did:w(h)e(y)b

This parallels atproto apps offering atproto acc signups via their own hosted PDS btw

What would it look like to have a did + domain handle ecosystem outside of atproto PDSes? Would it mean total PLC dominance? just because its cheap and accessible? Or would it just be like the services offering did web signup under their own domain, kinda like fedi instances, but risking lock in?

i wonder if i wouldve got travel support approved if i had applied (which i didnt cuz i assumed i wouldnt)

i like fluxer it looks much more polished than revolt/stoat but im really uncertain about its "federation" plans, like i really dont want to have my id to look like whey#0123@fluxer.app

this ended up being useless, cuz like i deleted the mediaanalysisd cache and i recovered 72gb just from that

Ideas on extending Open Graph embed displays for atproto I guess I also felt for the simpler feel of OGP and was unaware it’s a facebook creature that hadn’t evolved in ages. Schema.org is much less known beyond tech folks but it is solid at least and it do...

Holy shit its RDF discourse.atprotocol.community/t/ideas-on-e...

the personal independence idea isnt truly viable anyways because the nature of hosted websites and the responsibility that comes with it, no host of RD is (or should be) "neutral" (or unmoderated). which is why im leaning to medium size communities, removing the implicit bias and making it explicit

i think RD has always been a client that caters to hyper personal independence, as the client constructs and fetches the posts on your device, via a SPA. so extending it into medium sized communities that rely on big expensive infra is not too out there

screenshot of the AppView-First toggle in RD "Prioritize using an AppView to fetch posts before using microcosm" with an appview url input text box below it

featuring the scary existential settings toggle

ok whatever this is the next version of RD with its half-finished moderation screen test.reddwarf.app

also im writing this mostly because im pissed off at myself for still not releasing the moderation update for RD yet, which is a prerequisite for the independent authority thingy. and it was delayed because i hate how slow the app is, and im not sure how much further i can push this

screenshot of my bsky draft: a bsky client, with labels turned off, then use a fallback hydrator for the rest of the unavailable content from the appview / new types of content, and then re-apply labels according to community policy. is this a viable path for independent**** community platforms?

i was gonna post this but it sounds like a really weird bsky social-app fork instead of a rethinking of RD away from proving a point about infra independence into independence of authority and governance from bsky while still relying on bsky infra (because money)

all i use it for is the folder size visualizer. when i was on linux and windows i used KDE's filelight but it didnt work on macos so i found this similar thing that also has pretty visuals. Using these tools is much faster than manually cmd+i each folder to check the size

cleanmymac removing 4gb worth of node_modules and also whisper (openai)

die node_modules

not because im unaware, but like the sheer inaction and distance i have between things i care about

i feel like i live under a rock

screenshot of graze.social feed debug to the post above, its text classifiers for topics has a near 100% confidence for gaming for the above bait post

im gaming the gaming filter

i am #gaming my deadlock video games it is very awesome to join a match and play video games to play the new #level of the #stage of the #video #game to beat the #final #boss and game my way to win the @game

also graze is a whole lot more annoying to update the regexes because of how it separates the regex sources from the bsky post fields, instead of one copy paste its 6 or 7

current gripe is posts that dont show up in the feed but is valid when passed through the graze individual post tester debugger thingy

currently testing my graze migration for the deadlock feed.
i had to remove the gaming ml filter cuz while it was cool it filters out valid posts just because its not very #hashtag #gaming and the goal is for the new feed to not lose any of the content from the old feed

hmm it might be removed because the name was too similar to other terms? idk i had done this for some terms but idk about abrams. iirc character names are restricted to only match with "game" "deadlock" "match" and other specific matches

oh my god i didnt have Abrams in the feed this entire time. how could i forget about Abrams. poor Abrams

the new deadlock update changes a lot of lore too which means the feed's regex has a lot of things to be modified. iirc the sapphire flame and amber hand are obsolete terms now, replaced by whatever the new team names are. the problem is i dont play deadlock anymore so i need to read the wiki

cite your references for what caused the initial misunderstanding

i should do this today

the dangers of just saying things, llms believing in whatever you say

the actual version names is TBD especially regarding the idea of a 1.0.0 release being somewhat important ish, so maybe delay the actual 1.0.0 number for when layered moderation is finished

okay ive decided, no more delaying labelmerge sync, the roadmap will be strictly in this order, for my own sanity
1. finish labelmerge cache sync, polish up the client-sided label stuff, release as 0.8.0
2. do blocks and mutes and pref writes for 0.9.0
3. big refactor for 1.0.0

also this vague idea of rearchitecting RD is rooted in a desire of a native mobile app. i have a prototype of UPR rendering posts in RN with nativewind but not anything more than that. i have another test for a lynxjs port but its react compat is worse than RN despite supporting tanstack router