🔭 Search Space is out of beta! 🥂 New in version 1.0.0:

• 🎙️ Transcripts can be indexed!
• 🪚 Post content in search results is truncated at 300 characters.
• 🗂️ Indexing replies and transcripts are optional.
• 👋 The greeting text (Welcome to Search Space …) can be edited in the settings.

One wikipedian had fun with today’s Did you know … section. 🙃 Love it! Via Molly White.

Remember! There’s no such thing as having too many book-related newsletter subscriptions. If you’re a fan of sci-fi and/or horror, subscribe to Frights, Flights, and Literary Delights. Curated by @sannalund and yours truly.

Or read our first issue on the web. 📚

– I do a nude scene. Do you want to see it?
– Huh? Did I say yes?
– You didn’t say anything.
– I meant yes, my mouth didn’t speak.

🤩 From the trailer for Wes Anderson’s Asteroid City.

I’ve just rolled out a new version of my ✍️ Reply by Email plug-in. Now with email obfuscation to combat email harvesting & keep your inbox safer! 🛡️

Remember, no solution is 100% foolproof against web scraping spam bots. For best protection, combine with a service like Apple’s Hide My Email!

📚 Great advice from Oliver Burkeman:

[Treat] your “to read” pile like a river (a stream that flows past you, and from which you pluck a few choice items, here and there) instead of a bucket (which demands that you empty it). After all, you presumably don’t feel overwhelmed by all the unread books in the British Library – and not because there aren’t an overwhelming number of them, but because it never occurred to you that it might be your job to get through them all.

For those of you using my 🦄 Magic Preview plug-in in Safari, force clicking the 🔍 Preview link gives you a quick preview of what your post will look like. Works on handheld devices as well, just tap and hold the link to see a preview without opening the page. Like magic! ✨

Instead of manually adjusting hundreds of Markdown formatted microblog posts when migrating my blog from Jekyll to Hugo (Micro.blog), I asked ChatGPT to do it for me. How did it go?

👀 Read How ChatGPT made my migration from Jekyll to Hugo a breeze to find out.

The other day, after running my blog on Jekyll for 13 years, I switched to Hugo. Moving from one static site generator to another involves a lot of things to take care of. Among them was this directory with hundreds of microblog posts written in Markdown. The posts published dates encoded as Unix timestamps in the filenames: 1665322487.md etc.

Micro.blog hosts my new Hugo blog, and their import tools do not support Unix timestamps in file names. Instead, publish dates should be put in the YAML front matter, as follows:

---
date: 2022-10-09 13:34
---
Had a great day in the forest with friends. Grilling sausages, roasting marshmallows, and taking a short hike. Hair smells like a campfire now. 🥰

So, I basically had two options. I could manually fix this by snatching the timestamp from the filename, converting it to the right date-time format, and pasting that into the file’s YAML front matter. Rinse and repeat. Hundreds of times.

Or I could just write a script that automated the process for me. As a programmer, this is what I usually do. Even when it won’t actually save me any time in the end. 😅

But, instead, I tried a third option and asked ChatGPT to do it for me. I’ve played a lot with the chatbot and had it generate code for me in the past. But just toy examples, not, you know, anything useful in the real world. This task, though, seemed like a perfect fit. Easy to define and limited in scope. So, without further ado, here’s me and ChatGPT programming away. Complete with lazy typing and spelling mistakes (mostly mine).

My pair programming session with ChatGPT (GPT-4)

Sven: Write a Ruby script that loops thorough all files in the directory provided and adds the filename to yaml frontmatter without extension.

For example the file named 1676537577.md should end up like this in the frontmatter:

-–
date: 1676537577
-–

ChatGPT: Here is a Ruby script that loops through all files in a specified directory and adds the filename without the extension as the date field in the YAML frontmatter.

ruby

require 'yaml'
require 'fileutils'

def process_file(file)
  frontmatter = { 'date' => File.basename(file, '.*') }
  content = File.read(file)
  content.sub!(/---\s*\n.*?---\s*\n/m, '') # Remove existing frontmatter if present
  File.open(file, 'w') do |f|
    f.write(YAML.dump(frontmatter) + "---\n" + content)
  end
end

dir = ARGV[0] || '.'

raise "Directory #{dir} not found" unless Dir.exist?(dir)

Dir.glob(File.join(dir, '*.md')) do |file|
  process_file(file)
  puts "Processed: #{file}"
end

puts 'All files have been processed.'

Save this script as add_date_to_frontmatter.rb and run it with the desired directory path as an argument:

bash

ruby add_date_to_frontmatter.rb /path/to/directory

If no argument is provided, the script will default to the current working directory. This script assumes that your files have the .md extension.

Sven: Great! Now make is so that the file name (without extension) is processed as a unix timestamp and converted to a iso-8601 formatted date and time instead.

ChatGPT: To modify the script to process the filename as a UNIX timestamp and convert it to an ISO-8601 formatted date and time, you can use Ruby’s built-in Time class.

Here’s the updated script:

ruby

require 'yaml'
require 'fileutils'

def process_file(file)
  filename_without_ext = File.basename(file, '.*')
  timestamp = filename_without_ext.to_i
  iso_date = Time.at(timestamp).utc.iso8601

  frontmatter = { 'date' => iso_date }
  content = File.read(file)
  content.sub!(/---\s*\n.*?---\s*\n/m, '') # Remove existing frontmatter if present
  File.open(file, 'w') do |f|
    f.write(YAML.dump(frontmatter) + "---\n" + content)
  end
end

dir = ARGV[0] || '.'

raise "Directory #{dir} not found" unless Dir.exist?(dir)

Dir.glob(File.join(dir, '*.md')) do |file|
  process_file(file)
  puts "Processed: #{file}"
end

puts 'All files have been processed.'

Save this script as add_iso_date_to_frontmatter.rb and run it with the desired directory path as an argument:

bash

ruby add_iso_date_to_frontmatter.rb /path/to/directory

If no argument is provided, the script will default to the current working directory. This script assumes that your files have the .md extension.

Sven: Okay, I changed my mind, instead of ISO-8601 make it on the format yyyy-mm-dd hh:mm. For example 2020-03-28 22:10.

ChatGPT: Sure, you can modify the script to format the date and time as “yyyy-mm-dd hh:mm” by using Ruby’s strftime method. Here’s the updated script:

ruby

require 'yaml'
require 'fileutils'

def process_file(file)
  filename_without_ext = File.basename(file, '.*')
  timestamp = filename_without_ext.to_i
  formatted_date = Time.at(timestamp).utc.strftime('%Y-%m-%d %H:%M')

  frontmatter = { 'date' => formatted_date }
  content = File.read(file)
  content.sub!(/---\s*\n.*?---\s*\n/m, '') # Remove existing frontmatter if present
  File.open(file, 'w') do |f|
    f.write(YAML.dump(frontmatter) + "---\n" + content)
  end
end

dir = ARGV[0] || '.'

raise "Directory #{dir} not found" unless Dir.exist?(dir)

Dir.glob(File.join(dir, '*.md')) do |file|
  process_file(file)
  puts "Processed: #{file}"
end

puts 'All files have been processed.'

Save this script as add_formatted_date_to_frontmatter.rb and run it with the desired directory path as an argument:

bash

ruby add_formatted_date_to_frontmatter.rb /path/to/directory

If no argument is provided, the script will default to the current working directory. This scrit assumes that your files have the .md extension.

Closing thoughts

The final output from ChatGPT just worked and successfully completed the task at hand. That’s pretty cool! Is the implementation perfect? No, but it gets the job done. For a one-off script, I can’t complain.

The resulting program is short, and I could’ve come up with a solution and typed it out myself. But it’s fascinating to program in plain English. And I doubt I would have come up with the solution as fast as ChatGPT. I didn’t time our session, but estimate it took ChatGPT about a minute to generate each answer. So, from my first prompt to a job well done, it took no more than five minutes.

By the way, the first draft of the program ran fine as well. In the second draft, ChatGPT didn’t require 'time' at the top, but after human intervention, that version also ran correctly.

So, March 22, 2023 was the first time I paired programmed with a large language model to successfully solve a real world problem. And I’m certain we’ll team up again in the future to solve even bigger and more complicated problems.

Okay, everybody, today is a big day because my bff/gf/wife @sannalund just joined Micro.blog! She’s way cooler, smarter, funnier, and definitely prettier than me, so give her a follow and a warm welcome. 🥳

We’re doing wine, pizza, and newslettering again. If you’re at all interested in books – especially in the horror and sci-fi genres – you should stop what you’re doing right now and sign up for Frights, Flights, and Literary Delights. The first issue will arrive tomorrow.

A gpt-3.5-turbo powered poet-clock. Cute idea! Matt goes into some details in his post My new job is AI sommelier and I detect the bouquet of progress.

I made an AI clock for my bookshelves! It composes a new poem every minute using ChatGPT and mysteriously has this enthusiastic vibe which I am totally into

Here’s 3 × Game Boy (again):

• A Game Boy emulator device that harvests energy from the sun and your button mashing. More scientific papers like this, please. 🥺
• Travis Goodspeed’s Game Boy ROM Tutorial reminds me of my piece, How to Take Screenshots the Hard Way.
• Just look at 2bitboy’s custom-made Camera+ below. What a piece of art! 👌

Gameboy Camera, a lifestyle choice . . . . . . . . . 😂😂😂

Bill Gates on The Age of AI has begun:

First, we should try to balance fears about the downsides of AI—which are understandable and valid—with its ability to improve people’s lives. …

Second, market forces won’t naturally produce AI products and services that help the poorest. The opposite is more likely. …

Finally, we should keep in mind that we’re only at the beginning of what AI can accomplish. Whatever limitations it has today will be gone before we know it. …

The Age of AI is filled with opportunities and responsibilities.

Despite some slippery spots, we had a lovely Sunday stroll at Dumme mosse.

This post is just a test to see if we can abuse the newly introduced 600 characters limit for quotes, without actually, you know, quoting anything. Maybe we can! As I write this, I don’t know, and I have to keep typing to reach the 600-character limit. If you, dear reader, can read this post in its full glory on the Micro.blog timeline, it means I’ve succeeded. Let’s celebrate that:

🎉🥳

You might wonder how this tiny trick is achieved? I’m glad you asked! There’s this saying that a screenshot is worth more than a thousand words (or is it 600 characters?), so without further ado, here it comes:

My new Micro.blog plug-in is a minimal, no-frills post editor. But it has one pretty magical feature: live previews!

See it in action below, try it on my blog, or look up 🦄 Magic Preview in the plug-in directory. ⚠️ Read the important notice there before installing. Have fun!

I will not harm you unless you harm me first

Here are some hilarious (AI-powered) Bing encounters compiled by Simon Wilson.

I teamed up with OpenAl’s Whisper and had a lot of fun hacking on interactive transcripts this weekend. The neural net handled the transcription, so I, the human, could focus on programming and implementing the user interface.

See the demo below or live (in Swedish).

It’s #FeedReaderFriday again! And you should subscribe to Julia Evans' blog for nerdy posts, like Why does 0.1 + 0.2 = 0.30000000000000004?, and beautiful, informative zines, like Hell yes! CSS!.

Today I completed another task on the study to-do list: installing the ceiling light. Thirteen individual glass globes had to be attached one by one – tedious, but worthwhile! 💡

Summarize any content on the web - from articles to books - in a jiffy!

Universal Summarizer works on videos as well. I fed it my talk from Micro Camp 2022 and out came a pretty spot-on recap.

It’s #FeedReaderFriday already! Firstly, if you haven’t explored Maggie’s digital garden yet, now is the time. Yes, right now. 😊 And then, add the feed to your reader to never miss one of her visual essays (about anthropology, design, and programming). You won’t regret it!

Om du gillar böcker och befinner dig i närheten av Jönköping ska du så klart besöka SmåLit i morgon. Föreläsningar, författarsamtal, bokmässa och massa annat kul för bokmalar och läslöss. 📚

Där, vid ett av åttio bord kommer du hitta Sanna (och undertecknad) krängandes Sveriges kanske just nu bästa skräcknoveller. 👻 Vi ses!

Books and potted plants are slowly but surely finding their way back into the study. Along with one of @veronique’s original drawings. 🥰 Check out her lovely creations over at veronique.ink.