It is amusing to me that ChatGPT is running on Microsoft’s Azure infrastructure and Microsoft Teams has already integrated it into the product. Every company I know that uses Windows, Teams, and Outlook are the absolute last ones I’d expect to embrace AI in their workflows.

The week of 50F (10C) weather has the crocuses popping up! They are in for a rough time when the temperature hits 6F (-14.4C) tomorrow.

When ChatGPT is over capacity, it should at least let you log in see your previous chats while preventing new ones.

Mark D and I love jq for quick json processing via the command line.

curl -sH 'Accept: application/json' -H 'Authorization: Bearer [REDACTED]' https://public-api.wordpress.com/rest/v1.1/me/sites | jq -r '.sites[] | .URL' > sites.txt

Katz’s Valentine’s Day copywriting is on-point:

“Impress your beau with two pounds of hand-carved pastrami (the most sensual of all deli meats) with all the fixings, plus our rich cinnamon rugelach and an egg creme scented candle to set the mood.”

They hung it! Pretty cool. Now I need an excuse to go to Barcelona and visit.

The Museu de Matemàtiques de Catalunya (www.mmaca.cat) in Barcelona, Spain, is adding a photo I took of a cycloid to their interactive cycloid exhibit room!

Picked up a book on good hikes in the Hudson Valley for kids and am excited to take Charlie on some of these! I know we are a little ways off, but he walked roughly a mile on Saturday and did great.

Note for my future self since I always need to look this up: Here are the formats for tagging users in Slack when you post messages webhooks or the API:

• User: <@userid>, such as <@UUBE7MDML>. You can find this user ID in the web interface of Slack and going to their profile, then looking at the URL, which is in the format https://app.slack.com/client/clientID/something/rimeto_profile/userID
• Group: <!subteam^subteamID|@group-name>, such as <!subteam^S0406RGH26B|@chucks-fake-team> – You can find that in the web app, too: https://app.slack.com/client/clientID/browse-user-groups/user_groups/subteamID

On AI in blogging: I think I’m more interested in having AI handle things like categorization, tagging, and writing excerpts for me rather than writing the posts themselves. Blogging is personal and it should be from the author.

Imagine how cool it would be for an AI to consume your content and come up with a categorization and tagging scheme you hadn’t considered.

Meyer’s Olde Dutch in Beacon, NY, is excellent. Roadside stand style burgers and fries with really good sauces. Great selection of craft beer and ciders, as well as a craft cocktail menu that punches above this place’s size. Highly recommended.

Twitter removed the card preview from the validator tool (https://cards-dev.twitter.com/validator). That is annoying because half of the reason to use the validator tool is that it fetched fresh content from the site, which was a way to force updates on cards that needed updating. I suspect that the preview in Tweet Composer is working off of cached data and there is now no way to force an update. The announcement: https://twittercommunity.com/t/card-validator-preview-removal/175006

testing twitter syndication via Bridgy

It was a beautiful 50F degree day today, so Jon and I decided to go for a row on the Croton River.

We went out at high tide and rowed roughly three miles round trip. The water was quite swift today and we had some difficulty navigating a few bends in the river where the current picked up. Quite a workout!

I did some design updates on my blog this weekend during Charlie’s nap.

• Made new archive templates for the Likes and Microblog sections
• Made new single post templates for Likes and Microblog posts
• Updated the navigation across the site to finally include Likes and Microblog
• Added Likes, Microblog, and regular navigation to the homepage.

There is finally a birria truck in Peekskill! Paradise Taqueria Birrieria, parked on Brown St. Saturdays and Sundays. The standard quesabirria + consume is delicious, as are their salsas. Amanda likes the green salsa and I like the smokey red salsa.

Good post, Tom. I agree, we need more user-friendly tools to power use cases like this.

I love when people publish open CSVs on GitHub and have an associated GitHub Pages site where that data is displayed, but that takes some technical know-how. Heck, publish a SQLite file and I’ll be happy, but the moment the database isn’t viewable on the web, you’ve lost 98% of people.

Kaggle’s open datasets and notebooks are a partial solution, but more focused on data science. I’ve discovered some cool stuff in there, though. ObservableHQ comes to mind, too. You already mentioned Datasette, which I love, and it is also worth mentioning tools like Flatfile and OpenRefine

I’m not optimistic about storing niche databases like the ones you describe in Airtable or Notion because I don’t expect them to be around long-term or keep a consistent link structure with redirects while they are around. Are.na is nice for image collection, but not super sortable/filterable/searchable.

Regarding making a database of Brooklyn Artists and keeping it up to date, if you can find 1-2 profiles they all have in common (LinkedIn, GitHub, Instagram, Dribbble, Twitter), building a tool to occasionally check for updates would be a lot easier than free-form web scraping and parsing.

For a general use case: Since WordPress is my tool of choice (I work at Automattic on the WordPress Special Projects Team), I’ll explain how I’d go about setting up a niche collection of resources on WordPress.

• I’d start with setting up a custom post type with the appropriate fields for your collection.
  • There are two ways of doing things as the WP world transitions to Gutenberg. I think the Block way (where everything is added via blocks) is the best long-term, but there really aren’t tools for setting up custom collection blocks that do this sort of thing, so right now I’d still advise people to go the Custom Post Type UI + Advanced Custom Fields (ACF) or Pods route to set up the structure without code. Leverage things like Tags and Categories as much as possible to use the built-in functionality WordPress already has for these concepts.
  • Everything gets its own permalink.
• For outputting the content, you can use core blocks like the Query Loop for much of it, or leverage some of the blocks that ACF or Pods provides.
• Need auto tagging and more powerful term management? Use TaxoPress.
• For searching and filtering, use something like ACF Better Search and Search & Filter. Jetpack Search is really great, though a bit pricey for hobby projects.

I’m glad Buzzfeed is using ChatGPT now. Search engines and social platforms are going to demote it, as they should have long ago.

What is novel to you is just another Thursday for someone else.

• Having a baby is one of the biggest days of your life, but doctors and nurses in the maternity ward deliver multiple babies a day.
• Your huge data migration is small to Stripe, who handles bigger clients than you multiple times a day.
• Your website redesign project might be the biggest project you are working on right now, but the development team has 113 rebuilds in-flight right now.
• The garage tries to churn through as many tire changes as they can in a day.

All a matter of perspective.

Ivory for Mastodon by Tapbots is what I’ve been waiting for. I was a long-time Tweetbot user and it was my favorite third party Twitter app. Even the early version of Ivory is better than the other Mastodon apps out there.