downloadram

absolutely wild that HN juuust clips the visibility of the referral code appended to all of your links

well played honestly

canadian tumbler ridge scenario looks different, they dont appear to be fucking about with the LLM when its repeated direct planning type chats across multiple accounts even after a ban. theyre chats that were created for that purpose with far greater implications regarding intent whereas this australian has an ear infection and gets pissed off at the LLM and derails into an incriminating chat. now his name is posted worldwide BEFORE evaluation and his life/reputation is completely wrecked even if its absolved and dismissed. what if? sure but finish the work first then you can publish the name with the findings and evidence if they acually fail eval

i think that argument is a red herring here, wasnt the guy just fucking with the LLM? to me this reads like a bullying version of "my sick grandma" LLM manip

even reading the title you go in with a trama/rage before knowing. brokenmachine's take in this thread is probably spot on

(not GP) i'd be interested in knowing. i dont have any banking apps as i havent needed them yet, but i'd be curious to know if any fraction of these actions are impossible without an app VS going into a branch of your bank and spending time with a rep

tangential and legitimate are not mutually exclusive. the majority of the grievances are infact the latter; they are legitimate

what makes them also the former is that (and surely the reason for the rule) is that because these grievances are so commonly expressed when links are shared they create a near constant 'noise' in the signal to noise of each discussion, with signal being the idea or message intended to be shared by the poster rather than the medium it arrived in

a better way to discuss the tangent or the medium would be by submitting its own HN post. that way theres a better S/N ratio for everything

example: scientific breakthrough is shared on fbook

signal: any discussion about that breakthrough or its Nth order effects on anything

noise: any discussion about the fbook website, walled gardens, zark muckerburg, anything that would be on any link for that site regardless of the message inside the medium

those discussions arent illegitimate, they are better served by targeted discussion on their posts like..

some websites are stylized the point of illegibility

fbook website dark patterns

popups make alt F4

..rather than increasing noise across the whole of HN. this rule is closely related to another site rule: "use the best possible interpretation of the parent comment" because it too is a rule that is in place as an attempt to steer toward "curious discussion" (high S/N ratio)

( i will agree in this case the illegibility warrants a comment chain, holy smokes! )

godspeed

(not parent) their onboarding pushing is frequent, app is fine but web is popup-y with offers/perks

<rant>if i get popups slowing down my web visit it makes me less likely to upsell to, i'd rather a seperate pricing page that i have to manually click to see, my memory serves that most services ive bought have no popup modal (my memory is probably wrong statistically but right emotionally\

maybe op was a/b tested to something more irritating

carla!! of course how could i have forgotten, i tried to figure it out a few years ago when i was still new to linux but got completely lost on the basics

thank you for the resource!

i use EE on my parents tv (minipc connected to offline 4koled) to address the "movie too quiet then way too loud" for them but there is a cost, latency

fortunately its not large enough to notice unless you already know (i.e. problem for me but not for thee) AND fortunately MPV .conf can set audio offset globally to sync everything nicely, so its not bothering them; but i disable it if im going to play games where i need split second audio cue/reaction

the lack of vst support though, seems like a terrible oversight. obs audio filters with stacked vsts and routing filters can unironically do this programs job better because of the vst support, and depending on the setup choices, with better latency too

>she didn’t find a way to tip the AI into “just do it, write it now” mode

worse, the longer an LLM conversation goes on, but especially with constricted/free models (yes the simple chat interface they are likely using) the harder it is to get an LLM into this mode even *IF* you know the right words to say

at that point the best way forward is to terminate the exchange entirely, and to start off with the right initial message, instantly getting into coding mode. a non technical person will not know this and be stuck in feature theory crafting mode in perpetuity, or worse in an endless "excuses' mode as the LLM diverts ant attempt at coding into reasons why its not going to: "i wont output incomplete/broken code! that would require too many lines of code sorry i wont do it! i wont be able to get it perfect so i wont attempt it! but heres more features and theory crafting"

will a non technical person know to end the conversation and start fresh? not likely unless they have a lot of experience already with LLMs

another potential frustration is that HN (algolia) search also includes in the results any submission or comment by usernames/handles that match the keyword. ..so if a frequent commenter / poweruser matches your search term it becomes very messy to sift through unless you script/ublock out their matching sections

depending on the search intent it could potentially be useful, but a built in excluder option would be great for casual searchers that dont want their criterion to match handles

tracker control will be itself blocked by android auto, with a stonewall error DISABLE VPN TO USE ANDROID AUTO

not sure if this was caused by an OS update or an AA update because im certain it used to work fine

(not graphene, but friends otherwise stock samsung android)

i've read helldivers2 kernel component only runs on a normal windows install, you should be able to play on linix via wine/proton without any of that