https://en.wikipedia.org/wiki/Flattr figured out some parts of this. Notably, you picked your own total monthly donation, and then clicked a button on participating sites to allocate a fraction of your total to them. AFAICT it worked as advertised, but raised new issues with donation behavior. E.g. I obviously like curl every month, so should I click its button monthly? Twice monthly? If I am a developer of some other useful OSS software, should I click curl's button and the curl devs click my button? Does the money just slosh around between merchant-customers? Is that good?
I am trying to make efficient gst pipelines for security cameras. I only took a quick look at GPS so far. The UX for making connections is a little weird, but so far the whole tool seems to work as advertised.
By having an LLM name the chapters, I meant having whisper do speechrec on the chapter and then asking an LM to summarize the content into a name up to k chars.
Yeah, I play a VHS tape and capture the whole thing. Maybe I should be using a scene detector to split files on camera cuts, which would be roughly correct for home movies (but not for TV shows).
That's where I am now. I'd like to optimize out the retyping and duplication of time strings.
I want the player ui (I'm using mpv) to have a command that:
1. Remembers the last end time to use as this chapter's start time
2. Gets the current time to use as chapter-end.
3. Accepts the name (e.g. 'chapter1').
4. Runs the ffmpeg copy command.
Perhaps mpv+lua can already handle this. I see commands for setting a loop range and for calling a subprocess. Not sure how I'd input the chapter name. Maybe I'll have an LLM name the chapters for me :)
Can you recommend a tool for dicing up 2h digitizations of VHS tapes? I want to play the 2h video, seek around easily, mark 'chapters' and give them filenames, then do a no-transcode rough cut extraction of each chapter into its own video.
I seem to recall yet another one, maybe with a name that invoked a traveling group of helpers who would jump into projects briefly to fix them up?
On a related note, it would be cool if there was a way to leave a hobo sign equivalent if you find a project that is well-run and easy to contribute to. If the build and tests Just Work, etc., we should praise that project in a way that
1) encourages helpers to pick it since they'll have a good experience, and
2) provides a good example for other projects to follow.
I threw something together to address that goal. It was really hard to render obsidian-md into html that looked about the same. I started with general MD libs, then ultimately used obsidian-html. Obsidian's MD is not standard at all, though I'm not sure any MD is fully specified since there are so many edge cases.
107k photos here, and it's not great. I just clicked a month in the calendar and getting just the json back took 1.8 sec. Last of 12 images came in 2 sec later. Postgres, SSD.
Plus: all thumbs of non-square images are cropped, and it makes all this per source image:
The UI doesn't change the url to the photo you're looking at, so you can't share urls.
Images aren't links, so you can't do browser stuff like open one in a new tab. The back button doesn't work right at all.
The choice of metadata to display under each image in the listing is downright silly ('place', city, year, DoW, month, year again, time, zome, camera model, resolution, file size, city again, state, country).
The information scheme features 'albums', 'favorites', 'moments', 'labels', and 'folders', as if they were being forced to match a bunch of legacy systems. I want a powerful tag/group mechanism that I can use for many possible workflows.
The list under 'search' says 'Review 40193', clearly telling me I have that many pics to review. Clicking on it says:
No pictures found
Try again using other filters or keywords. In case pictures you expect are missing, please rescan your library and wait until indexing has been completed. Non-photographic and low-quality images require a review before they appear in search results
(And I had to disable some user-select:none styles to copy paste that)
It seems like something happened around 2026-05-01 (https://bigasterisk.com/post/delivered-email-rate-2026.png) where it became easier for spammers to make technically legitimately deliverable mails. Or my postfix is setup wrong.
Note "spf=pass" and "dkim=pass" in these headers: