#dataset
imo if you're an academic you should be downloading every free dataset you could ever want to ever maybe use right now, because LLMs are going to make the cost of large dataset linkage go way down, and that is going to get a lot of these things taken down
November 21, 2026 at 7:16 AM
lol
November 18, 2026 at 9:25 PM
Pangram hasn't even told us the PRECISION of their tools. They have one dataset for false positives and one dataset for false negatives.

There are so many other issues, but the fact that I have to be here spending my time arguing the issues to people who have already assumed that its reliable....
November 18, 2026 at 7:36 PM
It should have been $150k per title
And the dataset should have been destroyed to re-do from scratch without violation.
For all of them.
I can't speak for others, but to think I'd have accepted ~$3k (really, $1500! maybe! I don't know!) to just... give my story to the plagiarism machine for all of eternity is absurd. So what we're getting for this is essentially a scattering of birdseed and a punch in the mouth for our trouble.
November 16, 2026 at 4:53 PM
be the dataset you wish to see in the world 😍
November 19, 2026 at 7:05 PM
In fact, it’s the standard that Bluesky uses now!

Login, communication with users on other social apps, the ability to move your account between hosts, and extensive access to your own data and the public dataset
November 20, 2026 at 12:45 AM
Do you need data about video games? Here you have the most complete database (+110,000)

#gamedevelopment #steam #huggingface #gamedev #indiedev
FronkonGames/steam-games-dataset · Datasets at Hugging Face
Dataset with 124k rows (Image, Tabular, Text). Explore, preview and download it on Hugging Face.
huggingface.co
November 20, 2026 at 5:41 PM
"A global gridded dataset of unified onshore and offshore wind resources and power potential" www.nature.com/articles/s41...
A global gridded dataset of unified onshore and offshore wind resources and power potential - Scientific Data
Scientific Data - A global gridded dataset of unified onshore and offshore wind resources and power potential
www.nature.com
November 20, 2026 at 8:15 AM
Have you ever found PII in a dataset that was supposed to be clean?
November 22, 2026 at 3:04 AM
looking through an rna-sequencing dataset on this website called transxplorer and i'm so tickled by the name
November 20, 2026 at 3:11 PM
This dataset (mostly arrivals into the US, South Korea and Japan) is more of a random global sample, as it is not skewed by regional sequencing volumes.
🧵
November 21, 2026 at 1:14 AM
At some point, someone is going to have to launch an investigation into how:

1. SpaceX came to dominate the NASA launch market

2. How NASA did not recognise this trend, and manage / mitigate the risk that is very clearly demonstrated here (diversification risk management is high school economics)
November 19, 2026 at 10:26 PM
I also knew that publishing a dataset of machine-generated bird illustrations would put people out of work.

People would look at the hourly rate of an illustrator and then at this free dataset of machine-drawn birds and they'd make the easy choice. Or their boss would make the choice for them.
November 19, 2026 at 12:23 PM
Besides all of us who fell through the cracks when our books were listed in the big dataset but weren't in the lawsuit database and didn't even get this much and only got rage to show for it.
November 16, 2026 at 1:30 PM
Historicizing colonial Havana.
#slaveryarchive
November 21, 2026 at 10:14 PM
I also bet it's only prompt controlled not a fine tune much less a kid safe dataset.
November 17, 2026 at 6:36 PM
I didn't run these numbers myself (so caveat emptor on whether the underlying dataset is reliable), but Claude claims to pull these off Shiller's monthly returns dataset.

I am not sure how well people would deal with a redux of this (you dont have to cherry pick endpoints). Let's not even bring up
November 21, 2026 at 1:18 AM
This could probably be fought against by limiting the dataset to races that pre-election measures indicate will be close, but it's a constant cat-and-mouse game.
November 21, 2026 at 4:50 PM
And this being a public dataset that we own, it should just be available for download in its entirety if people want to scrape it. It’s public data, scrape away!
November 21, 2026 at 3:46 PM
holy

incredible dataset for the needlessly aggressive/rude classifier
November 16, 2026 at 11:07 PM
Try to not use a cherry picked dataset from 12 years ago
November 18, 2026 at 1:00 PM
i’ll send you the preliminary findings from the latest dataset before it goes to peer review 🫡
November 20, 2026 at 7:41 PM
A very nice way to celebrate #fluorescencefriday! Already having fun with with the 3D segmentation:

Dataset 🐟: www.ebi.ac.uk/biostudies/B...
Model used: hvit_b_cells
Animation: adapted from napari.org/napari-anima...

Congratulations @anwaiarchit.bsky.social @cppape.bsky.social and team!
November 20, 2026 at 12:41 PM
🤔 What kinds of higher education institutions are engaging in research assessment reform, and which remain underrepresented?

🙌 REALISE takes the first steps towards answering these questions by publishing its first dataset.

‼️ Read more about the dataset on our website: lnkd.in/gjU_cJTC
November 20, 2026 at 7:53 AM
this dataset is incomplete, you can help by expanding it
November 17, 2026 at 7:23 PM