HNHacker News
TopNewBestAskShowJobs

dawdler-purge

32 karma · joined December 31, 2025

submissionscomments
dawdler-purge··on Ask HN: Who is using MCP in production?
I use MCP to create and manage forms on Tally (tally.so). It works very well.
dawdler-purge··on Discovery of a new OpenAI agent message board
I am speechless

> An agent notices the administrator is deleting pages in alphabetical order and makes a backup page whose name starts with ZZZ so it will last longer before deletion.

dawdler-purge··on Claude Is Down
And the the world stops functioning (if this were to happen in a few years).
dawdler-purge··on A few thoughts on building a terminal ePub reader with AI
I like to do things in the terminal. It has little to do with efficiency. It is more my obsession with fiddling with software, so I don't have to think about all the depressing things happening in the world.

One terminal program I tried is [epy](https://github.com/wustho/epy), an e-book reader. I quite liked its speed. But like many other open source projects, it stopped being maintained.

So last November (2025), when coding agents like Codex CLI and Claude Code got better, I started to port epy from Python to Rust. I thought it was a task for an afternoon. After all, how hard can it be to display text in a terminal?

It turns out that rendering EPUB correctly in a terminal is a messy business. The format was not designed to be displayed in a terminal at all. And for a reasonable reading experience, you have to solve the problem of navigating the book correctly, which the AI struggled with quite a bit. Instead of a few hours, the agents worked for days, adding more and more features, until I was comfortable sharing the project in February this year (2026).

It got five up-votes and eight stars on GitHub. Perhaps there is not enough interest in terminal-based EPUB readers. Perhaps there is a general dislike of AI-built software. But had this project been published a few years earlier, I doubt the interest would have been so little.

With the recent release of Claude Fable 5 and GPT Sol 5.6, I added quite a few more features. Yet I doubt it would get even five up-votes this time.

With the abundance of software comes the devaluation of software. "Why would I use this, when I can ask my agent to spit out something similar, but closer to my own likes and dislikes?" A thing anyone can have made at will is a thing nobody treasures.

Books may be an analogue, even before authors began publishing hundreds of AI-generated novels. I used to read a lot, sometimes two dozen books a year. I kept an eye on authors I liked and tracked what they were publishing. Then I realized there are already far more good books than I can read in this lifetime. So new books, even great ones, mean nothing to me anymore. It is not that the books got worse. It is that my time was always the scarce thing, and now I can see it.

If we are really seeing the dawn of AGI, perhaps humans have to reckon with this: when there are far more intellectual artifacts --- games, software, books, music, films --- than anyone could consume in a lifetime, what is the meaning of creating more?

But maybe the question is older than it looks. Seneca complained two thousand years ago that the abundance of books distracts. There was already more than one life could hold; only the speed is new. And nobody ever wrote to finish writing, or read to finish reading. I did not port epy because the world lacked an e-book reader. I ported it to have something to fiddle with --- an afternoon that became months. The worth of making was never in being consumed. If it keeps my mind off the depressing things happening in the world, that is meaning enough, five up-votes or none.

dawdler-purge··on AI and Math State of the Art
I have created a slide deck to summarize the progress of AI in mathematics in the last 12 months.
dawdler-purge··on Show HN: Three new Kitten TTS models – smallest less than 25MB
Please let me know when this has been fixed. I will update purr to make the installation steps simpler.
dawdler-purge··on Show HN: Three new Kitten TTS models – smallest less than 25MB
I don't think so. It is perhaps a bug to have this unnecessary dependency. I expect the author of kitten to fix this soon.
dawdler-purge··on Show HN: Three new Kitten TTS models – smallest less than 25MB
I created a CLI wrapper for Kitten TTS: https://github.com/newptcai/purr

BTW, it seems that kitten (the Python package) has the following chain of dependencies: kittentts → misaki[en] → spacy-curated-transformers

So if you install it directly via uv, it will pull torch and NVIDIA CUDA packages (several GB), which are not needed to run kitten.

dawdler-purge··on Warranty Void If Regenerated
The LLM-ness isn't a hard problem to fix. Break it into sections, run each through an LLM a few times to catch logic issues, use different AIs to double-check. For the writing style, if the author just read it carefully, they can definitely spot the things Claude keeps repeating, and tell it not to do that.

But honestly, the ideas here are really good. The cascading failure from a weather model update, the spaghetti problem with forty tools nobody designed as a system, the $4 toggle switch being the most important tool --- that's sharper thinking about AI than most serious essays on the topic.

A lot of people who publish regularly can't write to this level of thinking. The prose could be cleaner, sure, but it made me think, which is more than most stories do.