DarthFreyr

An "error" could be like it did a grammar wrong or used the wrong definition when interpreting, or something like an unsanitized input injection. When we're talking about an LLM trying to convince the user of completely fabricated information, "hallucination" conveys that idea much more precisely, and IMO differentiating the phenomenon from a regular mis-coded software bug is significant.

Honest question: is that not just being complicit in their cover up? If they say nothing happened and nobody is allowed in to verify anything, then you say nothing has been verified yet so you'll assume nothing has happened, it's that not just doing exactly what they want? How are you supposed to prevent atrocities if all they have to do is deny and refuse outside observation for you to remain inactive?

If you're unable to explain something as basic as reading (and you clearly didn't bother practicing it either), I'm not sure if explanations are really the right thing for you, but I guess I'll try anyway: Even if those were her only motives (which I'd already have to stretch pretty far to accept), that would not refute the comment you replied to, nor would it support any other inferences one might make from your opposition. Thus, I questioned what meaning you had actually intended to contribute to the conversation, since I assumed that you were trying to communicate some coherent point and not just be a waste of space and energy. If this attempt at insulting me was all you could muster in response though, I don't have high hopes of every hearing that point, and I should probably rethink that assumption.

Yes, murder in all cases. Not like I fully analyzed my word choice at the time, but I think I did have some intent behind calling it that so I'll see if I piece that back together: I think my overall stance was pretty well established at that point, so not risking confusion there. For "journalist-killing" vs "-murdering", my guess is that the former just rolled off the [mental] tongue better, and I would've been more concerned about ambiguous attribution, for which both are basically equal, than morality, so just went with that. I believe "unaliving" supported the more casual off-tone of that question ("doing a little ... on the side") as part of my facetious approach towards the other comment, using that discongruity with the subject matter to emphasize the absurdity I was alleging in the other comment. The frog might be completely unrecognizable at this point though, so YMMV, but yeah, the intent of that word choice wasn't to obscure the nature of that [potential] violence or avoid censorship or other automated tools, I do stand behind considering it kidnapping and murder.

What, getting your cause in the news because you got arrested for supporting it, while also showing the wrongdoers doing wrongs, doesn't have a positive impact? Does getting in the news prevent a genocidal, journalist-killing regime from doing a little activist-unaliving on the side? Or is it that no one could possibly have motives less narcissistic and self-serving than our so-called "leaders"? What even are you saying?

Hey, if you've still got enough digits for blackjack, no problem

To this point, I want to rant a bit about the experience that most convinced me that they actively hated functionality. So I use Pandora for music, just seemed to feel the best. It's not the biggest player out there, not as much direct integration as something like Spotify, but I have simple needs. I could just tell Assistant to 'play Pandora' and the app would open on the one station I use and get going.

Well, Gemini rolls around, and it even says it can fall back to Assistant if needed, so no reason not to try it, right? Of course, if you ask Gemini to 'play Pandora', all it does is tell you "I can't use Pandora yet. Try YouTube Music or Spotify." or similar. How hard can it be to make it understand the user wants to open this mainstream music streaming app and hit the media play button? Too hard for an AI engineer, I guess. Oh well, if you tell Gemini to 'open Pandora', at least it will open up the app--after you manually unlock the phone.

Side note, the voice match is already so selective it only gets me like a third of the time (unless I'm reading a crossword clue from across the room), and I'm not exactly jamming out to a list of all my passwords read out over some looping beat, I think you can just go for it. I swear it used it to be better at detecting voice and snappier about responding when it did, but I guess these giant software companies have sped up devolpment so much it's actually rolled around and started going backwards.

Anyway, 'open Pandora' at least does something, though I've already gotten into a habit of saying 'play' instead of 'open', so it can be annoying at times. I know!, there's that feature to alias a trigger phrase to a routine of commands. I figure out how to get that set up to turn 'play Pandora' into 'open Pandora' (which IIRC was harder than it should have been for some reason). Alright, I go to test it out and "I can't use Pandora ..."!!! I literally went in to manually set up a workaround for this issue that shouldn't even have been an issue in the first place, but do you know what's more important than actually doing what the user wants? Telling the user how absolute garbage our product is!!!

Alright, time to dump this in the trash where it belongs and go back to the old Assistant like Google promised you could. Sike! You forgot that was a Google Promise(tm)! The old Assistant just hangs immediately and crashes. That's actually really convenient though, because my headphones have a button you can use to directly activate a digital assistant on your phone, but it can only be set in the mfr settings app to use Alexa or Google Assistant. Since I'm not gonna use Alexa, I can just leave it set on Google Assistant and--wait, did I slip and say "really convenient" earlier? I guess what I meant to say was "a huge pain since Sony isn't going to correct their short-sighted buffoonery on their (aging, but still going strong) flagship headphones any time soon". Funny how you can just miss a couple keys and completely change what you meant to say.

I have no idea what Google is thinking with their product development team, but they've clearly learned that the best thing to do with an established, functional tool with a good userbase is to just toss it all out the window for the next piece of trash they can put a clever name on. So the question is: Would it be unethical to round up everyone who supports pulling the sort of nonsense I described and "re-educate" them until they reach some baseline level of rationality, or should it go past just non-counterproductive until they can make actually good decisions?

He is saying 1+1, then you are adding your own +1 to get 3 and calling him a liar because it's not 2. That's all there is to it. What sort of "conception" or "context" are you going to add that puts words in his mouth instead of your own?

Even if that is the most common "interpretation", it's already been explained how that is not actually part of his words, but you've done nothing to refute that except double-down on baseless assertions. Innumerable riddles, mind-benders, word games, and garden-path sentences demonstrate how inaccurate the first or most common interpretation of a statement can be. You say he is responsible for his words but blame him for others' misunderstandings of them. And to keep track of the goalposts, it isn't lying if you say one thing and someone else misinterprets that to mean something else you didn't say, even if you weren't flawless in your original phrasing.

I believe that this is the same point: this is not a case of justice being won, and we should still fight for the right thing to be done in the first place (presumably just being paid for as part of a criminal investigation, among a broader context).

I think "plotting" doesn't see a ton of use in that more neutral sense outside of a few idiomatic cases like "plotting a course". I definitely did not naturally associate a presidential run with that navigational sense of "plotting", but instead the "plotting an evil scheme" connotation jumped out. I'd think of planning a presidential run to be more similar in activity to plotting a scheme, another literal plan of actions to achieve a goal, than to plotting a course as a figurative map of those actions. That's why I interpreted pretty sharply that way, at least.

"can be interpreted" would mean that he is not inherently lying, but that you are choosing an interpretation (twisting his words) to try to say that he is. Otherwise I could say you are lying about calling me bad faith because you don't know anything about my religious practices. See how absurd that is?

Is coming into a conversation and clearly laying out my points along with giving reasoning and explanations "bad faith" now? What conventions or norms am I breaking, other than taking a fact- and logic-based approach to reality? Are those not allowed any more?