3 ms·
> It's ultimately a gradient It's actually not. In global politics actors rarely engage in things their peers don't. It seems to me that this is rather ultimat
by foverzar 5y ago
> It's ultimately a gradient
It's actually not. In global politics actors rarely engage in things their peers don't. It seems to me that this is rather ultimately something more like finding the most wildly looking thing about how other culture manages information flows and PR and focusing optics on it.
> I honestly worry that with the development around language models like GPT-3, the internet as we know it will be completely blanketed with astro-turfed content by language models and low-paid contractors.
Tbh I think that we had already passed this point, with all the low-grade SEO and myriads of copy-paste clickbait internet-tabloids. The internet has already drowned in low-grade media-content and language models would hardly change anything fundamentally in that matter.
- dirtyid 5y ago>focusing optics on it. Basically. Western reporting trying to overindex PRC influence campaigns as part of manufacturing consent has gotten so bad that useful idiots in the west start to reflexively accuse 50C everywhere. And brainwashed enough to genuinely believe it. Official 50C operates domestically, not abroad. Even then, 50c posting is characterized by spamming low effort platitudes instead of high effort engagement. PRC isn't wasting valuable human capita with English proficiency to argue/troll on platforms like HN or Reddit. Millions sympathetic to PRC opinions like diaspora / VPN users do that on their own free time. To date, PRC influence campaigns are limited to a hundreds / low thousands of accounts boosting discourse of official accounts on Twitter and Facebook. So far, they've been low effort campaigns with poor reach. But repetitive headlines tries to insinuate otherwise. PRC will probably harness GPT-3 to exploit asymmetry in censorship ability. They just haven't yet.