@BlackHC is this really your takeaway from the story? 😄 if it wasn't for GLM-5.2 the infra team would have harder time containing the attack open models also have safety training on top in any case
@mervenoyann But safety training can easily be abliterated in any case, can't it? And yeah that's my takeaway bc we're about six months away from that world iiuc
Negative users challenge assumptions of good intent by AI labs on safety training for open models, citing risks from concentrated power and inadequate governance approaches.
Based on 2 visible X reactions from 2 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@BlackHC the issue is that you assume constant good intent with large AI labs, so I won't discuss further too much concentrated power is a big issue that can't be compensated with an oopsie if the world you want existed then we wouldn't be able to fix such cases
@mervenoyann lol, seriously? have you seen what I've said about AGI governance in general? what exactly is your approach to the risks open models will pose in the near term or what is your mitigation idea to avoid them posing risks in the first place?
@BlackHC is this really your takeaway from the story? 😄 if it wasn't for GLM-5.2 the infra team would have harder time containing the attack open models also have safety training on top in any case
@mervenoyann But safety training can easily be abliterated in any case, can't it? And yeah that's my takeaway bc we're about six months away from that world iiuc
Negative users challenge assumptions of good intent by AI labs on safety training for open models, citing risks from concentrated power and inadequate governance approaches.
Based on 2 visible X reactions from 2 accounts; directional sample.
Ask a question below.
Published answers will appear here.
