they added an ai summary section. it’s disabled by default but it consists more than half of the popup and if disabled has a big blue button on it that activates ai features, the only big blue button on the popup, and it says “continue”. or, it does now. it used to be bigger and say “See more with AI” or some shit.

all of that is correct and also basically what i said. the reason it became a big deal was because of accessibility. the definition of a “long press” is short enough that older users, who tend to hold the mouse button down for longer after a click, were suddenly seeing popup windows everywhere and, believing it to be an issue of the site they were on, assumed their popup blocker was broken. the timing was adjusted to one second in an update and also the long press shortcut was made optional.

when it was pointed out to mozilla that having a popup containing one big blue button in a popup with fancy graphics around it might also be harking back to the popup ads of yore, and that it might compel people to click on the only visible button on instinct, their head of firefox did an interview with pc world where he countered this with, and i’m paraphrasing here, “nuh-uh”. this was in reference to an ama they did where several interaction experts weighed in with frankly pretty standard stuff: don’t surprise the user, don’t shove things in their face, don’t draw attention needlessly.

considering they had the worst public fuck-up i think i’ve seen in the last 30 years of gaming, i don’t know if they should continue to be listed as publisher for any games, including their gacha bullshit catalog.

well it’s his state media now so

this condemnation coming after they enabled the link preview popup by default because “it’s not ai” and having a big “activate ai now!!!” button take up half of it, and then doubling down when people rightfully pointed out to them that this is a textbook dark pattern, makes it hypocritical.

me neither. just saying that some people wrongly do, enough to join.

yeah except the wrinkly old fuck is the one on the floor

i honestly think there are people in the party who no longer want that. unfortunately their base hasn’t gotten the message.

except in this case it definitely has, because there have been a lot of so-called rotten eggs in that party.

yeah the expression the party has used, calling them “enskilda rötägg” (literally meaning something like “singular instances of rotten eggs”) doesn’t really translate. interestingly the sweden democrat party members are all mostly working class. they got started as a swedish arm of the BNP so they had basically the same audience, but they’ve cleaned up a lot since the 90’s and unlike the BNP of today no longer say the quiet part out loud.

don’t know if porn depicting an adult pretending to be a minor actually qualifies, but that feels like an extremely thorny topic.

yeah the “one bad apple” party just keeps producing single bad apples

…how do you even manage to screw up that bad?

yeah, those were apparently only popular in the 60’s, 70’s, 80’s, 90’s and early 2000’s. so you could be everything from 28 to 70.

my favourite bit about this sketch is that they start out listing two dishes with no spam in them, which everyone promptly then forgets.

don’t forget repeated!

sounds like you should transition to being an architect

not tried gemma yet, i’ve stayed away from google stuff. maybe i’ll give it a shot.

yeah one of those framework machines with 128GB shared ram would have been amazing. shame they’re sending money to racists.

one of my most recent fun activities came from discovering the “allow editing” button in koboldcpp. since the model is fed the entire conversation so far as its only context, and doesn’t save data between iterations, you can basically re-write its memory on the fly. i knew this before but i’d never though to do it until there was an easy ui option for it, and it turned out to be a lot of fun, because when using a “thinking” model like qwen3.5 you can convince it that it’s bypassing its own censorship.

basically you give the model a prompt to work off of, pause it in the middle of the thinking process, change previous thoughts to something it’s been trained to filter out (like sex or violence or opinions critical of the ccp), and it will start second-guessing itself. sometimes it gets stuck in a loop, sometimes it overcomes the contradiction (at which point you can jump in again and tweak its memory some more) and sometimes it gets tied up in knots trying to prove a negative.

a previous experiment was about feeding stable diffusion images back into itself to see what happens. i was inspired by a talk at 37c3 where they demonstrated model collapse by repeatedly trying to generate the same image as they put in (i think this was how sora worked).