My next project with n8n automation is to build a replacement for Mailbrew. 🫣 I’m facing many architectural decisions:

• How do I fetch content (web or RSS feeds)?
• How do I extract articles for more efficient summarization?
• How do I combine the results?
• How do I control the size of the summary?
• Do I need some form of temporary data persistence within the workflow?
• How do I minimize LLM credits usage?
• Should I use an n8n data table for storing data sources and loop through them one by one?
• Where do I send the summary and how (Telegram, Discord, Email)?

…and probably a few more decisions to ponder!

From MacUpdate website:

Unfortunately MacUpdater 3’s promised lifetime of “until 2026-01-01” is now over. There will be no MacUpdater 4 or any continuation of the MacUpdater product from us. Our daily maintenance has been stopped and we don’t verify updates anymore. MacUpdater 3.5 is now unsupported but free-to-use including all previous “Pro” features.

WTF?? 😩

This is my first post of the year on Micro.blog. Despite the overall global, political, and economic challenges that don’t seem very promising, I am personally looking forward to 2026.

Travel-wise, I have four planned: Egypt, Mexico, France & Thailand. This could also be a productive year for photography. I’m looking forward to those trips as we celebrate our 20-year relationship, my wife and I. 😊

Tech-wise, it’s the year of the iPhone upgrade. After skipping the iPhone 16 and iPhone 17, it’s time to upgrade my capable but aging iPhone 15 Pro Max. I expect to stay on the Max. It could also be the year when I replace my M2 MacBook Air with the M5 version. It will depend on the available money and other factors. Lastly, if Apple finally release a HomePod with a screen, I might get one, too. 💻

I wish you a good one!

I’m already turning my attention to 2026, in no particular order:

• Ghost.org next moves, now that are a better Fediverse citizen.
• Plausible Analytics, which seems to be overkill for my needs but I might find a use for their service exposure via their APIs.
• Craft because they finished the year with a bang and I’m super anxious to learn what’s next.
• Apple because of Apple Intelligence and Siri promised updates. Will they deliver? They are the underdog right now but it could serve them well in the end. Remember the Mp3 players market before the iPod? I do. I don’t want another features rush but a bug-fix bonanza. Will they deliver? I have my doubts. What’s next for Photomator and Pixelmator?
• Micro.blog might also surprise me with features like RSS reader integration. Who knows.
• Anthropic and OpenAI offerings… will they keep the pace? Will they slow down? Will they surprise us? Will they crash? Is enshittification on our way?
• Inoreader is also evolving at a steady pace and I wonder if it will continue in 2026.
• Things to-do manager: will it turn to version 4?

So much fun is awaiting, I’m sure.

In 2025, two notable apps or services received numerous and meaningful updates: Craft and Inoreader.

Craft received long-awaited tag support, with APIs and MCP support added. The latter two are quite transformative, and I expect 2026 to bring many new users to the app. Personally, I’m barely scratching the surface of Craft APIs. Craft is at the center of everything that I create, and I couldn’t think of a better app to support my creative journey.

Inoreader also got a bunch of updates, many of them focused on AI-based empowerment (article summaries, podcast and video transcriptions, tag suggestions, etc.). More than ever, Inoreader is an essential part of my digital toolset.

It’s a web app (on Vercel) just for me to use so that I can write blog posts on the go with a clean UI. It’s my second web app on Vercel built entirely with Claude AI.

Apparently, web analytics is not an exact science. Here are three web analytics versions of the same period: from November 23rd to December 23rd (Top: Ghost Analytics, Middle: Plausible Analytics, and Bottom: Tinylytics). Plausible feels conservative, with about half as many unique visitors as Ghost, while Tinylytics seems to overestimate. The patterns are barely the same, too. Who’s right?

While experimenting with n8n and LLM services, I realize that using artificial intelligence can become a very costly hobby. The fact that the consumption of these services relies on two separate offerings — the subscription to the interactive service and on-demand billing for APIs — requires careful management and wise choices of providers.

Currently, I use ChatGPT and Claude AI in interactive mode, but I also need a provider to access AI via APIs. This latter mode of consumption is particularly expensive if you’re not careful.

Still exploring n8n, slowly but systematically. Because I’ll be using all sorts of external services like Craft, Micro.blog, Ghost, Inoreader, my strategy is to do individual integration tests instead of trying to build a biggy workflow and find all sorts of errors. Each of these micro experiments is forming the building blocks of something bigger. This approach is not different from the one I used for building iPhone apps and learning Objective-C and Interface Builder back in the day. I guess I’m a bottom-up kind of guy when it comes to understanding this kind of stuff.

It was a memorable, long journey. I officially shut down my IFTTT account tonight after over five years of use. It wasn’t costly, but I have the feeling that IFTTT started to trail behind competitive offerings like Zapier, Make and now n8n. It wasn’t the most user-friendly for debugging issues. Now, I’m turning my focus towards n8n. In the coming days and weeks, it will be my next digital playground for experimentation.