The Nexus of Discussions

Yep. People running fedi instances, just like people running Bluesky, have to decide whether they want to take that risk -- and how much geoblocking decreases the risk.The tradeoffs may well be somewhat different here but it's still very messy, especially for fedi instances either based in the US or with financial ties here (like Mastodon's 501(c)(3)). Bluesky's high-profile, so they might be a more tempting target; then again they fedi isntances have a lot less resources to fund a legal defense, so maybe we're more tempting targets? There's no way to know at this point.@jsit @Gargron

@ophiocephalic great pointsl. @rwg and @rra 's https://www.cell.com/patterns/fulltext/S2666-3899(23)00323-9 has best practices for researches on fedi, and goes into detail on the expectations of privacy and history of poor research.(Of course, fedi is not monolithic, and the norms on Lemmy quite possibly do assume consent. But it's more complex than that: Mastodon posts, from peoiple who aren't consuming consent, can federate to Lemmy, so get swept up in this.)Also worth mentioning: Google's Perspective API also has known racial biases - see https://civic.mit.edu/index.html%3Fp=2402.html for example.

@rancidrabbit good point, thanks. In practice both are potential vectors (as is RSS), so useful to lock them all down.

@fellofreak and neither are elders with cognitive issues - https://www.reuters.com/investigates/special-report/meta-ai-chatbot-death/

it's a dirty job but somebody's got to do it. the fireproof undergarments industry is critical infrastructure!@thegibson a

Yeah totally agree. And Friendica / Hubzilla's multi-protocol approach is the future as well as the past!@quillmatiq @gabboman

Blacksky has been hosting their own relay for several months now, it really isn’t that expensive. And they’re developing their own AppView. They’ve spent a lot of time investigating and building stuff on the AT Protocol, and clearly don’t think it’s madness to use it as a base - in fact one of Rudy’s early posts talked about why he saw it as a better match for his community-focused goals than ActivityPub.@frumble @mastodonmigration

Jon-Pincus hey check out that topic loading from within this NodeBB cc admin@discussions.thenexus.today

No other community deserves whatever is happening more than the research community that brought this on us. What you unleashed coming back to eat its own. "We note that in the extreme case where an LLM might be used to produce an entire piece of research, we still require a human author for accountability."What's the point of submitting papers or singing up as a reviewer at these conferences? Why don't we just have LLM generated "papers" be met with LLM generated "reviews" and call it a day?

If there are people interested in grassroots privacy organizing in Oregon, happy to connect!

Details, details.@mcc @deadsuperhero

"We are planning to publish a feature explainer detailing our decisions (for example, how to let users control if they want to be included in a Starter Pack) in the coming weeks ..." -- and really glad to hear you're thinking about that! Bluesky's failure to provide that control has really been a problem.FYI @onepict -- Mastodon thinking about consent, gotta like that!@MastodonEngineering

No joy on the attempted crosspost to #SocialHub. Oh well. Here's a bit more context

Specifically with Meta, here's a post from @cuchaz with some tools to set up firewall-level blocking. https://gladtech.social/@cuchaz/115004304985099620Also, Meta's user agents include Meta-ExternalFetcher and Meta-ExternalAgent (which is the one they say they use for AI training). https://www.businessinsider.com/meta-web-crawler-bots-robots-txt-ai-2024-8And for instances that use a separate domain for media, make sure that you've got your robots.txt and firewall blocks in place for that instance as well -- I've heard from a couple of admins who recently realized they didn't.@sw_isac

@thenexusofprivacy @JenniOlsonSF Remeber: Linkedin is Microsoft.

@quillmatiq must have been a day that ends in Y

@clarinette unfortunately the majority of US policymakers are in the "censorship good" camp at this point, so I'm not sure how helpful the article's framing that they need to "take heed" is.