ai search is insult to the injury of them taking away commands and filters

an IT exec complaining about Indians probably doesn't mean native americans, there's not a ton of us in tech, let alone compared to the Indian expats in the country or the big tech offices they have back home

if palpatine can come back, cinta can too

we've had almost twenty years to match what dji is doing, and this is what they come up with?

the propaganda machine on this website is running pretty hot most days. and the loudest voices and the biggest trends are pushing a lot of misinformation

i'm partly hoping for the bubble to pop so that the credulous trolls will drop their moral crusading

bump stocks are old news, did this legislation also cover forced reset triggers?

next year in particular with the election year in the u.s., all those grainy infographics of dubious origin are getting an "upgrade" for sure

again with the capitulation journalism. it's the ACA.

the "within 30 days" part makes me wonder/hope he's planning on leaving the country for the holidays and never coming back

the real crime is using the app's light mode with the brightness way up like that

also is that first picture an AI hallucination or am i just seeing too many hands and ankles?

whatever feed he has open seems kind of cursed

but the general idea of "repeat an approximation of what you've seen before" fits both stochastic parroting and style theft. i thought that would be obvious as well.

i do know what the term means and i know it was coined to be inflammatory, but not terribly accurate. the whole point of LLMs *is* to encode meaning and mimic the process of "understanding" meaning

i hope it's obvious that i'm NOT saying these systems "think" or "know" or "experience"

which i'd agree is still not outright "theft", but it's pretty obviously IP/copyright violations, etc... sometimes i feel the whole internet is repeating history and old conversations about piracy

i will add though, that there are some red-teamers out there that have managed to make some models reproduce whole passages verbatim. some of the models are definitely capable of and vulnerable to doing full-on plagiarism