20 ms·

Framing & launching LLMs as a "chat" interface is the source of many ills. I don't have a simple solution, but leaning away from conversational interfaces would

Framing & launching LLMs as a "chat" interface is the source of many ills. I don't have a simple solution, but leaning away from conversational interfaces would lead to less anthropomorphizing.

Terr_ 6mo ago
I feel "document generator" is the best and most-grounded framing. Right now, lots of people get caught in a trap of asking things like "does MyFreeBestFriendAI feel remorse?" If we're already looking at it as a document artifact, we can evade the implied-ego trap: The document generator took a chat-like document where two fictional characters are talking, and predicted that there would be text where one fictional character is associated with apologetic words.
xg15 6mo ago
One aspect is that chat LLMs are trained to talk like a person - "If I should do X, just say the word" etc. It would be interesting to train an LLM that consistently "talks like a computer" or a command line utility instead, i.e. passive sentences, relatively bare results of the tasks given, no reference to a self, etc.
- Terr_ 6mo ago
  There's a webcomic where the author wrote some backstory before LLMs, where the world over-trusted in statistical models until it went wrong in a ghastly "stupid Skynet" way. This led to international requirements that all "AI" have explainable/auditable decisions, and must have clear non-human features.
  - xg15 6mo ago
    Do you have a link? That sounds interesting (and might be a genuinely good idea, especially with the explainable or auditable decision making)
    - Terr_ 6mo ago
      Here's the strip [0] where the author commentary goes into it. In the niche of long sci-fi comedy webcomics with worldbuilding, I suggest Schlock Mercenary [1]. Early art is rough, but it gets consistently better. [1] https://forwardcomic.com/archive.php?num=200 https://forwardcomic.com/archive.php?num=200 [0] https://www.schlockmercenary.com/ https://www.schlockmercenary.com/
      - xg15 6mo ago
        That looks interesting. Thanks a lot!

Framing & launching LLMs as a "chat" interface is the source of many ills. I don't have a simple solution, but leaning away from conversational interfaces would lead to less anthropomorphizing.

Terr_ 6mo ago

I feel "document generator" is the best and most-grounded framing. Right now, lots of people get caught in a trap of asking things like "does MyFreeBestFriendAI feel remorse?" If we're already looking at it as a document artifact, we can evade the implied-ego trap: The document generator took a chat-like document where two fictional characters are talking, and predicted that there would be text where one fictional character is associated with apologetic words.

xg15 6mo ago

One aspect is that chat LLMs are trained to talk like a person - "If I should do X, just say the word" etc. It would be interesting to train an LLM that consistently "talks like a computer" or a command line utility instead, i.e. passive sentences, relatively bare results of the tasks given, no reference to a self, etc.

Terr_ 6mo ago
There's a webcomic where the author wrote some backstory before LLMs, where the world over-trusted in statistical models until it went wrong in a ghastly "stupid Skynet" way. This led to international requirements that all "AI" have explainable/auditable decisions, and must have clear non-human features.
- xg15 6mo ago
  Do you have a link? That sounds interesting (and might be a genuinely good idea, especially with the explainable or auditable decision making)
  - Terr_ 6mo ago
    Here's the strip [0] where the author commentary goes into it. In the niche of long sci-fi comedy webcomics with worldbuilding, I suggest Schlock Mercenary [1]. Early art is rough, but it gets consistently better. [1] https://forwardcomic.com/archive.php?num=200 https://forwardcomic.com/archive.php?num=200 [0] https://www.schlockmercenary.com/ https://www.schlockmercenary.com/
    - xg15 6mo ago
      That looks interesting. Thanks a lot!