Notices by picofarad (picofarad@noauthority.social), page 3

ohhhh, the loud support for VP Harris is *virtue signalling*

now it all makes sense.

@icedquinn 7/10 times the UPS lets me ride temporary breaker trips upstream (like when it does the brownout 3 times before a blackout because it's trying to nuke a tree branch or something). My triple X platinum PSU on this computer can usually ride a quick brownout (less than 2 seconds), but i don't trust spinny drives and network gear to not s the bed, so they all get UPS.

If i was "gaming" when this PSU would normally save me from a brownout it would just shut straight off, tho.

lightning!

anyone remember the name of music sharing software for windows, where you'd join a server/room and you'd tell the software what folder you wanted to share, and it just worked? you could search across all "rooms" you were in, and there was a log window where it showed you all the searches hitting your computer.

this was like 2-4 years ago i used it, it was not: kazaa, limewire, emule, edonkey, XDCC, etc.

It's on a virtual HDD on a dead machine so it's a PITA to get to, so crowd sores!

Item 1 China banned most gacha game mechanics including randomized loot lootboxes and incentives such as daily login bonuses and "buy 5 get 50% off." Investors and representatives from the two largest Chinese game companies said this would cost them $80,000,000,000
Item 2 lawsuit that began in 2005 finally in court, involving Visa and Mastercard. A judge rejected the deal offered. V & M, in 2023, from terminal fees alone, took $60,000,000,000 to the bank.

@colonelj @bot
having rescued a few animals in this sort of condition, it is a good feeling when they finally deprogram whatever happened to them and become part of the family.

chatgpt got it

I didn't even know it could handle images on my version, huh

@icedquinn i'll go demo those services by putting a few sentences in that will immediately make it obvious if they're to blame.

@icedquinn I don't like anything AI that anyone else creates, and maybe 0.1% of stuff i've "run off" on AI is postable. This is to say i'm extremely critical of it, and most people publish garbage (i mean, anyhow, but moreso of genAI stuff).

sonauto is another one

"@thereiruinedit" on youtube et al runs his own genai stuff and the secret sauce is very strong. If it were mainstream there wouldn't be any reason for record labels.

@icedquinn what service generates these? the services i've tried out only output like 40 seconds, and they've got weird quirky glitches. I'm wondering if the people who make/release these just run off 100 and pick the best one, the same way i do with stable diffusion

here's an example from udio

"content creators" that use dubbing/voiceover software bug me so much.

not once in my life have i ever wanted to hear a computer or robot read to me for 20+ minutes.

I am unsure what software they use as i only know of 3 TTS softwares, and the quirks that are common in those three aren't present. I'm wondering if capcut does it or there might be a paid service somewhere using something like tortoise that has been customized.

anyhow, i just watched 10 minutes of a video and i'm annoyed.

@PNS isn't the loophole "in order to 'work' most of the 'therapeutics' have to be done before puberty because [reasons]"?

When did the CAST thing come out, because i don't have it yet, i got sidetracked by that other huge medical report put out in the UK last month

@kirby this is why i set all my malware to do nothing for at least 280 seconds

@icedquinn there was some commentary that mixtral (with an x) may be better since it *may* have the full 32k context window and not require moving window batching.

I've never tried summarizing, but i wonder how hard it would be with this one corpus i have, maybe split it on a phone number or something and summarize each section.

@icedquinn yes, i hope someday they release better models but for now it's have some errors and a searchable corpus or zero errors and no searchable corpus.

diarization hasn't worked for me, yet. I'm willing to try anything once, although using docker but also requiring .sh runnery is almost a non-starter.

https://github.com/transcriptionstream/transcriptionstream

hmmm, will have to see

@icedquinn I'd feel so bad for whoever edited it; if i edited it i'd never publish, as is evidenced by there not being any podcast episodes already.

We used to record our hour long conversations on the phone and since we did that twice a day, 4-5 days a week, we had enough content to do a podcast, but ugh, my voice.

i've kicked around the idea of doing a podcast with a friend or two. I don't like my voice much - there's a few videos with my voice in them on youtube (including ones uploaded by other people) and when i'm in "concentration" mode i repeat "ticks"; one video i am in was edited to cut out uninteresting parts, so the other person would ask me "filming?" and i'd go "yuh yuh yuh" which i learned from a marine i was friends with.

ugh.

Maybe i can record the podcast, transcribe, then TTS it?

@sun Grand Internet Mersenne Prime Search?

okay, but why

@r000t like a mastodon relay or nostr?

or, and i hesitate to ask, an x.25 packet relay, or APRS...