@[email protected] my school used free teams and free zoom because who cares
until the govt provided it with paid teams
though its sso was so inconvenient that nobody really liked it
@[email protected] why not generate slug at creation/edit instead of having to parse the thing? db will ensure you won't run into any edge cases, and you would also get custom urls for free
also their federation model is pull-based.
unlike activitypub, which is basically glorified email, afaiu atproto federation expects you to basically have the entire (or most of the) network locally and continuously crawl them for updates. making it much more expensive to host personal instances.
which they solved by... uhh.. relays
> The Relay handles "big-world" networking. It crawls the network, gathering as much data as it can, and outputs it in one big stream for other services to use. It’s analogous to a firehose provider or a super-powered relay node.
whats the point of federation then honestly
side note: they also support `did:web` identifiers, which (obviously) do not use that, making the dns the trust root.
which is kinda fine, but i'm not sure if their frontend supports creating such accounts, and im not really willing to test that via atproto right now.
but the issue is that the `did:web` identifiers will only work for as long as you have access to that domain, basically killing the whole point. and afaiu the protocol doesn't really support migration across did-s yet
there's one major thing that concerns me, though (as far as i understand, i may be wrong, please correct me if i am)
so there's such thing as "PLC Server" (https://plc.directory). which is basically a global thing responsible for resolving their `did:plc` identifiers to pds and <i>user's public keys</i>. a *centralized* thing in a federated network, yep
and im not really a security expert, but aren't we putting too much trust on that plc server? like, what if an attacker happened to take over that server? there's literally nothing stopping them from changing the public key in the database
like yeah, there are publicly available signed audit logs (which are basically a per-did blockchain), but how can a third party know they were not tampered with too?
the only (kinda) solution i could think of is to continuously validate changes to the directory and reject improperly signed ones. and i haven't found the reference implementation doing that.
@[email protected] @[email protected] that's not necessarily a bad thing though. bad products are basically what drives us forward.
mistakes happen, even tech giants make them, but only if they don't improve on them that becomes a problem
@[email protected] never knew lenovo made "work tech", huh. they indeed market them towards businesses, never noticed that
also i still don't get how "having a sim card" is a business requirement for a tablet. maybe its true for *your* business, but definitely not for most businesses
@[email protected] because actually most of the time you don't really want cellular on tablets, which allows to save some money (both for you and the manufacturer). even apple sells non-cellular ipads
i honestly never had a need for cellular on tablet, i can always just hotspot from a phone. and it's also much cheaper (basically free, since i already pay for my cellular plan) than having a separate plan just for tablet
i dont need your fucking life advice
i dont need your fucking money
i dont care how you struggled when you were my age
my life isnt fucking awesome either
stop fucking guiltuing me into doing what you want me to do, i want to live my own fucking life
dont you love it when your ci/cd pipeline takes 2 fucking hours, so you have to overtime on friday just to release your 3 lines patch
also its run on like 10 year old xeons that go on aliexpress for 5$. best fucking cpu to run webpack