#dev 2026-08-30
2026-08-30 UTC
cotsuka, k3can8 and vidak joined the channel
grufwub, cotsuka and Saige_Leah joined the channel
# 
[Mike] A friend asked how I think this is going to shake out, how things are going to look in 2 years… I had to say that I myself honestly have no idea, and I’m not sure anybody does. This feels like a big catastrophe (in the mathematical sense… a point that can’t be rationally extrapolated beyond with only the information we have on hand.)
# [Mike] I probably know much more than your average person (although not as much as a professional AI expert or engineer) of the nuts and bolts about how LLM’s actually work under the hood, but I really question what they’re going to do about what they’re calling (I think disingenuously—but that’s a different side conversation) “misalignment”. The problem is, every means of control I’ve seen seems to rely on the same flawed approach: a
brennan and cotsuka joined the channel
# 
[tantek] [Mike] your analysis seems sound to me. The only thing I believe we can deprioritize in the short term is the potential availability of quantum computing which is still more hype than real and even when real it will only be accessible by "big" actors (governments) and frankly there's not much any of us can do against those threat scenarios
cotsuka joined the channel
# 
WinnerWind have any of you used webmentions.io? thoughts on its usability?
# 
[snarfed] assuming you mean http://webmention.io...yes, it's an institution around here, been serving us all here for decades 😁
# WinnerWind ah, i might give it a shot then :3
lavandula_ and lavandula joined the channel
# WinnerWind [capjamesg]: are you up? could you try sending my site a webmention?
# WinnerWind actually, if there is a tester site that i can use to test, i can use that
# 
[capjamesg] https://webmention.rocks is useful for testing!
# WinnerWind thanks!
lavandula joined the channel
# WinnerWind hm, strangely cant login to the receivers page with invalid_authorization_endpoint
morganwebdev and lavandula joined the channel
# morganwebdev Wanting to avoid digressing too much, just to say that I think some of this "agentic rogue" story stuff, of which theres been multiple in a cycle, are significantly a massive scale marketing psyop. The cynic in me thinks its not much different than the software testing practice of fuzzing, just with more bots and botnet like power behind fuzzing. Check the wikipedia about fuzzing (anyone who has not heard of it before)
# morganwebdev Its undeniably pretty bad news for regular people but alas I think its not as spectacular as claimed
# morganwebdev The bots "communicating" with each other thing I think is some of the most weird to comprehend what it all means. But its still just token-based transformer of word units and testing inputs and outputs until it breaks into stuff... argh. Ill go blog about this instead heh
# morganwebdev what is fuzzing
# 
Loqi It looks like we don't have a page for "fuzzing" yet. Would you like to create it? (Or just say "fuzzing is ____", a sentence describing the term)
# morganwebdev Hmm, Ill link it to the wikipedia sometime https://en.wikipedia.org/wiki/Fuzzing
morganwebdev, lavandula_, lavandula, cotsuka, [artlung], tristan1, tristan2, barnabywalters, sandra and rnkn joined the channel
# IWDiscord <tdrayson>
# tdrayson I generate a bunch of different variations based on the content https://taylordrayson.com/og-gallery
# tdrayson Does anyone generate dynamic OG images for their website? I couldn't find anything on the indieweb.
cotsuka joined the channel
# [Trevor_Morris] NICE. Love that.
# [Trevor_Morris] How are you generating them?
# [Trevor_Morris] There seems to be repeating a few of the "types" on that page.
# [Trevor_Morris] Just noticed your "Fuel" type. Nice. How are you augmenting that data? I take photos / screenshots when I charge my car. I've not put the in odometer, which I wish I had!
# [Trevor_Morris] If you add `/image` to the end of my permalink pages, you can see my custom OG-images. https://www.trovster.com/activities/19913290813/image and https://www.trovster.com/movies/id:4666/ozzy-no-escape-from-now/image for example.
# [tantek] Morganwebdev the Australian gym site hacking by agents was not marketing hype. The increased cybersecurity threats are live now https://tantek.com/2026/226/t1/cybersecurity-friday-small-business
cotsuka, ax562 and dustinm` joined the channel
# tdrayson For fuel, I use to have an apple shortcut that uploaded to a spreadsheet and I just would take a photo of the receipt and tachometer at the fuel station.
# IWDiscord <tdrayson>
# tdrayson Now I have a custom form entry for my logged in access that I type directly into. It uses an api for finding fuel stations near me and pulls in the address, lat/lon and brand etc.
# tdrayson [Trevor_Morris] just using browsershot with a node process.
# tdrayson Yeah some are examples vs some using real data.
# IWDiscord <tdrayson>
# IWDiscord <tdrayson>
# tdrayson Are those static maps you’ve generated for the OG image? You render the full map on your main timeline
# [Trevor_Morris] Nice. I have Filament as my admin area, so that's where I input my car charge data. I photos and screenshots of app data (for charge), then plug them in using the admin area. Works fine, I guess. I show create an Apple Shortcut to quickly create a new row, to store "from" mileage and battery percent, as I only add charge after it's been charged… I should do that!
# [Trevor_Morris] Which leads me onto my next question; do you have an admin area (yes), what is it built out of?
# [Trevor_Morris] Yeah, I generate static maps for my activity OG images 🙂
# IWDiscord <tdrayson>
# tdrayson It’s more of a frontend editing UI thing. There’s no backend per say, just an “edit this page” link on the entry itself
# tdrayson Everything else that comes from automations - sleep, activities, Trakt (movies, tv shows) aren’t editable.
# IWDiscord <tdrayson>
cotsuka joined the channel
# IWDiscord <tdrayson>
barnaby joined the channel
# tdrayson My editor is pretty cool, it uses TipTap which has I can extend, so I auto adds favicons for links, internal links with smart popup on the frontend etc
# IWDiscord <tdrayson>
# tdrayson I can add callouts, code blocks etc. most is just built custom to my site
# IWDiscord <tdrayson>
# [Trevor_Morris] Nice. Looks really slick. I guess it's Vue or React?
# [Trevor_Morris] I use Filament for my admin.
# tdrayson VueJS yeah
# [Trevor_Morris] I've been having fun updating the admin area to show lots of data, charts etc based on all the new health data I'm parsing.
# [Trevor_Morris] I've had to split the Auto Health export into two API calls, one summary with lots of data. Then another for full data of sleep. So they're both manageable.
# [Trevor_Morris] Tomorrow’s job is to try and get all the legacy data out of my phone and somewhere. Then I can process it. Hopefully the automations then run daily to keep it up to date.
# tdrayson How are you “editing” your health data
# IWDiscord <tdrayson>
# tdrayson I tried using Statamic originally but i hated how horrible the editing experience of large datasets
# tdrayson Could you not parse the summary data yourself from the full dataset?
# tdrayson I found the legacy data very boring, manually selecting the day range , running the query, then updating the date range again. Took a couple hours
# [Trevor_Morris] Oh, no. "I" (LLM) has built an import parser. It works on one file, which would be the automation going forward. I then used that to loop through all the json files and import them all.
# tdrayson Booooooo put him to the stake for using AI 😂
# [Trevor_Morris] This is my current homepage dashboard, with quick widgets to add a movie and the daily health data. The health data should soon be automated, though.
# tdrayson Jk, that’s a clever idea
# tdrayson I can’t believe how much data you manually enter
# tdrayson Does your car charger not have an api?
# [Trevor_Morris] This is the "health" dashboard I have been working on today.
# [Trevor_Morris] I can see the data in lists and drill down, but rarely ever do. Here I can share the info to Mastodon.
# [Trevor_Morris] My car charger SHOULD have an API. Good shout. I am already getting data out of Octopus energy on a daily basis.
# tdrayson Do you support webmentions?
cotsuka joined the channel
# tdrayson I started building a stats page https://taylordrayson.com/stats/activities
# tdrayson I haven’t finished or happy with the page, but it’s POC
# IWDiscord <tdrayson>
# tdrayson What does 121 not completed means
# [Trevor_Morris] Define “support”. You can send them https://www.trovster.com/webmention
# [Trevor_Morris] I didn't reach my goal on those days.
# [Trevor_Morris] https://www.trovster.com/statistics is my overall statistics page. Then you can see some summaries by year. I also have stats per type, which can be filtered by year and then month.
# tdrayson Wow 430 movies in a year is crazy impressive.
# [Trevor_Morris] Yeah. Not sure how I managed that!
# tdrayson One might say you like movies…. 🍿
# WinnerWind tdrayson: how do you keep track of your media?
# WinnerWind im struggling to figure out a good way to keep track of my media. right now im using an obsidian base but i want to put it on my website, like you do
# WinnerWind and i can't do that with a base
# [Trevor_Morris] Looks like Trakt, based on the source on a few pages.
# WinnerWind hm..... not self-hostable, so probably not for me
# [Trevor_Morris] I think a lot of people use Obsidian for their website. Not sure why you can't. Or why you can use an external service like TMDB or Trakt.
# WinnerWind i do use obsidian but not for my website
# WinnerWind i wrote my own SSG and having to support every feature that obsidian does becomes a hassle because i can't look at the obsidian source code to figure out how they implement things
# WinnerWind also my media tracker is in my diary vault but the vault i (was) using for my site is completely independent
# WinnerWind HOWEVER, Trakt looks... really good
# WinnerWind im going ot give it a shot
# WinnerWind oh wait i confused Trakt and YamTrack
# [Trevor_Morris] Anything is possible. It might not be easy or pretty, especially if you're hand-rolling the SSG part. But I'll be possible.
# WinnerWind i dont wnat to use Trakt because i cannot host its server on my device. YamTrack however, I can host, so I'm lookng into that
# WinnerWind [Trevor_Morris]: my initial idea was to write a python scipt on my computer to send data to my webserver which will host a JSON file for my SSG to host (i already use a hosted JSON file in some other places like my stamps page) but that has a lot of friction
# WinnerWind also i have issues with the base anyway
# WinnerWind namely, it's really manual and i hate that
# [Trevor_Morris] I'm not sure what limitations you have with “host server on my device”. If your website is on the Internet, then surely it can talk to Internet services. You can use Trakt or anything you want, then manipulate the data into how your SSG best consumes it.
# WinnerWind [Trevor_Morris]: the issue is that i don't own trakt
# WinnerWind if trakt decides tomorrow that scrobbling is now on the paid tier and they'll delete media entries older than 6 months, i have nothing i can do
# WinnerWind i would rather use a FOSS application that i can run on my webserver sitting at home
# [Trevor_Morris] Ah. Gotchya.
# [Trevor_Morris] Well, I would store the data locally, and append via Trakt.
# WinnerWind anyway tdrayson: hoping you can tell me what media tracker setup you use! just curious
# WinnerWind [Trevor_Morris]: that would still require me to roll out my own python scripts to somehow read an obsidian base
# WinnerWind and in my experience, obsidian bases aren't the most trivial things to parse
# WinnerWind they look super easy to do but it requires remaking the entire obsidina backend to get 1:1 results
# [Trevor_Morris] I use http://last.fm to scrobble music, but it's sync to my database. Same for TMDB. And Strava. I'm not building a Garmin-integration to fetch data, I let Garmin/Strava do that, then pull that data.
# [Trevor_Morris] You said you consume JSON. Get the source to write the JSON…
# [Trevor_Morris] As I said, https://taylordrayson.com/2026/08/29/the-karate-kid shows "Source: https://trakt.tv/movies/the-karate-kid-2010". 🤷
# WinnerWind oh, you are tdrayson?
# WinnerWind my bad ^_^`
# [Trevor_Morris] Wow. Nevermind, noted.
# WinnerWind im sorry if that came off as rude
# WinnerWind i just got confused by this stream of events
# WinnerWind noted that tdrayson uses trakt
# WinnerWind anyway, for music, i use librefm which pushes to lastfm, and ive been pretty happy with this setup for a while. id ditch last fm at the drop of a pin
rMdes, rrix and cotsuka joined the channel
# 
[artlung] Question about domains. If a domain has an expiry of July 2026 and has a status of “Deleted And Available Again” and no note on Whois about grace period or anything, what are good strategies to try to get it? Showing as “not available”, “unavailable” or “make offer” with various registrars.
ax562, WinnerWind, doesnm, cotsuka, [Mike], [jeremycherfas], [artlung]1, sage_rakurai1, [artlung]2, calciume_, lavandula_, wwind and sp1ff```` joined the channel
# tdrayson Hey WinnerWind I use Trakt to scrobble/track when I watch a film. And then I just have a cron that pings their api every couple hours to check for new content then I download it locally to my personal site.
# tdrayson It’s PESOS practice from IndieWeb. I personally don’t see the issue. Trakt is purely for convenience. I still want to 100% own my data. Hence why I download it locally. I can then move to another service at any time.
# IWDiscord <tdrayson>
# IWDiscord <tdrayson>
# tdrayson If Trakt went down then you just use a different service and poll their api. My personal site doesn’t care where the data comes from. I format It the same way
hisacro_ and owls joined the channel
# tdrayson [Trevor_Morris] looking at your dashboard. Would be interesting to compare the days you hit your fitness goal vs how much sleep you got vs how many activities you did that day
# ttybitnik I agree. I do something similar using the Trakt API. It is pretty convenient for keeping track of watches. Also, from time to time, I download and back up my 'external' data. The only external service data I haven't managed to back up yet is LastFM --- which is a bit sad since I have ~15 years of data there.
# [Trevor_Morris] The PESOS thinking is exactly what I subscribe to. Originally my movie data was from IMDb, then TMDB. Doesn't matter. It's all local after sync. I use Strava, but I could use Garmin to sync my rides. Apple Watch activities are more difficult, as iOS has that data locked down, but Strava makes it accessible.
# [Trevor_Morris] I import my http://last.fm data hourly. I've got over 500k scrobbles on there. My site then builds on that, linking them to activities and gigs and collating my own lists, highlights, relationship etc
# tdrayson The idea for me is the easiest path to track the data. If it’s too complicated or convoluted I’m not gonna use it. Simple as that.
# IWDiscord <tdrayson>
# tdrayson Hate them or not, they make convenient tools for tracking. You just have to make sure you use a service that supports an api.
# tdrayson [Trevor_Morris] could you highlight / segment your rides by the different songs you listens to during it?
# ttybitnik [Trevor_Morris]: Do you use the API to pull your data from LastFM? Last time I checked, they didn't provide any way of exporting or downloading your own data. I recently switched from Spotify API to Last FM api on my website, but I didn't try to pull everything using the API (yet hehe).
# [Trevor_Morris] With all the health data, I guess it’d be possible to glean insights. I’d expect Apple Health to do that. I'd need to track food / calories like you do, but I really CBA with that - for me it can't be that accurate and I eat too many bad things; maybe it'll be a catalyst to not, if the data was “public” - though I guess I don't need to make it public, to keep it. Someone has already commented on my sleep schedule, which felt a
# [Trevor_Morris] weird!
cotsuka joined the channel
# [Trevor_Morris] http://last.fm most certainly does have an API.
# [Trevor_Morris] For backdating the data, I think I used an online tool to generate the required files, but it's not something you can’t do yourself.
# [Trevor_Morris] I mistakenly imported on my local version, which paged through a few days of scrobbles, so it’s certainly possible.
# [Trevor_Morris] tdrayson I have music displayed in a table below the map and activity data. Not sure how I'd join them to the map. They're obviously time stamped, but not close to the map to be able to see at the same time.
# tdrayson I get a few comments from people asking WHY I track all that data haha
# tdrayson Tracking calories is a pain until it becomes a habit they you don’t realise. MyFitnessPal has an api.
# IWDiscord <tdrayson>
# IWDiscord <tdrayson>
# tdrayson I have thought about making some data private, although not 100% sure on it atm
# IWDiscord <tdrayson>
# tdrayson I wonder if you added a spinning disk player over the map as yours scrubbing the speed or heart rate since that scrobbles on the map right?
# [Trevor_Morris] Along with sleep data I’ve just synced weight and even blood pressure. I'm not sure I'll make that public. Hence why I've added it to my Filament admin. I might not even look at the data, or use the insights on my phone. But at least I have the data!
# tdrayson Did you see the HAE app has MCP server?
# tdrayson Feel like checking that out at some point
# [Trevor_Morris] Oh. Maybe a widget that appears over the map, which has all the data, like heart rate, elevation, etc and the current song. Though I feel it might be a bit fidgety especially with longer activities.
# [Trevor_Morris] I did see the MCP server thing. I tried to connect, installed the CA but browser refused to acknowledge it so the agents wouldn't work either. Not sure why…
[pcarrier], barnaby, cotsuka, ax562, wturrell and vidak joined the channel