The whole issue with LLM models right now is that they are notoriously difficult to control. If it were up to Google, every single response would include a reference to some sponsored content - but they can't do that without completely destroying the usefulness of the output and besmirching the sponsor's brand.
Of course, as time passes, we're going to refine this technology, until we have enough control to implement these terrible/profitable ideas. Like any aspect of life under capitalism, we can really only enjoy it while it lasts.
I hope at that point we have enough capable alternatives. Like, hopefully around the time they add ads is also the time when open-source models and apps have caught up again.
It’s about unlearning empathy. About learning to other and put up walls of concern.
I saw parents bring their children through a seafood shop that had live crabs tied up in a basket, and the parents corral them away from the crabs and direct them to not care about them. It was training. Learning to not care about obvious suffering, strengthening one muscle annd atrophying another.
AFAICT, mastodon's decisions, which are arguably problematic (on which see: https://lemmy.ml/post/14973403) are literally trickling down to other platforms and infecting how they federate with each other as they dance around mastodon's quirks in different ways.
It seems like masto is ruining "the standard" with its gravity.
None of that matters if Mastodon doesnt implement these suggestions or standards. And from past experience its extremely unlikely that they will. Thats why I think its best to ignore what Mastodon does, its not our concern how they decide to render things.
That's kind of what I meant too, if there's a standardised and correct way to implement things, that's how projects should implement it instead of trying to do it the "Mastodon" way
The problem is that Misskey code is bad. This is the main reason the forks exist at all.
Iceshrimp is rewriting from scratch in C# and it's my main hope for the *keys.
So I just learnt that J Physiol requires figures to be made with Biorender (https://www.biorender.com/), a VC backed subscription SaaS extend and extinguish of scientists drawing pictures!
Is this a thing now!? Complete privatisation of the publication workflow!?
Am I the only one enraged by this!? Pictures? We could have just made our own shared repository of useful graphics. Our own open source software. Uggghhh!
Bizarre, yes, but as I said in the original thread, it seems well designed to trick researchers into thinking that it’s a standard that they may as well just use like MS Word.
A relatively tech illiterate lab leader will often just tell their students to use it in the same way they do everything else to please the journals. Because as far as they know illustration already works like documents and MS Word.
@maegul @NicoleCRust@academicchatter
I find all the bad PowerPoint illustrations in biology charming, but I personally take great pride in my illustrations in my work (probably wont be doing biology for the forseeable future) and it makes me sad when people dont, whether that be because they are pressed for time or bc they dont see communicating ideas visually as part of the task of science.
I have tried to introduce diagramming markup like mermaid or graphviz/dot ti my lab to limited success, they are just used to PowerPoint I guess. My cracked copy of illustrator 2019 is basically always open. Turning figure design into some drag and drop biorender task (in addition to the platform capture element as u say) is sad to me bc it feels like the final flattening of illustration as a proud tradition in biology.
