Piero Bosio Social Web Site Personale

@whitequark@social.treehouse.systems It's interesting to note that there is zero mention of the ethical problems with LLM code generators mentioned in their linked post.

(Or license-ability.)

I don't care that it's good <insert product>, if it's made with <insert unethical process>, I'm not interested.

So even if there were no quality issues and no deskilling, I still wouldn't want it.

Thank you for the alternative.

@Umbreon @minmi No worries, the landlords can be eaten in turn.

@rcombs @mcc Keeping the status quo means keeping all of its horrible sides too.

No future means no directly making it worse (it can rot into a worse condition though), but it also means not making it better.

@kitkat_blue @mcc Some of the politicians are plausibly aware but don't care because it serves their authoritarian desires, which is another issue.

@De_Minimis @dalias @pmdj @scottjenson For now, it was to try and make you rethink this interaction in the future. I gave up on any notion of a more productive conversation.

@De_Minimis @dalias @pmdj @scottjenson I'm glad to know I come off as young at heart.

The rot of my flesh argues otherwise, sadly.

"seen not heard" is something that was said of women and slaves in the past, did you know? That's the kind of culture you're perpetuating unquestioningly.

@De_Minimis @dalias @pmdj @scottjenson Wow, that tells me quite a lot more about your grasp of ethics and personhood than I think you intended.

More importantly, resorting to ad hominem is ignoring the point of my post and doesn't deserve a reply not making fun of it.

@De_Minimis @dalias @pmdj @scottjenson I could reply the same. In fact, I do.

@pmdj @scottjenson To be clear, @dalias's response answers my concern entirely and is fine by me.

@De_Minimis @scottjenson @pmdj @dalias > Don't hear anyone complaining about ai application in the science fields. People are just focused on the slop side of things, not the tangible.

You haven't seen the reports about medical errors and the whitepapers about failing reliability & deskilling of professionals?

From the sound of it the video you're liking is a major case of the latter.

@dalias @pmdj @scottjenson I dislike the notion of mutating the objects, as followers of the one that got blocked may prefer to see the replies by one they explicitly follow (probably unlike the other party).

(This also becomes a question of who one trusts more and that's not a choice I think should be made for the users.)

@dalias @pmdj @scottjenson > By "working", I mean a block function that detaches all past replies by the account you're blocking from your posts, so that you're not serving as a billboard for their opinions every time someone expands your toot.

Wouldn't this mostly be a UI thing?

The objects would still have the "in-reply-to" field pointing the same way.

@GrapheneOS @tedstechtips Yeah, with apps it gets even more complicated given how little introspection mobiles usually permit.

Often a whole self-contained browser you're not allowed to modify. It's incredibly gross.

@GrapheneOS @tedstechtips > That doesn't do anything to address the privacy invasive behavior built into the app's own services providing functionality.

Indeed, that's what I refer to as the '"legitimate" destination' first party being malicious.

Some however are lazy and externalize the enactment of their malice. They can possibly be used with reduced harm (through allowlisting) for some amount of time before they correct & apply their malice everywhere.

@schm43cky @GrapheneOS @tedstechtips That's somewhat similar, yes.

It doesn't quite reach the same granularity.

Both of course also lack per-script granularity, presumably under the assumption that at point the host could just make them all malicious at any given moment.

@tedstechtips@mas.to @GrapheneOS@grapheneos.social Allowlisting requests & assets by default (yes this also breaks everything by default until one allows strictly what they need).

Unfortunately umatrix died a while ago and I'm not aware of anyone else doing it to anywhere near the same degree as it did.

An additional problem is that if the "legitimate" destination is also malicious, umatrix cannot help.

@GrapheneOS > Il y a la surface d'attaque, là pour le coup on est pas des spécialistes de la sécurité

C'est pourtant assez simple. Arretez de couper les coins rond et à peu près tout fonctionne comme il le devrait.

Ah, mais ne pas bâcler en informatique c'est si difficile et peu populaire avec les corpos...

@GrapheneOS@grapheneos.social So they're just compromising on security and quality of their product for literally no reason (besides being collaborateurs, anyway).

And people hire those clowns for something other than comedy routines?

@GrapheneOS@grapheneos.social > /e/ and Murena have repeatedly claimed GrapheneOS is for drug dealers, pedophiles, terrorists and spies

Wow, they used the old tired canard practically verbatim. How pathetic.

@davidrevoy @windigo Someone would make it, of course.