Tessera
Folgen

Robert Gay

@paleorob.bsky.social

Educator and paleontologist. Here for science communication. Personal account, all postings are my own and do not reflect the views of any employer. Most recent pub: https://giw.utahgeology.org/giw/index.php/GIW/article/view/82

131 Posts380 Follower320 Folgt
Clarkesworld@clarkesworldmagazine.com · 6mo ago

Today is the 3rd slopiversary of us having to temporarily close submissions. The problem of "AI"-generated submissions hasn't gone away and is still a huge waste of our time. House policy remains the same: #noAI stories, translation, art, or anything else.

Post from twitter on our @clarkesworld account from 11:48AM on Feb 20, 2023. "Submissions are currently closed. It shouldn't be hard to guess why."
Robert Gay@paleorob.bsky.social · 6mo ago

I think the most wtf moment of #WAVP was the talk suggesting that Troodon was an obligate herbivore based on eggshell nitrogen and nothing else. The same talk implied that Maiasaura was specializing in eating rotten conifer logs when gravid. Definitely puzzling but now I want that Maiasaura paleoart

Dr Rebecca Wragg Sykes@lemoustier.bsky.social · 6mo ago

🧪🏺 Update - authors have new paper showing how useless gen- #AI is for archaeological illustration. All 400 images were multiply inaccurate (physically, socially, technologically, environmentally), even with improved prompts. JUST USE HUMAN EXPERTS & ARTISTS www.cambridge.org/core/journal...

screengran of figure from paper showing four different generated images of a "day in the life of a Neanderthal". All images contain significant inaccuracies from physical characteristics (extremely primitive) to material culture (glass containers), and the lack of women and children is also obvious.
Dr Rebecca Wragg Sykes@lemoustier.bsky.social · 8mo ago

e.g. to produce these two – which are still problematic – took at least 80 batches of DALL-E 2 generation (each batch making multiple images), and on top of that, required direct editing of the #AI images themselves. Paper: Magnani & Clindaniel 2023 🏺🧪 www.cambridge.org/core/journal...

Dr. Jens Foell@jensfoell.de · 6mo ago

Using AI makes science communication worse. Period. AI use necessarily reduces accuracy, transparency and authenticity. But sci comm *needs* to be accurate, transparent, and authentic. Otherwise you’re just goofing around, which will destroy your audience‘s trust in you and in other sci comm.

Dr Rebecca Wragg Sykes@lemoustier.bsky.social · 6mo ago

🧪🏺 Update - authors have new paper showing how useless gen- #AI is for archaeological illustration. All 400 images were multiply inaccurate (physically, socially, technologically, environmentally), even with improved prompts. JUST USE HUMAN EXPERTS & ARTISTS www.cambridge.org/core/journal...

Dr Manu Saunders FRES@manusaunders.bsky.social · 6mo ago

On this day that supports inclusive futures for women and girls in science, let's not forget all the women that were sidelined, ignored, or forced out of science because they: - had children - prioritised family/caring - did science differently - questioned boundaries & norms - fought the patriarchy