Mastodawn

Anuj Ahooja 10h ago

In the last five years, we've gone from "employees will never have to go into an office" to "employees need to be in the office because creative and innovative work can only be done face-to-face between humans" to "lol we don't need humans"

Evan Will 8h ago

@zackbatist Anyway, I spent a bunch of time figuring out how to turn off copilot auto suggestions in a bunch of different contexts, because it is just too distracting, constantly takes my mind away from the task at hand.
They make it way too hard to have copilot on ONLY when you ask for it. VS Code just keeps adding more places for it to "helpfully" popup and distract you

Evan Will 8h ago

@zackbatist I have gotten copilot suggesting new rows in a CSV of digital collection metadata. The suggestions are very interesting, but clearly missing the context of describing an object that you have digitized! I guess next step it will auto generate objects for the collection at the same time...

#CAQDAS #QualitativeDataAnalysis

Zack Batist 8h ago

So I'm using a CLI-based qualitative data analysis system (it's great! https://qualitative-coding.readthedocs.io) that I integrate into my VSCode workspace. I happened to start experimenting with copilot last week and today when I started qualitative coding I noticed it was offering me suggestions. Seems to be based on my prior codings rather than the actual file from the corpus, but still a bit eerie and not sure if I want to keep this on or not.

Andreas Wagner 10h ago

@kfitz Congrats, well deserved!!

Andreas Wagner 10h ago

Andreas Püttmann 10h ago

Ein Symbol für die freie Entfaltung der Persönlichkeit (Artikel 2 GG), für die Menschenwürde (Art. 1) und für gleiche Rechte gegen Diskriminierung (Art. 3) kann ja auch schwerlich die Neutralität des Staates des Grundgesetzes verletzen.

RE: https://bsky.app/profile/did:plc:x7vfp7rt4gbywara7ljm5e2s/post/3lsjblyjnqk2n

Andreas Wagner 15h ago

Antonin Delpeuch May 15

I'm between jobs and open for short-term missions until the end of the year :) . I like jumping in old #opensource code bases, tackling tech debt, working on #governance, and trying out new things! Get in touch :)

Lately I've mostly been working in #Java and #Rust. Based in #Leipzig, open for in-person or remote freelancing gigs.

Andreas Wagner 15h ago

Mathieu Jacomy 16h ago

In that sense, LLMs will abide by one view or another depending on how you prompt them.

But it's another thing entirely to pretend that inconsistent world views are self-consistent just because authoritative figures pretend they are. If some political discourse is incoherent, don't blame a model for missing the coherence it lacks.

The King is naked, and it's not the tool's fault if it can't measure the thickness of his clothes.

9/9

seems to imply that the liability in that case would rest with Hugging Face or whatever service runs the model / uses it to infringe. Having created a model that *could* infringe is not actionable here. Finally, not a practical point, but this analogy will piss ppl off and I'm glad.

Oh, and if you want to consult the original, see below. Glad this did not settle out of court. That would have really screwed over academic researchers and libraries without big legal staffs. bsky.app/profile/eriq...

RE: https://bsky.app/profile/did:plc:i6w2uuipovifyhfu62viemv7/post/3lse3ncj4qk2q

My overall reading: if this is the law, we can feel confident that 1) libraries that physically possess copyrighted books 2) can let researchers train models on them 3) without a lot of fuss about exactly what % of any book a model would be able to reproduce 4) and even share those models! +

All of that, by the way, before we even address the additional factor that the intended research use is not-for-profit. It's a slam-dunk — and unfortunately these things have to be a slam-dunk when you're an underfunded institution worried about lawsuits.