In 2020 I started writing a novel. It was an ill-fated endeavour, but I learned a lot along the way.

One of the things I learned is that it's useful to have a separate "space" for writing.

I used to cycle into my city office on Eagle St., then take my laptop and a coffee somewhere else to write. There was this one particular spot, a chair in the lobby of an office building, that had people bustling all around and a view of the river. I'd put in my headphones, play a very specific playlist, and write.

I don't live in the city anymore, and it's not so practical to drive 10 minutes down the road to the closest café every time inspiration strikes. So instead, I unplug my laptop from it's charger, and move to my writing chair.

We are spatial thinkers. We think with our mind, our physical body, and our environmental context. Our brains reorient and reorganise each time we move from one place to another.

If you'd like to write more (and you should!), try sitting somewhere new.

I've been having an ongoing argument with a friend about AI's role in recruiting. Whether you're a job-seeker or employer, things are getting pretty dire.

If you're looking for a job, one in five advertised positions are "ghost jobs" – fake, or already filled. The number of ghost jobs are probably much higher on LinkedIn, where recruiters are priming their talent pools and employers are gaming the algorithm.

If you're looking for an employee, you're already drowning in AI-generated CVs and cover letters. From the Financial Times:

Candidates are turning increasingly to generative AI — the type used in chatbot products such as ChatGPT and Gemini to produce conversational passages of text — to assist them in writing their CVs, cover letters and completing assessments.

Estimates from employers and recruiters who spoke to the Financial Times, as well as multiple published surveys, have suggested the figure is as high as 50 per cent of applicants.

A “barrage” of AI-powered applications had led to more than double the number of candidates per job while the “barrier to entry is lower”...

What's the solution to this problem? According to my friend: MORE AI!

• Employers can use better AI tools to identify and weed out the AI slop
• Employers can use AI to test the proficiencies and capabilities of applicants
• Employers can use AI to automate early interviews
• Job-seekers can use AI to improve their CVs
• Job-seekers can use AI to tailor their applications to the job description
• Job-seekers can use AI to receive personalised career guidance and skill development recommendations

It doesn't take a crystal ball to see how this whole thing just loops in on itself. It's the enshittification of the job market. Employers get better at identifying AI, job-seekers get better at hiding it. Employers get better at funnelling applicants with AI, job-seekers get better at staying on the list. Round and round it goes… until?

Well, until someone pulls the plug on AI recruitment entirely. Like Anthropic, one of the world's leading AI companies, does:

AI Policy for Application: While we encourage people to use AI systems during their role to help them work faster and more effectively, please do not use AI assistants during the application process. We want to understand your personal interest in Anthropic without mediation through an AI system, and we also want to evaluate your non-AI-assisted communication skills. Please indicate 'Yes' if you have read and agree.

We're in the "storming" phase of AI. We haven't figured out how it fits into our lives yet. There are no guarantees LLMs will take over every aspect of our lives, nor that they will replace people's jobs. The more I use AI tools in my day-to-day work, the more I appreciate real contact with real people.

The next big trend to land in our adscapes will be imperfectly, strangely, gloriously human.

I'm building an app called Pita. It makes publishing and managing your podcasts easy (and pretty). This is the second update. The first is here.

One of the problems with hosting your own podcast is that you still need the audio files to be served from somewhere. Depending on the podcast, these files can be quite large, especially after you're 50 or more episodes in. You also need to host your RSS file somewhere (though this can be somewhere different). And most of the time, you'll want a simple marketing website too – maybe one that has your podcast episodes listed with an audio player.

That's a lot – and that's why all these paid SaaS options exist (Buzzsprout, Blubrry, Podbean, Libsyn, the list goes on…)¹. But what if you could do it all for free?

Well you can! There's a bit of setup involved, but here's how I've got things setup for the Digging for Fire Podcast.

1. Audio hosting: the Internet Archive offers free mp3 hosting!
2. Website hosting: you can host a static website for free
   • Use Jekyll or Hugo to create your blog, with each podcast episode as a post
   • Static hosting like GitHub Pages or Cloudflare Pages to build and serve the website
3. Podcast RSS hosting: You can modify Jekyll or Hugo to serve an RSS file for your podcast episodes

---

The goal for Pita is to automate this whole process. Connect your Internet Archive account and your Cloudflare account, and Pita will look after the rest.

In the meantime, I've hacked together some scripts.

Each new podcast episode goes through a four step process:

1. AI cleanup and edit

I've been using Auphonic for this. It does a great job of getting rid of background noise and reverb. This is essential for me because I record in a dome in the rainforest, so I get lots of both!

2. Prepare the MP3

Whether I use Quicktime or Voice Memos to record, the output is always a plain MP4. I've been using ffmpeg to create a 64k mono mp3, and embed the episode artwork, title, and podcast name.

3. Upload to Internet Archive

There's a great command line utility called IA which does the heavy lifting for me here. It uploads the episode to archive.org, and returns a URL. I copy this URL into the front-matter of my Hugo post.

4. Get additional file data

The podcast XML standard also has an option for filesize and duration, to help podcast players show the episode correctly without needing to download it. This last script gives me the data I need, and I also copy that into my front-matter for the episode post.

Then, in Hugo, my front-matter (a bit of meta data text at the top of each markdown post) looks like this:

---
title: "🥙 Pita: Devlog 1"
date: 2025-01-20T20:30:00+1000
lastmod: 2025-01-20T20:30:00+1000
audio: https://archive.org/download/digging-for-fire/pita-devlog-1.mp3
audiolength: 3721724
audioduration: 00:07:31
categories:
  - Apps
---

Thanks to great RSS support in Hugo, I was able to add a separate RSS feed to my blog which only includes new posts that have the audio included in the front-matter.

And that's it! A completely free podcast.

---

So my next step is going to be pulling all that into a very basic Mac app. It should move an audio file through a workflow of:

1. Collect episode info from user
2. AI Edit
3. MP3 Packaging
4. Upload MP3 to IA
5. Create a GitHub PR to a Jekyll / Hugo repo with the new episode

This will still require manually setting up the site to begin with, and auto-deploying it via GitHub / Cloudflare Pages.

My plan is to test this whole thing on a new podcast network that Joshua, Jonathan, and I are working on: Pretty Good. My goal is to get it to a point where trusted friends can run the app on their machine and have it publish episodes for them.

On the 25th of February in 2015, I backed what would become my most beloved kickstarter project. The Pebble Time smart watch.

I chose the Pebble Time Round edition, which I still believe to this day is the most beautiful smart watch that's ever been built. I was a long-time Apple Watch holdout simply because Apple didn't make it as a circle.

PebbleOS is pure delight. My memories of the software are laced with nostalgia, and it's only 10 years ago! It had this quirky interface that was as fun and beautiful as it was unique. It covered all the bases for me in a Smart Watch:

• Lightweight
• Multi-day battery
• Easily readable outdoors
• Extensible (great developer API and documentation)
• Notifications

The interface for this thing was genius. It's a watch, right? So, the buttons all relate to time. There's a column of three buttons down the side of the watch. The top one is for "things that already happened", the middle on is for "things that are happening now", and the bottom one for "things that are coming up".

I loved this thing so much. Here's a product shot featuring my newborn (at the time) daughter.

And then… acquisition happened. Fitbit stepped in after Pebble filed for insolvency. Then Google bought Fitbit. The founder, Eric Migicovsky, went on to start Beeper, which was then acquired by Automattic. And my beloved Pebble moved from gadget of the year to the annals of techno-history.

And another thing happened. The Apple Watch came along. It solved a very different set of problems. I love my Apple Watch, but I've got a list of complaints, too.

• It's heavy
• It could have better battery life
• WatchOS is kind of terrible
• Something has always bothered me about the always-on backlit screen
• I get a bit bored of the watch faces

The Apple Watch is good, not great, but nobody makes anything quite as good as the Pebble once was. So Apple Watch is now my daily driver.

But hold the phone! It's just been announced that the Pebble is making a comeback!¹ Turns out, Google open-sourced PebbleOS, a community (aptly named Rebble) has been maintaining an ecosystem of apps and watchfaces, and Eric Migicovsky is down for giving it another spin.

I'm so pumped for this². The particular set of tradeoffs and design principles that Pebble offers is perfect for me. I have a bit of a wishlist for the new Pebble:

• Bring the Round edition back!
• Some basic fitness tracking (integrating with Apple Health)
• Better, brighter, colour E Ink screen
• (Never gonna happen) NFC Payment support

And I want to point out a part of Eric's post that I think says a lot. What's going to be different this time? No VC funding.

This time round, we’re keeping things simple. Lessons were learned last time! I’m building a small, narrowly focused company to make these watches. I don’t envision raising money from investors, or hiring a big team. The emphasis is on sustainability. I want to keep making cool gadgets and keep Pebble going long into the future.

---

Here's another product shot, this time for Apple Watch. This is a photo of my son just moments after he was born!

Try asking DeepSeek about Tiananmen Square or Uyghur Muslims or Taiwanese Independence. I'm not using it. ChatGPT has better voice recognition and can remember things about me. I like that. Can't login with Apple over the web, just in the app. It should be banned like TikTok. Exporting Chinese propaganda to the world. Something there about some level of censorship being useful and good, but this crosses a line. What line? Political? Historical? First piano lesson this week. First in 30 years or so. Went really well, playing some neoclassical pieces (early 20th century French stuff) and trying to disconnect my left and right hands. Pulling apart neurons by sheer force of will. Same on drums, man, this A Funky Thesaurus by Charles Dowd is pure groove. So hard but once you get it it's enlightenment. Also trying to learn to read that bass clef. Never quite got the hang of it. Duolingo Music has been really helpful for reading music, surprisingly. I've been sick. Slept most of yesterday. Fever and talking nonsense. Practicing my inversions up and down the piano especially for that top Circle of Fifths quarter FCG and their corresponding minors DAE. Key signatures man, one day I'll be able to look at a piece of music and see a whole row of ♭s and go ah, this is written in E♭m. Pool needs a clean. Caravan too. Garden needs weeding. Too hot to work outside during the day, too tired in the afternoon. Zohar and I have been waking up at 5:30am everyday. Meta Quest is a great hardware product but damn, the OS and software sucks. So sick I don't want to eat. Thank g-d for Talia. She made a delicious chicken soup for breakfast. I've been watching Victoria in unfortunately small chunks (about 30 minutes at a time). It's amazing so far (I'm ¾ or so through), love a good one-shot. A friend of mine is credited as the Visual Effects Supervisor. Big recommend from me. More blogging, please Tobias. Do you have a writing chair? I find it hard to write at my desk. I pick up my laptop and move to one of the lounge chairs in my office. How do you grow a podcast audience without social media? Or a blog audience, for that matter. I was thinking about this, and maybe the answer is email. I don't mean email newsletters, I mean personalised email to other podcasters and blog authors. Start real conversations. I might take this post down later, I'm not quite thinking straight. No audio for this one sorry. Well, yes audio, but not conventionally.

We use the word "goal" to talk about so many things.

I want to get better at piano. I'm committed to doing yoga every morning for 3 months. Someday, I'd love to run a marathon. These are all "goals".

Business school teaches that goals should be measurable, within a set timeframe. Sales targets and KPIs and OKRs are all goals with multiple axes: time and value. We plot our course as we go, adjusting as necessary, until we reach our final destination (up and to the right). These types of goals are maps.

But back here in the real world, life teaches something different.

Maps are rigid ideas. We place a destination marker and chart the shortest route to those coordinates. Maps don't leave room for exploration or serendipity. Changing course is difficult, and you certainly can't stop half-way!

These types of goals may yield some short-term success, but ultimately they lead to banality, boredom, and burnout.

A better tool is a compass.

A compass points you in the right direction, but doesn't dictate any particular path. It opens a door to adventure and possibility. We may get lost, but in those moments we often stumble upon unexpected treasures - hidden pathways, chance encounters, and new perspectives.

While a map goal has a time and value component (the x and y axes), a compass goal includes the territory and the direction.

• I want to run (territory) more often (direction)
• I want to improve (direction) my health (territory)
• I want to learn more about (direction) cooking techniques (territory)

Compass goals remind us that the journey is more important than the destination.

Compass goals are still goals. It is possible to wander aimlessly with no direction at all, embracing a sense of wonder and delight in every experience. This can sound like:

• I love being in nature
• Spending time with my kids is important to me
• There's nothing better than a book and a warm bed

---

That way of thinking is not without worth, but it doesn't move a person forward, it doesn't help with goal setting. In a kabbalistic sense, this sort of self-wandering relates the sefirah of Chesed, which represents devotion, love, and desire.

Map goals, on the other hand, are like the sefirah of Gevurah, representing discipline, limitation, and judgement. Like Chesed, Gevurah is an important and helpful part of life, but alone, it has ultimately negative outcomes for goals.

Sitting between these two sefirot is Tiferet. Tiferet is all about integration, beauty, and balance. Our goals serve our needs best when they strike a harmony between the whimsy and desire of Chesed, and the discipline of Gevurah.

Compass goals, or we could call them Tiferet goals, retain the excitement and enjoyment of our task, while continuing to point us in the right direction.

There's something special about black and white photography. We cannot help but perceive colour all around us, so the black and white image doesn't capture the world as we see it, but rather, as we remember it.

It is said that we dream in black and white. I don't think that's true. Dreams, like memories, are a non-visual sensation. We might remember some colours that might have stood out, but that's a memory of the experience of colour, not colour itself.

As Andri Cauldwell puts it:

To see in color is a delight for the eye but to see in black and white is a delight for the soul.

I've loved shooting black and white film for a long time, especially on my medium format Mamiya 7II. Recently, I've been re-inspired by the work of Adrian Vila (I highly recommend his blog).

The birth of photography occurred in 1826 when French inventor Nicéphore Niépce created his groundbreaking View from the Window at Le Gras. The process involved coating a pewter plate with bitumen and exposing it to light for eight hours. The light exposure hardens bitumen, so in areas where less light reached the plate, the bitumen remained soft and could be washed away. After polishing the plate, Niépce was left with a negative image.

We've obviously come a long way since then, via the daguerreotype and the calotype, and eventually colour photography. Yet, despite the popularity of colour photos, it's still black and white photos that we remember best.

Here's some of my favourite photos that I've taken over the years.

---

---

---

---

---

---

---

---

---

---

---

The dust and scratches in the images below a result of me bring a lazy while processing the film.

---

---

I've always been a troublemaker. It's in my blood.

You see, through my father's line, I am a descendant of Cornish pirates.

You may have heard of the Pirates of Penzance. The Gilbert & Sullivan opera was written about real pirates who harboured in a town named Penzance, at the southern tip of the United Kingdom. If you drive 15 minutes north from there, along the A30, you'll come to a place called Carbis Bay.

That's my namesake. I'm a Carbis.

Carbis Bay, nestled in the larger St. Ives Bay, faces the Celtic Sea. From here you can easily sail to Cork or Cardiff, Plymouth or Porto. It's an excellent place for a seafarer to setup shop. And many did. Cornwall became known as a place of refuge for pirates.

An elaborate system of piracy was carried on intermittently during the whole of the [Elizabethan] reign… A few of their havens-one in Dorset, one in Cornwall, three or four in Ireland, and one or two in Wales-were immune from sudden attack. They were nearly always privately owned, some of them were never used for legitimate traffic, and they were all more or less secret.

No search was ever successful. In Cornwall and Wales it was impossible to muster such forces as the Crown controlled without arousing suspicion. No pirate captain was ever taken while unlading a cargo in secrecy, and it was above all the secure possession of these harbours that made the traffic successful.

The world needs troublemakers. Or, as Seth Godin puts it, people who "make things better by making better things". The status quo is like an old freezer. It can get jammed, and sometimes needs a good shove to break the ice, so that it can open smoothly again.

Steve Jobs penned it beautifully, in the famous 1997 "Think Different" ad.

Here's to the crazy ones. The misfits. The rebels. The troublemakers. The round pegs in the square holes. The ones who see things differently. They're not fond of rules. And they have no respect for the status quo. You can quote them, disagree with them, glorify or vilify them. About the only thing you can't do is ignore them. Because they change things. They push the human race forward. And while some may see them as the crazy ones, we see genius. Because the people who are crazy enough to think they can change the world, are the ones who do.

This really resonated with me as a teenager. It was foundational. Another foundational text was the Hacker's Manifesto (which in hind-sight is a little more angsty).

This is our world now... the world of the electron and the switch, the beauty of the baud. We make use of a service already existing without paying for what could be dirt-cheap if it wasn't run by profiteering gluttons, and you call us criminals. We explore... and you call us criminals. We seek after knowledge... and you call us criminals. We exist without skin color, without nationality, without religious bias... and you call us criminals. You build atomic bombs, you wage wars, you murder, cheat, and lie to us and try to make us believe it's for our own good, yet we're the criminals.

This is our world now... the world of the electron and the switch, the beauty of the baud. We make use of a service already existing without paying for what could be dirt-cheap if it wasn't run by profiteering gluttons, and you call us criminals. We explore... and you call us criminals. We seek after knowledge... and you call us criminals. We exist without skin color, without nationality, without religious bias... and you call us criminals. You build atomic bombs, you wage wars, you murder, cheat, and lie to us and try to make us believe it's for our own good, yet we're the criminals.

My pirate heritage has had plenty of opportunities to play out in my tech career. Standard software and media piracy, of course (information wants to be free, man!). The very spirit of open-source software is pirate-like. Taking fancy Spanish wine from Galleons and distributing it cheaply to the masses – is that really so different from what the internet did to IBM and Oracle?

Beyond that, breaking rules has defined my career. I'm known by friends and colleagues for ignoring expectations and experimenting with new ideas. I have little respect for ladder-based leadership (I don't tend to do well at large companies) and embrace meritocracy – those who do the work get to decide. That's been my history, but also the history of the tech sector. Tech has been taken over now by men in ties, but it was started by rebels and rogues with dirty jeans and dirtier hair.

In education, mischief is more important than ever. Our school systems reinforce conformity. It rewards students for the right answer, and punishes a better question. The world walks on tip-toes, hoping not to offend anyone. This stifles new ideas, hard questions, and growth.

It's time we taught our children to make a ruckus. Our kids need to learn how to pursue their own goals, not some curriculum designed to make everyone think the same way. Let's fly our skull-and-crossbones flag high, and raise wild, fearless, dangerous pirates.

I recently wrote about the way AI is making it possible to build software essentially for free. That's not exactly true – at least not yet.

Starting work on a new project is hard. It's the classic blank page problem. While AI doesn't yet do such a great job of programming, one thing that it's great at is giving me a starting point.

From there, it's much easier for me to say "that sucks, it should be done this way instead", and rewrite the whole thing.

I was chatting with a friend of mine, Aaron West, about this phenomenon just now, while he worked on fixing a whole bunch of poorly generated React components. He put it this way:

The review and edit workflow is much easier than building everything from scratch. Even if it means you end up rewriting the whole thing.

I've found this to also be true for writing. Occasionally I might ask ChatGPT to write something for me, only to find that the attempt inspires me in what not to do. I'll read it, abandon it, and feverishly begin my attempt to show the LLM how a real person does it.

Take that, ChatGPT.

I don't ever take that approach with this blog, since I only write what I'm inspired to write. I do occasionally prompt with "this paragraph doesn't quite work, can you fix it?" or "finish this thought…". Even then, I take the suggestions and run with something entirely different.

AI hasn't yet proven itself as a great writer or programmer, but it's an excellent tool for getting the creative ball rolling.

On June 28, 2005, Apple release iTunes 4.9. From the release notes:

With iTunes 4.9, you can now browse and subscribe to podcasts from within the iTunes Music Store. Podcasts are frequently updated radio-style shows downloadable over the Internet. You can also transfer podcasts to iPod, for listening on the go.

At launch, Apple's podcast directory was home to 3,000 shows. The BBC at the time described podcasts like this:

Podcasts are downloadable audio shows that can be created and listened to by anyone with the right equipment.

Who could have imagined that podcasts would grow to become an entire media category, right alongside film, television, music, and books?

In 2020 over 1,000,000 new podcasts launched! That number has died down quite a bit – only a quarter that many in 2024. Even so, there remains 260,000 active podcasts as of this writing.

Despite the popularity of podcasting, publishing a podcast still requires a good amount of time, skill, and money. There are some great tools to help, but there are also some gaps.

The way I see it, a podcast production workflow requires three keys steps¹.

1. Record

This step is well covered by existing tools. You could use music production software like Logic or Garageband, a simpler tool like Voice Memos or Quicktime, or a SaaS that also handles video calls like Riverside. Most of these tools have native Mac and iOS apps. Most are easy to use and well established.

2. Edit

When it comes time to edit, things get a little trickier. You could learn some of the more advanced features of Logic or Garageband, keeping the editing and recording software the same. Or you could try AI tools like Descript which automatically trims silences and remove filler words. Many podcasters swear by Ferrite, an iPadOS native app that makes podcast editing famously simple using the Apple Pencil. Again, there are many native apps, and while editing software can be more complex, there's lots of help to be found online.

3. Publish

This is where things take a bit of a turn. You've recorded and edited your podcast in Garageband, and managed to export it as an M4A file. Now what? Or maybe you recorded in Voice Memos on your iPad, and edited using Ferrite. Again, what to do now that you've exported that file?

There's really only one answer: Find a podcast hosting platform to take care of the rest for you. You're kicked out of these simple native experiences, and suddenly find yourself logging into a web page (each host with a different UI), with top-bar navigation and an upload button.

Good luck doing it on iPhone. And don't forget the monthly subscription.

Publishing is really quite a departure from the user experience of recording and editing your show. There's a real friction there, while you wait for web pages to load so you can manage the description and meta data for any one of your episodes. This is not what it should feel like to publish a podcast!

---

I'm working on an app to fix this. It's called Pita. It's based on 3 Principles for Podcast Publishing.

1. A Native Interface

Podcast publishing should feel native, like editing an Apple Note or using the Mail app. There should be workflow support to drop a new episode in through the Share Sheet, or a Shortcuts integration. Our podcasts are important to us. We want to see how they're faring with widgets and notifications.

2. Hosting Independence

A podcast publishing app should support any hosting provider. Why learn a whole new interface for each of the different hosts you engage with? Publishing a podcast should be like opening your favourite FTP client and choosing which server configuration to connect to. Or an email app that supports multiple accounts. All your podcasts, managed from a single app interface, no matter where they're hosted.

3. Support to Self-Host

You might not want to pay a $12 monthly subscription to host your podcast! Perhaps you have a cheap S3 bucket ready to go, or would be happy to host your podcast with a Creative Commons license for free through the Internet Archive. There should be an app which enables these self-hosted options, with an auto-generated RSS feed based on your episode data.

---

So that's what I'm building. It's still early days, but a work in progress nonetheless. Stay tuned, and if you like the idea, please let me know! Those feedback loops help keep me motivated.