to us, potentially look at enabling others to use them for their journeys (in whatever category they are seeking to more deeply understand) and evolve into a product company over time, all dogfooded by us along the way. Lots of other ways this could head as well, but that sure seems like a possible contender.; tools: Replit Agent. Human role: The original post does not say how a person stayed involved.; limitations: Replit generated a prototype quickly but failed to handle custom domain deployment without significant manual effort or external help. The workflow relies on multiple distinct tools (Canva, Riverside, Descript) rather than a single integrated platform. The author admits the process felt 'inefficient' and 'painful' initially, suggesting that AI tools may not fully replace engineering expertise for complex deployment tasks..
Tools
Published on Agent Work Atlas:
Overview
Another action packed week!
Workflow
Another action packed week! This week was a roller coaster, but ultimately very productive. As you may recall, I am heading on a journey to unpack how AI is changing how we live and work, and specifically, how startups get built and funded (while I am in parallel figuring out where I will fit). As I have been making the rounds, it is humorous how many other middle-aged founders like me have been thinking similarly :) I had already determined that, while I will lead with a podcast and newsletter again, from day zero I want to be a data company and be thoughtful about how to dogfood AI tools internally to mine insights from our content and seek insights externally that we can bring back into the fold along the way. What I hadn’t thought about is that there are opportunities to get my hands dirty in the interim that are far more basic than that. I went to set up a squarespace website to house the pod and newsletter, and realized I should be “coding” one with AI tools, whether I need to or not. I started with replit, had my mind blown in the first 10 minutes when it spit out a clickable prototype, and spent the next 10+ hours getting spun around in circles as I tried to get it to loaded on my custom url. The kind folks at Val Town swept in and saved the day, offering to pair program with me using their tools to get the website up and running quickly (which we did). You can find my shiny new (MVP version of a) website here. I then turned my attention to the podcast, and got a (starter) logo made using canva, recorded a trailer using riverside, and edited it using Descript. These may seem like little things to you, but I assure you I am already getting my hands dirtier this time around than I ever have in the past :) I booked the first show guest (which is a great one!), and I will soon be off and running with the podcast content engine. I can’t wait! My schedule is getting loaded up with hammers from the space to meet with. My hands are getting dirtier every day with these emerging tools. And the pod is getting closer to lift off, meaning I can get the content machine really humming soon. Seems like a lot of progress to me! I do still think long-term I will want a junior engineer who LOVES working with AI tools to partner with. But no rush on that, and there is great value to me getting my hands dirty in these early innings, however painful and “inefficient” it may feel. Focus over the next week: getting the pod fully ready (recording trailer, stock intro, stock outro, getting the first few guests scheduled, getting the hang of editing using Descript) getting the pod fully ready (recording trailer, stock intro, stock outro, getting the first few guests scheduled, getting the hang of editing using Descript) Figuring out how to populate my new website with episodes as they come (Simplecast may do it automagically, but I haven’ t tested yet) Figuring out how to populate my new website with episodes as they come (Simplecast may do it automagically, but I haven’ t tested yet) Starting to map out a gameplan to start to incorporate the tools themselves more into the journey Starting to map out a gameplan to start to incorporate the tools themselves more into the journey Maybe starting to put some feelers out for some junior engineering help, but again, probably premature Maybe starting to put some feelers out for some junior engineering help, but again, probably premature Asks: guest ideas of people you want to hear/learn from guest ideas of people you want to hear/learn from A peer group of other founder-types building software using these tools A peer group of other founder-types building software using these tools Smart podcast people that may have feedback around all aspects of show production Smart podcast people that may have feedback around all aspects of show production Also, I am going to go ahead and call my shot here for one potential way this could all play out. I could start down the path of podcast/newsletter. Could dogfood internal tools that help us make this flywheel hum better over time. When we build ones that add value to us, potentially look at enabling others to use them for their journeys (in whatever category they are seeking to more deeply understand) and evolve into a product company over time, all dogfooded by us along the way. Lots of other ways this could head as well, but that sure seems like a possible contender.
Reported result
The website MVP was successfully launched with a custom URL after external technical assistance. The podcast trailer was recorded and edited, and the first guest was booked. However, the source does not provide quantitative metrics on audience growth, revenue, or time saved, and notes that the initial Replit deployment attempt was inefficient.
Human role
The original post does not say how a person stayed involved.
Limitations
Replit generated a prototype quickly but failed to handle custom domain deployment without significant manual effort or external help. The workflow relies on multiple distinct tools (Canva, Riverside, Descript) rather than a single integrated platform. The author admits the process felt 'inefficient' and 'painful' initially, suggesting that AI tools may not fully replace engineering expertise for complex deployment tasks.
Try it
Act as a technical assistant for a new podcast launch. I need to build a website and produce a trailer. First, use Replit to generate a basic HTML/CSS/JS prototype for a site hosting [replaceable: podcast name] and [replaceable: newsletter description]. Do not attempt to deploy to a custom domain yet; just provide the code structure. Second, guide me through creating a logo in Canva using keywords [replaceable: brand vibe]. Third, outline a script for a 2-minute trailer to be recorded in Riverside. Fourth, provide a checklist for editing the audio in Descript. Before I publish any code or send any emails, I will review the outputs. If Replit deployment fails, suggest a manual workaround or identify if external help is needed. Measure success by having a clickable local prototype and a recorded audio file ready for review.
Source reports are not independently verified. Publication date is the catalog publication date, not the original article date.