Yes.

@U2FRKM4TW Are you opting out?

despite my well-voiced skepticism of LLMs on this slack, I’m generally ok with useful Clojure information on this slack team being used for training data. My concern is over the results: it would be nice if we could benefit in some way from the embeddings or models produced from our data so we can build Clojure-specific language applications. I doubt we’ll have that ability, realistically.

Dunno. Maybe @U04V70XH6 has sent that email? Just in case someone doesn't know - all (or most?) of the public data on this server is publicly available online, without having to sight up. And also, Slack doesn't feed the data on its servers to some LLMs as-is: > When developing AI/ML models or otherwise analyzing Customer Data, Slack can’t access the underlying content. We have various technical measures preventing this from occurring. Please read our Security White Paper for more info on these controls that protect the confidentiality and security of Customer Data.

I have not opted us out because the topic is still being discussed. The policy has been in place since September last year so it's nothing new. It's just coming to some people's attention now because of some noise on Twitter about it.

As noted above, nearly all the content here is already publicly logged on the ClojureVerse site and that's linked from various places so I expect public LLMs have already vacuumed this content up. As also noted above, Slack's own LLM models are respecting per-workspace privacy. So, bottom line, nothing has changed and I think it's a bunch of noise about nothing much.

I think that, had I been in the admin team, I would think of erring to the side of more privacy, in respect of, well, people's privacy. Even if not everyone in the workspace cared about that or some are even pro-sharing, it seems more unfair (to me) to opt those in that have no choice in the matter of privacy than to opt those out who might get some vague, future advantage from some (presumably commercial) LLM product. No offense meant 🤓

Good point about the publicly distributed Clojureverse copy, of course. Even so, one could opt out out of principle, and have one's cake of supporting LLM (indirectly) and eat the pro-privacy cake, too 😁

@U0AQ3HP9U Well, a channel here can either opt in to logging to ClojureVerse or not. Individuals here have no such option to opt in or out of anything, because that's not how Slack works. And as noted, this has been Slack's policy for eight months and these "privacy" concerns -- which Slack address in the linked privacy document -- have only just been raised, because of some agitation on Twitter. (and as I've pointed out a couple of times, Slack's terms of use mean that everything posted here becomes the property of the primary owner... so by posting any content here, under Slack's ToU you've already lost control of it)

What is the point of the opt-out, then? Virtue signalling?

(I did think that Slack owning everything was already the case, but thought to argue the point, anyways 😊 ) I'm not going to battle over this or anything, just curious.

Thank you for the insights 🤓 I realise I know little about Slack, the org, and consider myself slightly more schooled now. And thanks to you and the rest of the admin group for keeping the wheels turning ⚙️🛠️💪