3 ms·

I don’t recognize the described risk. If spam gets so advanced that individual articles are indistinguishable from human-produced material, reputation steps in,

I don’t recognize the described risk. If spam gets so advanced that individual articles are indistinguishable from human-produced material, reputation steps in, forcing much more ability from the bots to keep up with human content producers. If language models or their descendants are able to overcome this barrier, then it doesn’t really matter who produces the content.

mdale 4y ago
We already live in a ecosystem flooded with human "human level" spammers at a mass scale. We are only grounded in systems of reputation & relationships. These systems will remain important as content becomes fully synthetic. I.e Linkedin, Google, Social networks already heavily depend on social and reputation graphs towards keeping spam at bay; if anything new ideas will be harder and harder to go viral as the social and reputational filters are forced to assume synthetic. Maybe AI learns to leverage/buy endorsement "content" posts from your friends and colleagues like celebrity endorsements operate today (as personalized commercials/text become very low cost). AI/people trying to drive value with AI will have to find ways to leverage social relationships to squeeze into relevance where content/ideas are very low cost.
mef 4y ago
the issue is that the large language model doesn’t understand what it’s saying, cannot reason or come up with novel ideas, yet it can convince a human that it _is_ doing so. so then if the world begins to prefer AI-generated content, any question you ask the internet will only show you AI-generated answers, which can only offer you answers based on its training set, and over time a system with no new inputs just ends up being static generated from static, albeit static which is convincing enough for a human to accept it.
- 20after4 4y ago
  I feel like we've already reached this point, or nearly so.

mdale 4y ago

We already live in a ecosystem flooded with human "human level" spammers at a mass scale. We are only grounded in systems of reputation & relationships. These systems will remain important as content becomes fully synthetic. I.e Linkedin, Google, Social networks already heavily depend on social and reputation graphs towards keeping spam at bay; if anything new ideas will be harder and harder to go viral as the social and reputational filters are forced to assume synthetic. Maybe AI learns to leverage/buy endorsement "content" posts from your friends and colleagues like celebrity endorsements operate today (as personalized commercials/text become very low cost). AI/people trying to drive value with AI will have to find ways to leverage social relationships to squeeze into relevance where content/ideas are very low cost.

mef 4y ago

the issue is that the large language model doesn’t understand what it’s saying, cannot reason or come up with novel ideas, yet it can convince a human that it _is_ doing so. so then if the world begins to prefer AI-generated content, any question you ask the internet will only show you AI-generated answers, which can only offer you answers based on its training set, and over time a system with no new inputs just ends up being static generated from static, albeit static which is convincing enough for a human to accept it.

20after4 4y ago
I feel like we've already reached this point, or nearly so.