How I Produce the Unsupervised Learning Podcast

u87ai 1

I have a show called Unsupervised Learning, where I do Content Curation as a Service. I consume 5-20 hours of books, podcasts, and articles per week, and then curate that into a 15-30 minute summary.

A number of people have asked me about my setup, so that’s what this is. Most of these types of posts are from super podcast / audio pros who are basically giving a reference architecture.

This isn’t one of those posts. I basically fumbled my way into a workflow that seems fit for purpose, but it’s not necessarily ideal. I would appreciate any pointers from people who know what they’re doing.


The show is actually both a podcast and a newsletter, and the newsletter serves as my podcast show notes, so each podcast episode has an image and link to the newsletter.


  1. Gather content throughout the week via Safari Reading List, and Feedly’s “Read Later” functionality. I gather content by either sharing to Reading List on iPhone, right-clicking links to send them to Reading List, or within Feedly I simply bookmark items

  2. Bring all that into the week’s newsletter (I use Mailchimp)

  3. Record the show by talking through the stories in the newsletter, using Hindenburg as the input software4. I use DeVerberate to reduce echo in my small, hard floor room where I record.

  4. Export song to disk (44Khz WAV)

  5. Publish podcast in OmnyStudio

  6. Publish blog episode

  7. Publish newsletter

Recording takes 15-30 minutes, depending on how much content I have. I try to keep the shows short, but if I have something I go into depth on I can get up around 60 minutes.


A look at my view within Adobe Audition

Post production, i.e. steps 5 to 8, takes about 3 minutes or so. The vast majority of the time comes in steps 1 and 2, which happens throughout the week and probably takes 5-20 hours of collecting and organizing.


That’s my current process. Like I said, I’m not claiming it’s a good one, but I’ve been doing this since January of 2015 and this seems to be the best system I’ve had so far.

If I had to say one thing I’ve learned, it’s that sound production matters far less than:

  1. Content quality

  2. Show structure

  3. Consistency

People would rather have a bad-sounding show that’s super high-quality and that comes out every week, than a great sounding show about nothing, or one that constantly skips weeks or comes out at unpredictable times.

Anyway, I hope this all helps someone, and if you have tips on how to optimize or improve the process, do let me know.


  1. Nov 21, 2016 — Moved to Audition and OmnyStudio from Audio Hijack, Garageband, Blurbbry, and Soundcloud.

  2. Jan 15, 2017 — Moved from my Yeti mic to the Electro Voice RE27.

  3. Jan 10, 2019 — Moved to Hindenburg as my DAW.

  4. Jul 15, 2019 — Moved microphones from my Electro Voice RE27 to the Neumann U87Ai.

  5. Sep 15, 2019 — Added my Neuman KH 120A Monitor speakers.

Related posts: