Notices by Wolf480pl (wolf480pl@mstdn.io), page 2

@argv_minus_one

@kimapr @ak @erik @shiz

@lynne
AI powered FOIA redaction automaton.
It puts in just enough blanks that sometimes it's possible to guess what was in them.
I wonder if it was trained on language-learning textbooks...

@lanodan
so basically the industry discovered rubber domes and gained the ability to make keyboards much cheaper?

wait a sec...

the Space Cadet keyboard had linear hall-effect switches?

Not a fan of linear but like

they were making hall effect keyboards in the 70s, even though there wasn't a broad customer awareness of keyboard switch types?

and nowadays the keyboard manufacturers are only starting to discover hall effect after a failed attempt at optical switches?

iwd, why...

https://git.kernel.org/pub/scm/network/wireless/iwd.git/tree/src/wiphy.c#n832

@domi
> loadavg 52

:blobcatfearful:

@lanodan @ayo
I was thinking of Naruto: Shippuuden taking forever to get to the conclusion, with every 100 episodes covering half of the in-universe time than the previous 100 did (though I haven't watched it until the end so I don't know that first-hand)

@ayo
Anime Movie: Final Season Summary
Anime Final Season Four (For real this time)

Bonime

since memory is expensive, can I get a job optimizing memory usage by rewriting things in C?

@lynne
Streetlights should have a warm light to tell my brain that it's nighttime.

Cold/"neutral white" streetlights should be forbidden.

@lanodan so no journal to replay?

@lanodan
does ZFS have some form of crash recovery that runs on the first mount after an unclean shutdown?

@Suiseiseki
what in that slide portrays them as a good thing?
@bagder @via4 @p

@Suiseiseki
is it an endorsement to acknowledge your invention, in addition to being used for good, is also used for evil?
Are you morally required to pretend the evilbusecases don't exist? Or treat their names like taboo?
@bagder @via4 @p

@via4
I think a solution would be to include on the slide someone so obviously evil, that it's clear to everyone that the slide is merely yo ilustrate the spectrum of users and not an endorsement.
@p @bagder

@jannem @mntmn
AFAIK the way these LLM tools work is they have an embedding of words into a vector space, they index text by converting every word in a every document to a vector, and storing it in a database together with ID of the document it came from, and then when you search, they turn each of the query words into vectors, and search for K nearest neighbors in the vector space for each of them.

Then they feed the documents they found to an LLM.

What if you skipped the last step?

@mikoto
how does it work?

@mikoto
>IPv6 NAT

oof

but I guess assigning the IPv6 addresses to wireguard endpoints by hand has become tedious?

@Suiseiseki
like wtf why would that be per card? Is that per card in ath10k too? what are they saving by making some cards not support it?

@Suiseiseki
> depending on the card

so your particular ath9k might not? yay, how useful