I think it's very useful to split centralisation from 'centralisation you can't escape' - like I *want* centralisation a lot of the time, but I also want a credible exit (e.g. we see this in big OSS projects where central effort is good and forking is your credible exit)
07.03.2026 01:17
π 3
π 0
π¬ 1
π 0
if we look at verifiers, etc. and consider inverting that relationship - "I am a member of this community, here is my token, my root is <x>" and then let features hang off that (discoverability, etc.) that would also help move the control 'out' so it's less about 'trust me first, then...'
07.03.2026 00:22
π 4
π 0
π¬ 0
π 0
not sure if that solves the problem wholistically - but it helps make community more durable (like a subreddit can survive the exit of a chunk of the mod team)
07.03.2026 00:19
π 6
π 0
π¬ 0
π 0
whenever I have discussions around what would work, the conversation tends toward co-owned repositories (that would survive a user leaving/losing interest/etc.) and ways to express degrees of trust (so I can express "I want to use this labeller but only to label never block")
07.03.2026 00:19
π 7
π 0
π¬ 1
π 0
one of the clearest gaps in the ecosystem - working against the lofty decentralisation goals - is how power concentrates in tools: lists/feeds/labellers
the implicit decision by bsky rn is "don't let these things become powerful", which also means "don't let these things become *too* useful"
06.03.2026 23:28
π 35
π 3
π¬ 1
π 0
but they are automated - idk that there's a better word that means "no human at the controls"
but I suppose it matters what people think the label is for - like is it for filtering, spam reduction, avoiding arguing with those AI auto-repliers, etc. etc.
06.03.2026 19:30
π 2
π 0
π¬ 0
π 0
why why?
05.03.2026 10:16
π 7
π 0
π¬ 2
π 0
just makes it lag a bit
the sunβs out and Iβm riding my bike, so it is a good day
05.03.2026 06:20
π 1
π 0
π¬ 1
π 0
yea it should sort itself out after a little bit - itβll see the migration as a change and then should clear on the next post; but will have a look when I get home
05.03.2026 06:12
π 2
π 0
π¬ 1
π 0
A push notification from copilot:
A Poetic Start to Your Week
See your name transformed into a poem by M365 Copilot.
my corporate dystopia now gets copilot-induced whimsy
02.03.2026 01:10
π 46
π 7
π¬ 3
π 1
Had to Reinvent the Sailing Block (To Replace Them All)
YouTube video by Reversing Entropy
I don't know anything about sailing but I love a good information-wants-to-be-free story
28.02.2026 09:09
π 5
π 3
π¬ 0
π 0
I'm kind of assuming but the random power brick I was using was very very warm
28.02.2026 02:37
π 3
π 0
π¬ 1
π 0
does the rpi use vmlinuz-6.8.0-101-generic? no
so TIL I was separately remoted into the wrong machine
28.02.2026 02:36
π 2
π 0
π¬ 0
π 0
well TIL why the RPi people get particular about USB power
28.02.2026 02:33
π 14
π 0
π¬ 3
π 0
well they do it on a computer I suppose
28.02.2026 02:27
π 1
π 0
π¬ 1
π 0
my feeds broke so I added a watcher and now I need a watcher watcher and well, that's something they claim is a real job
28.02.2026 02:24
π 1
π 0
π¬ 1
π 0
we call this site reliability engineering
28.02.2026 02:21
π 1
π 0
π¬ 1
π 0
ok I think we're recovering
so the LAST time things broke I added a monitor and restart job and now I need to add a monitor and restart job monitor and restart job
28.02.2026 02:11
π 10
π 2
π¬ 4
π 0
Panic! at the vmlinuz-6.8.0-101-generic
28.02.2026 02:04
π 13
π 2
π¬ 1
π 1
yeah something is broken
some people might be seeing newer stuff, there's a cache so it might be flipping between broken (not refreshing) and not broken
28.02.2026 02:00
π 3
π 1
π¬ 0
π 1
oh stuff is broken I think cf went down and now things aren't coming back π
28.02.2026 01:52
π 3
π 1
π¬ 1
π 1
hmm let me check
28.02.2026 01:41
π 1
π 0
π¬ 1
π 0
it's a challenge π
24.02.2026 20:36
π 1
π 0
π¬ 0
π 0
amendment to the RFC: Records MUST match the vibe of site.standard.document.
24.02.2026 20:36
π 2
π 0
π¬ 0
π 0
a male and female gang-gang cockatoo on a dead branch, they are looking at the camera
fierce lil guy
14.02.2026 20:14
π 25
π 4
π¬ 3
π 0
The maximum current through the power cord is 16A, and the maximum power used by the power strip is 3500W (can't be used for a long time).
idk much about electrical safety but I'm pretty sure this power board is all about insurance fraud
24.02.2026 20:23
π 6
π 0
π¬ 2
π 0
I think any model that has some kind of either consensus or verification can work - ultimately you just need to indicate "do not trust this source", and the ecosystem needs some alternatives to go to in that case
22.02.2026 06:01
π 1
π 0
π¬ 0
π 0
or at least from the new plc foundation - there are probably a bunch of workable scenarios, but they can't work without initial cooperation
22.02.2026 05:59
π 3
π 0
π¬ 0
π 0
there are systems that could work - like CRL lists, and could introduce an external validation to the resolution (i.e. βthese 3 directories agreeβ¦β)
but in terms of βtake this record with meβ, no I think youβre right
21.02.2026 22:20
π 1
π 0
π¬ 0
π 0
it should be possible with an βIβm over thereβ entry, but in the adversarial case they may just not publish that - so you also need a PDS attestation or smth, and means to re-point records which is where it gets complex because you have to care for *every* record (client caches, etc - very messy)
21.02.2026 22:13
π 2
π 0
π¬ 1
π 0