💾 Archived View for bbs.geminispace.org › u › requiem › 19036 captured on 2024-08-31 at 15:51:44. Gemini links have been rewritten to link to archived content
⬅️ Previous capture (2024-08-19)
-=-=-=-=-=-=-
Re: "An Idea about having public Gemini llm"
Personally I want to keep AI off Gemini as far as possible. The inaccuracy and the content theft are huge problems and actually I don't want machine generated random nonsense simply because that'd mean more bytes of data floating around Gemini. It'd be extra bytes but it wouldn't be "content" and I would certainly never refer to the insanity generated by llm's.
Aug 09 · 3 weeks ago
@requiem agreed. I want a lot of content on Gemini, but it should be valuable content, and most AI generated content has very little value.
I don't know. For adding Q/A style to content it would have little value unless the output was fact checked. Another problem is the relatively small input afforded by a Gemini request; you'd basically have to support Titan for prompt submission for a lot of use cases such as requesting a document summary. Plus, I personally wouldn't want to be the one footing the bill for a VPS with decent GPU or the electricity if selfhosting at home.
This would not add anything useful or valuable
Just leaving this recent news article here about the family who ended up poisoning themselves because unbeknownst to them the mushroom identification handbook they bought online was in fact LLM generated garbage. AI based “knowledge bases” are actually life threatening:
Sorry I should have said “post” rather than news article. It’s a reddit post and not a news article, my bad.
🚀 random2934 [OP] · Aug 21 at 06:09:
@requiem that's really sad I didn't know that. I only knew about Google's Gemini telling people to put glue onto their pizza
An Idea about having public Gemini llm