Ian Bicking
ianbicking.hachyderm.io.ap.brid.gy
Ian Bicking
@ianbicking.hachyderm.io.ap.brid.gy
Software developer from Minneapolis, long lapsed open source author, working at Brilliant.org - previously Mozilla and Meta.
searchable

[bridged from https://hachyderm.io/@ianbicking on the fediverse by https://fed.brid.gy/ ]
I rode passenger today on an ICE patrol in my neighborhood in Minneapolis.

Our ride was mostly uneventful, the neighborhood we patrolled has been a target but just not this afternoon.
It took me a while to follow everything that is going on, and I felt rather incompetent even as a passenger […]
Original post on hachyderm.io
hachyderm.io
January 13, 2026 at 11:10 PM
RE: https://hachyderm.io/@ianbicking/113235462740592053

"Kristallnacht as housing policy" was not taking it too far
hachyderm.io
January 12, 2026 at 7:53 PM
ICE activity and aggressiveness increases each day in Minneapolis, and expands further into the state with reports of large actions in Rochester. Many observers being arrested, citizens being abducted off the street, entering and searching hospitals.

No activity at schools as they are all […]
Original post on hachyderm.io
hachyderm.io
January 9, 2026 at 3:58 PM
Oh my, I lost a lot of followers there
January 8, 2026 at 8:10 PM
On January 6th ICE began a full scale attack on Minneapolis. On January 7th they murdered a woman, Renee Nicole Good, they abducted people off the street, broke into homes without warrants, terrorized and attacked a high school. Today they got up and are doing it all again.

The murder happened […]
Original post on hachyderm.io
hachyderm.io
January 8, 2026 at 7:55 PM
I wrote a blog post: On The Subject Of AI Poetry
When language shows up on cue what meaning is there in it?

https://ianbicking.org/blog/2025/10/ai-poetry
On The Subject Of AI Poetry
ianbicking.org
October 1, 2025 at 7:44 PM
I've been thinking about #llm sycophancy (the tendency of #chatgpt/etc complimenting the user excessively), and wondering if it has more purpose than simple aggrandizement of the user. That is: is it serving other valid conversational purposes? First an example:

Another LLM tendency is to be […]
Original post on hachyderm.io
hachyderm.io
September 29, 2025 at 8:49 PM
An algorithmic puzzle I thought of:

Imagine you have 5 DIP switches. You want to be able to represent all 32 values, but instead of 1's and 0's you will label them with letters. So you might imagine:
A<->B
C<->D
E<->F
G<->H
I<->J

One of the values might be called ADEGI, or another BDFGI, and […]
Original post on hachyderm.io
hachyderm.io
August 30, 2025 at 4:22 AM
I wrote up a review and a bunch of constructive suggestions for the LLM-driven game Hidden Door:
https://ianbicking.org/blog/2025/08/hidden-door-design-review-llm-driven-game

In large part the summary is "they need to do way more prompt engineering"
Hidden Door At Launch: Design Review of an LLM-Driven Story Game
ianbicking.org
August 27, 2025 at 6:18 PM
I'm watching Technical Connections (a technology explainer YouTube channel) and I'm enjoying it. I think a big part is that the guy is willing to making strong claims like "this is ALL you need to understand" or "this is how EVERY device of this type works". It's comforting to feel like you can […]
Original post on hachyderm.io
hachyderm.io
August 26, 2025 at 8:09 PM
So with the Pocket service down, what should I use now as a cross-device bookmark storage service? Ideally where I can upload my Pocket export. Mostly I save things to assuage my guilt at not reading them, but sometimes I really do revisit things. But my needs are simple.
July 24, 2025 at 5:18 PM
Is a recursive async function in Javascript effectively doing tail call elimination?

Hrm... I guess only if you return the recursive call and don't await on it. I think?
July 21, 2025 at 2:42 AM
SudoLang (https://medium.com/javascript-scene/anatomy-of-a-sudolang-program-prompt-engineering-by-example-f7a7b65263bc) was one of several attempts during the GPT-4 era to create formalized prompting languages, where there's no execution engine, just something slightly formalized for the LLM to […]
Original post on hachyderm.io
hachyderm.io
July 19, 2025 at 9:42 PM
Here's an article about a 60s UK educational experiment to teach reading with a modified (purely phonetic) alphabet called ITA: https://www.theguardian.com/education/2025/jul/06/1960s-schools-experiment-created-new-alphabet-thousands-children-unable-to-spell

As large and radical as the program […]
Original post on hachyderm.io
hachyderm.io
July 17, 2025 at 6:54 PM
I wrote a (long) blog post on a side project called Intra, a text adventure game (interactive fiction) with both real mechanics and run by an LLM: https://ianbicking.org/blog/2025/07/intra-llm-text-adventure

There's lots of LLM collaborative storytelling games, but nothing I've encountered has […]
Original post on hachyderm.io
hachyderm.io
July 2, 2025 at 4:57 PM
[#uspol protests]

I went to the No Kings protest today; it was good counter-programming and the least I could do at this moment. But attending protests mostly bums me out, they always feel like a demonstration of impotence.

In part it's because protest culture is incredibly stale. Walking […]
Original post on hachyderm.io
hachyderm.io
June 15, 2025 at 12:13 AM
Late to the game here, but having just taught some Python and JavaScript to a kid, I think whitespace sensitivity is not a good choice for learning. Braces and automatic formatting is much better.

An important part of learning is to see code and just copy it (I agree with Zed Shaw's advice from […]
Original post on hachyderm.io
hachyderm.io
June 1, 2025 at 9:25 PM
There's a trope among educational (especially mobile) apps to encourage a "learning habit" or some regular cadence of learning, like 10 minutes a day or something.

This is promoted as being the "best way to learn" or something. But is it? At all? None of my best learning was like that. At all.
June 1, 2025 at 12:33 AM
I got into my mind that I should think about better typed interfaces between TypeScript and my PostgreSQL database.

Two libraries seem like good fits:
https://github.com/adelsz/pgtyped
https://github.com/pg-nano/pg-nano

Both look pretty cool, and slightly different but with the same approach […]
Original post on hachyderm.io
hachyderm.io
May 19, 2025 at 11:54 PM
Reposted by Ian Bicking
Alright, I'm back with what I assume will be a week long rage thread about a book I hate. It's really not important for current events in any way, but I signed up for a talk about it in the before-times.

So mute this, it's absolutely not interesting […]

[Original post on social.vivaldi.net]
May 12, 2025 at 12:43 PM
I really dislike table aliases in SQL, but data people seem to love it. (And as a result AI loves table aliases because it learns from them.)

That is:

SELECT *
FROM user_table u
WHERE u.name = $1

Or even worse I've seen some people do this consistently:

SELECT first_name, last_name
FROM […]
Original post on hachyderm.io
hachyderm.io
May 12, 2025 at 10:22 PM
I was looking up an old blog post of mine, and now that I've read it I think it's still a good blog post:

https://ianbicking.org/blog/2019/01/overengaged-knowledge-worker

The unrealistically idealized model of the user that I complain about there remains more powerful than ever.
The Over-engaged Knowledge Worker
ianbicking.org
May 12, 2025 at 2:29 AM
I posted a video walkthrough of a side project I played with last year: an LLM-driven text adventure, Intra

https://youtu.be/CWReNVF3Ark

The game is not complete by any means, but I haven't seen anything else that goes as far with using an LLM to drive a game while also using a grounded game […]
Original post on hachyderm.io
hachyderm.io
May 6, 2025 at 10:33 PM
I have a particular programming task that I keep starting but gets interrupted by life. And each time I think "well this is kind of a big thing, how am I going to change the system to make this work?" Then I remember it surprisingly already fits into the system. Then I get interrupted and next […]
Original post on hachyderm.io
hachyderm.io
May 2, 2025 at 12:45 AM