Discussion
It's very important to call out, criticize and push out LLM usage for creative expression (writing, art, music, code).
But we also should not make "hunting LLM usage" our identity because it can also bind so much energy we could invest in building something better.
(I am in the target audience for this post)
@tante Hard agree. I am not the AI police. Everyone is free to do as they please in that regard — they just lose my respect and I will not touch any of their work with a 10ft pole.
@tante
Serious question: Are you including "asking if LLMs were used in a product you're using/interested in and refusing to engage if so" with "hunting for AI use"? Because in the context of last week's discourse it seems like yes.
If so, I'm not sure how you square that with "we should push out LLM use".
Obviously, seeking out projects you have no use for/ interest in to check for LLM usage isn't constructive (ideally that'd be a government job) but this messaging seems to conflate the two.
@tante how about teaching people ways to express themselves creatively and how LLM can be a tool with potential to both support but also limit that process. It's a technology with a narrative and currently big corp writes it. Instead it needs to be a common good whose uses, harms and limits need to be learned and educated about. By making #NoAI an identity we will neither prevent nor improve that technology.
@randogoth I don't see "AI" as just a narrative problem, "open source community AI" (based on similar logics and structures as the one we see from Anthoic/OpenAI etc) would also be just as bad. I went into my reasoning a bit here:
@tante AI is as fascist as 'the internet' being militaristic war technology. It depends on who writes the narrative.
@tante I read the experience of an indie game developer that, first, researched a problem, then contacted the original authors of their game engine, before finally consulting AI and arriving at a 10 lines-of-code fix to their problem, and you know what, I can't blame them. Software developers failed them and the LLM didn't.
Us software developers really haven't been doing ourselves any favours the last two decades.
@tante This is a great reminder. I had to tweak some of my follows on here because my feed was starting to lean heavily into that second mindset. I think I struck a healthier balance.
@tante look you're not alone and that scrutiny side of things is needed too
@tante I can't directly verify this, but there may be measures taking place to further develop the Tor network, to establish LLM's to decipher the most abusive application of Tech Bros AI, and potentially develop further countermeasures.
@tante Agreed. The only models I am using are machine learning models on-device that don't require access to the internet, and even then I am only using them to crunch massive data sets to find regression trends to then use when plotting user selections to return a function action.
LLMs have no place in doing things that I should be able to code myself. #noAI