2 ms·
Plot twist: all these people sharing on twitter yet another creative way of mentioning Xi and Tiananmen in a conversation without triggering the protection (cou
by alexey-salmin 2y ago
Plot twist: all these people sharing on twitter yet another creative way of mentioning Xi and Tiananmen in a conversation without triggering the protection (count to 11 in roman numbers, leetspeak etc) were in fact collecting the training data for the nextgen LLM-based protection. Well played!
- gruez 2y agoPeople had the same theory for chatgpt, except rather than Xi and Tiananmen, it was how to make meth and anti-"woke" topics.
- cscurmudgeon 2y agoA US Tiananmen-comparable example would be ChatGPT censoring George Floyd's death or killing of Native Americans, etc. ChatGPT doesn't censor these topics
- lmm 2y agoHuh? TPTB in the US do not try to censor those topics; if anything they encourage discussion of them (or at least did until this year). US "AI" systems censor much the same topics as US social networks, just as Chinese "AI" systems censor much the same topics as Chinese social networks.
- ta988 2y agoYes, they probably all do that. Anthropic primised to pay the winner that broke all their protections. That way they get tens of thousands of free workers trying to get the money. Much cheaper than $300k engineers.