pull down to refresh

108 sats \ 2 replies \ @SimpleStacker 31 Aug \ on: GPU World AI ideasfromtheedge
LLM use is permitted, but discouraged; we remind participants that LLM use tends to reduce originality and writing quality, and the flaws are especially obvious when LLM outputs are read as a group---unskillful use of LLMs will reduce the odds of being the best entry. We request disclosure of AI use.

IMO, AI disclosure requests are themselves a bit of a minefield to police. What constitutes use? If I use an AI to just bandy ideas about, but don't copy and paste any of its text, is that AI use? What if I use it as a thesaurus to look up a better word choice?

Overall though, I agree that you can't really police AI use and in the end what really matters is the end-result quality.

write
compose
97 sats \ 0 replies \ @freetx 31 Aug
If I use an AI to just bandy ideas about, but don't copy and paste any of its text, is that AI use? What if I use it as a thesaurus to look up a better word choice?

Probably the best use of LLM in this case would be to help formulate a premise that you know will resonate with Neal Stephenson.

Having read some of his stuff, I think a premise like this would work well:

Local LLM use becomes broadly a 2nd Amendment style issue....nearly everyone in future uses LLMs for personal/family safety both in physical sense and also to combat governmental / big-corp data harvesting. Tied in with this, local LLMs now generate most of the media and entertainment that people consume, hyper-individualistic styles like Beatles+Space Jazz+Baroque+Gregorian Chant.

Likewise totally new "dynamic entertainment" platforms exist that are generated on-the-fly by your local LLM. So instead of going to youtube and watching a pre-recorded video. You scroll sites like "YouPrompt" which are collections of constantly updated prompts (tailored to your history) that your local LLM then generates into a videos on-the-fly.

reply
48 sats \ 0 replies \ @Scoresby OP 31 Aug

I will be curious to see if they just get overwhelmed with submissions. If they do, I imagine they will use AI to sort, categorize or rank them in some manner...or they'll simply resort to old fashioned human reputation (find submissions from humans who are known to be serious writers and only read those).

I clicked through the submission form and it doesn't seem to have any really strong mechanism for preventing the tidal wave of slop in submissions. I really hope they write something about the process of dealing with submissions.

reply