Main Site

today, in Trakaido/Barsukas development: I want to add 20-30 cities to the word list 💡 with translations in 14 languages.

• Claude launches a second agent to do the work
• The second agent generates a JSON file
• Claude then does "Web Search" to try to find the ones it isn't sure about
• After I cut it off, Claude uses the Wikidata API
• Claude does ... a lot of verifying, much of it needless

🔥 the layers of abstraction are dizzying ...

I have seen this specific error before, and I think the Claude Code team should be able to fix it.

🔥 do you see the bug? ⚙️ it is going up another directory. "cd .." is not idempotent. 💡 maybe the local Claude directions should make it clear?

Somehow Barsukas has become the other server as part of the Trakaido project. It contains translation data, sentence data, etc.

It is, in effect, a multilingual dictionary.

Instead of sorting alphabetically, we have to categorize words and then sort them roughly by linguistic complexity. This is a blurry term, but:

• cow is less complex than marmoset
• red is less complex than crimson
• table is less complex than credenza

The categories themselves have evolved. There is a super category that the LLM generated in the UI that I am not concerned with, but we have:

• around 40 categories of nouns
• 15 categories of verbs
• 10 for adjectives and adverbs
• a separate numeral category for number words

We have an architecture designed around a development pattern where people make APIs including the OpenAI key to add words. As a single-user project, this is a way to avoid putting the LLM key on the server. For a multi-user project, there are obviously security risks associated with this design. We are hoping to solve it better later.

Some of the agents have turned out to be more useful than others. This is fine.

• The low-hanging fruit is having agents do things like type checking. 💡 there is no reason to not have your main codebase include type hints for JavaScript, Python, etc. any more. On the other hand, there is no need to force developers to use them. An agent can modify the pull request to add them.
• Having an agent take every "Swift" ⚙️ for an iOS app app change, and apply it to "Kotlin" ⚙️ for an Android app, doesn't work extremely well. And, as I have mentioned, the $20/month LLM subscriptions will work for an individual developer doing one version, but not for two 💡 or three with web; or four if you count documentation versions of an app.
• Audio generation is still expensive if you have an LLM API do it, and low-quality locally. Neither the LLM nor non-LLM audio generation tools give high-quality output. And, even "gpt-4o-mini" 💡 still the newest model in January 2026 has about a 1% error rate.

the Trakaido stats server is in the Atacama repo largely by accident.

in addition to "word" stats, we have grammar interstitial views and user config data on Atacama as well. ⚔️ well, actually, it's a different process, though the same monorepo

There is much talk about whether LLMs prefer a monorepo or microservices.

The answer is, generally, it doesn't matter. You can run an LLM across multiple Git repositories.

Having the iOS and web 🔥 and Android apps in the same repo is a slight convenience for a one-person development team. But would probably become large and cumbersome later on.

https://github.com/DigitalPhonetics/IMS-Toucan

Another text-to-speech platform. Supposedly supports thousands of languages.

But: the audio quality is too metallic, the diction is too blurry.

Claude seems to be standardizing on a new "agent session" method, and I think I like it.

Step 1: Design a plan.

Step 2: Write the plan.

Step 3: Forget everything other than the plan.

Step 4: In parallel, execute the plan.

My list of "firmament words" ⚙️ words that are too difficult to define without already understanding language at an advanced level, or that are so fundamental they are often embedded in grammar rather than stand-alone words from a few years ago:

Click title to read full message...

I met a man who carried a Bible everywhere he went, he said that he walked with Jesus.

He did not call himself a preacher, but he talked mostly in Bible verses. I took no offense at that, but my romantic partner was somewhat agitated by it.

I must agree with her on one count: he did use a lot of words to say very little. He seemed to have taken inspiration from a certain type of low church ⚙️ "High Church" is the bells-and-smells formal liturgy of an ancient and hierarchical church. "Low Church" is just a man with a bible and a voice. preacher, who can spend an hour talking about six lines of the Bible 💭 the Catholics have 3 readings in their Sunday services. the Jews, never to be outdone, do seven ... in the original Hebrew. Many of the Protestant denominations can only find time for 1 reading, and not say very much interesting in that time.

He was of a persuasion that suffering leads one closer to the LORD, so one should strive to live a life full of suffering. This may not be as wrong as some readers may think it is, but I find it the conclusion of flawed logic.

It may be true that one must suffer in order to know the LORD.

It may be true that the LORD's presence is more easily reached by those in suffering.

But one must not treat it like opiates; living for nothing but the joy of suffering.

And one must not reach the point where one loves the suffering more than the LORD.

There are those who are called to a life of asceticism. But it is best if they know it comes at a cost; to themselves and those around them.

One might think that it is good that they live a life away from others; a hermit in the woods, or at least a celibate monk in a mostly-silent commune.

✨ of those things that cannot be spoken-of: do not speak of them.