Of course it isn't a perfect analogy. The perfect analogy would be someone stealing the wallet from a little old blind lady when she placed it on a seat next to her in public. Harming those who are disabled, even if they don't know they're disabled yet, is reprehensible. FFS, those who do it are treated worse in prison than those who have literally raped children! When even the supposed "lowest" in our society deems such behavior wrong, it's one hell of a bad look to be defending it in any way whatsoever. Ya may want to consider that moving forward.…..I don’t know man.
I don’t think a minimum expectation of personal responsibility makes someone an “asshole”. An example I always come back to is someone who left their wallet on the seat getting their car broken into. The person/people responsible for the bad actions should of course be held responsible. However I think it’s as important to rebuke the car owner for not doing due diligence.
It’s not a perfect analogy in this case since LLMs aren’t sentient but people have a responsibility to understand what the tool they’re using is doing and it’s risks.
Sort of, yes. It was his father's abuse and the neglect by every adult in his family which resulted in his issues. That's not far removed from someone programming a chatbot to behave this way."ChatGPT told student he was “meant for greatness”—then came psychosis "
Is this what happened to the currently sitting POTUS?
This stuff isn't just a matter of insufficient guardrails. They've literally been trained on websites where this sort of bullshit is said by humans in response to others. They're not well known because they're very shitty places but this is the problem with programming a synthetic text extrusion machine on the entirety of the Internet. All those bad things online also get thrown into the mix, which ABSOLUTELY WILL cause harm to some percentage of folks. It isn't a question of it. It will happen. We know this because it has happened before but the websites are shielded by Section 230 immunity. There is no such immunity for OpenAI because these chatbots aren't someone else posting bullshit on their website. It is the program they created doing it.Who would have thought that there could have been adverse consequences from optimizing systems with the objective of being pleasing to, and indistinguishable from, humans (i.e. supervised fine-tuning through reward modeling, with user preference and "natural" communications as the reward metrics)?
Human-like communication is impressive, and pandering to users keeps them coming back, so these things have been prioritized over correctness and safety guardrails. It makes for a superficially impressive system, but the potential consequences are obvious, particularly when these systems are used by vulnerable individuals.
Just to expand on this "hot coffee" case a bit, Ms Liebeck was not driving when she spilled coffee on herself. If I remember the relationship properly, her nephew was driving. He then parked the car ion the McDonald's parking lot so she could put the cream and sugar McDonald's had handed her in the cup of coffee. To do this, she necessarily needed to remove the lid. This predated cup holders being anything close to standard so she removed the lid in the safest manner possible while sitting in a parked motor vehicle at the time. Everybody, at the time, did this if they wanted to drink the coffee while in the vehicle.And for those wondering about the particulars of the McDonald’s hot coffee case:
https://www.tortmuseum.org/liebeck-v-mcdonalds/
That also was not the first time McDonald’s had been made aware that their coffee was served dangerously hot, yet they persisted in doing so.
There are parallels with cases like the one in the article. People will claim these lawsuits are frivolous, that they are somehow the victim’s fault, that they just needed to handle thecoffeeAI more carefully, and so on. Reality is that these systems are deeply flawed and dangerous, and those responsible for them have been warned and know it.
In my original post I screwed up my phrasing and implied the burden of proof was on the defense, not the prosecution. Nobody seems to have noticed, but I just want to correct that error. Now let me try taking a different approach to explain my reasoning on why I couldn't vote to acquit.Ah but the burden is asymmetric. The defence doesn’t have to show 'this full chain is likely'.
They only need to show that, given the frailties in the prosecution’s proof, guilt isn't established beyond a reasonable doubt.
A juror can have a coherent doubt about one or two pivotal items ( identification, timing, ability to hear/see, knife etc.) and that alone can block conviction, even if other items 'point the same way'.
An acquittal does not require believing an elaborate alternative narrative: it can rest on 'I'm not sure it was him'.
No, what she ordered was "coffee with cream and sugar" and the company chose not to hand her that. Instead they handed her a cup of coffee with some cream and sugar for her to assemble herself. Which, sure, that's fine if your coffee isn't so hot it will fuse your genitals together in under 3 seconds if you happen to spill it on your lap. The problem is this: that coffee was that hot and the company damned well knew it because it had happened to other people literally hundreds of times before that case! This wasn't a one-off. It was a company continuing a pattern of behavior they knew caused harm. The really fucked up part of that case is there was never a criminal charge because that's literally gross negligence, which is criminal in all 50 states and was back then, too.But she requested hot fresh coffee.. and she received hot fresh coffee.
1) You were kind to me in the same thread as Orwell a couple years ago. Thank you to you as well, Grayl.One of the things I appreciated about the Cumberbatch/Freeman series was that they gave Holmes' probabilities plays a tweak on the "House" treatment, where the initial diagnosis is incomplete, and that his wily foes make him play the Vizzini game. Nor, do I think, was the literary version quite as big on the mathematics of probability, but more on positioning things in binary terms. To the extent Conan Doyle took liberties with that I'd have to pick them up for a re-read. It has been a while.
Not as big a corpus, but the Benoit Blanc movies show a sense of him knowing when he can prove something and when he knows he has to bluff someone into giving away their game.
What's even worse is while it's not a killswitch, OpenAI has a system that is supposed to flag cases such as this for human review and intervention. In 2 other cases, discovery showed that very system flagged literally hundreds of individual chats with particular users yet OpenAI literally did nothing at all. So they're even ignoring the very guardrails that were supposed to catch this shit in the first place, albeit after the potentially problematic output had already been extruded.(And before anyone objects with a description of the modern table saw failsafe where the blade is instantly arrested on sensing flesh, you'll need consider the fact that AI tools don't have a corresponding instant killswitch on treading into psychologically problematic territory.)
I haven't either but humans are fallible. Nobody has to be perfect to be worthy of that. Seriously, you are more than the sum of your flaws.Well, you've obviously never dated me, lol
Well, you've obviously never dated me, lol
I haven't either but humans are fallible. Nobody has to be perfect to be worthy of that. Seriously, you are more than the sum of your flaws.
It was making light of how many women have thrown rocks at my head during the breakup, lolHey, don't you talk about one of my friends that way.
It was making light of how many women have thrown rocks at my head during the breakup, lol
Trust me, I'm loooooooooong over it - haven't bothered dating in about 25 years now (except for 3 months with an alcoholic lesbian - another long story, but quite fun).
And to be clear: I'm autistic, so I'm not the problem - the social interaction is far more hassle than it's worth unless Sofia Vargara shows up.
I need to pull you up on this.Well, you've obviously never dated me, lol
It was a joke. Seriously people lighten-up - self deprecating humour is a good thing. And there's much more important things to be worrying about right now than dates.I need to pull you up on this.
Being "date worthy" and being worth of compassion and kindness are not the same thing. There are expectations within a relationship that are somewhat orthogonal to being compassionate and kind; it's as much about the give and take between the couple as anything else.
I have plenty of issues that mean that I'm a difficult romantic partner. I'm aware of them, and I try to work on them, but they're there. But that doesn't mean that I can't - or won't - extend kindness and support to those around me when it's needed. The reverse holds true. Irrespective of whether somebody is a good dating partner, they are still worthy of compassion and kindness. (Unless they're a complete psychopath like Stephen Miller. Seriously, that guy deserves all the bad luck he gets and then lots more. In fact, I would not be unhappy if the world's supply of bad luck ended up on his doorstep for the next decade. But that's an exception, not the rule, and the default assumption on my part is that a given individual is not such an exception.)
Be kind to yourself, and when you're in a position to do so, extend that kindness to those around you. Life's too short to do anything else, in my opinion.
To a point. Past that point, it becomes damaging, because it can easily slip into negative self talk. Been there, done that.It was a joke. Seriously people lighten-up - self deprecating humour is a good thing.
It's all good. Folks around here mostly just like to be sure is all. Things going unsaid are sometimes be what folks most need to hear. It can also be useful for lurkers at times.It was a joke. Seriously people lighten-up - self deprecating humour is a good thing. And there's much more important things to be worrying about right now than dates.
Which, ironicly, was half the reason I mentioned a small kindness from a couple years ago - which was on my mind because the apartment managers that got fired by the landlord in October turned out to be MAGA levels of corrupt and incompetant (they're getting sued by half the tenants and the landlord and criminally investigated and literal forensic accountants have had to be hired to try to fix all our accounts. They even stole all the goddamn mops from the maintenance closet when they left - as we discovered I woke up Christmas morning to a busted kitchen sink and slightly flooded apartment.)It's all good. Folks around here mostly just like to be sure is all. Things going unsaid are sometimes be what folks most need to hear. It can also be useful for lurkers at times.![]()
Among all this kindness, let's pause and offer our respects to the stupid druid, who gave his all so you might muse.ETA: And I forgot to hit pause while typing this and my stupid druid just died![]()
I am sure that someone could, to go back to the McDonald's case, go over the transcripts and jury instructions and evidence exhibits and come to a different conclusion than the jury did. Heck, if McDonald's thought it was a definitively losing case, they would have settled rather than going to trial.
But insisting that the case was unfairly decided having done none of that work, simply because "hot coffee is served hot," and therefore the jury must have ruled wrong... It's an arrogant disregard to the fact that life and legal issues are often more complicated than a ten-thousand-foot overview might suggest.
WTF?!Yes, judges/juries see more evidence than we do but that doesn’t make their conclusions immune from criticism. Expertise increases probability of correctness - it doesn’t guarantee it.
It isn't arrogance to have coherent, substantive, articulable differences with an expert or with 12 non-experts selected at random.
If it never happens and is a moot point, why did you bring it up?That the jury believed that the officer’s use of deadly force was objectively unreasonable but they acquitted anyway. Anyway, this is a moot point - this never happens.
Okay, this establishes that the subject of the conversation is the "hot coffee" case.This is compounded by the fact she expressly ordered hot coffee, and not iced coffee. This is like ordering a knife and complaining that its sharp.
And I'm complaining that the specific people who tend to criticize this case usually fail to have any respect whatsoever for the intelligence of the jury.Why do people arguing that this was a frivolous lawsuit insist on treating the jury (again, the people with access to all the facts of the case) like they were unintelligent and had no agency?
My specific complaint about specific people treating a specific jury like they were idiots is immediately strawmanned into "a jury's decision is somehow sacrosanct."Same people who argue that a jury's decision is somehow sacrosanct will decry the general idiocy of the American public and question the juries when they rule to exonerate police of police brutality and defend qualified immunity.
Pick a lane...
My point here is, "acquittals (or no-bills) are not necessarily attributable to the jury." It's a recurring theme in this thread!Not at all. Generally, these are no-billed at the grand jury stage, where only the prosecutor presents evidence, and the cliché is that a prosecutor could get a grand jury to indict a ham sandwich.
If there's a no-bill on a police brutality case, that's probably more on the prosecutor presenting the case weakly than any fault with the jury members.
Okay, you know what? Fair. I'll concede the point: acquittals in police brutality cases are not just a grand jury thing.Philando Castile shooting, Terence Crutcher shooting - there are so many cases where police are acquited by a jury not a grand jury.
Except, contrary to your expectations, I'm not critical of the jury. I'm critical of the prosecution, because so was the jury.But again, I'm not sure why you're blaming the jury and not the prosecutors.
From the Crutcher case, a juror 'said the jury felt that the state’s prosecution “was shoddy,” and that they could have gotten a conviction “had they done a better job.”'
You basically reiterate that I should be critical of the jury, because a juror who thought the case was provable but not proven didn't vote to convict and force a mistrial, which the juror regrets.Cherry picking quotes to support your narrative I see. From the article you cited:
"[The anonymous juror] said that while he never came out and said he felt Shelby was guilty, he will always regret not “hanging the jury.”
“At one point I talked with another juror about just hanging the jury, and making the state try the case again,” he said. “We really agreed that if they did a better job, they could have convicted her. And maybe the right thing to do was just make them do it again, maybe they do something different and a different jury convicts her.
“I’ll always feel like a coward for not doing that.”
I point out that's not the jury's job and not what they swore to do. If they are not given enough evidence to convict, they acquit. And now we get into the "null verdict" thing. Reminder: Shelby was acquitted of Castile's murder.I'll accept that that's something a jury could choose to do. But there's a reason why they aren't given the option of returning a "mistrial" verdict. They're supposed to be deciding whether the prosecution proved their case or not. If they didn't, the Fifth Amendment's Double Jeopardy clause prevents the state from trying the case a second time.
Yes in that case the jury can ask for additional questions to ascertain facts and/or return a null verdict.
What is a "null verdict," in this scenario?
To summarize:That the jury believed that the officer’s use of deadly force was objectively unreasonable but they acquitted anyway.
Yeah, it's not clear how much blame is correct to assign. Would they have had the same breakdown later from something else? Would we blame that something else? What's the aggregate added mental stress from these things?The title put me in mind of the Self-Esteem Movement from the 90s, but I'm pretty sure parents weren't trying to convince their kids they'd become oracles or religious figures, just high-achieving workers. I also have to question what DeCruise's mental state was prior to using ChatGPT, because I have a hard time believing that reading text output from a chatbot gave an otherwi0se mentally-healthy man bipolar disorder.
I wasn't contradicting your entire post: I was reinforcing an epistemic principle since you used the word 'arrogance' in there.I said all of that. In the part that you quoted immediately above that paragraph you wrote.
What the heck are you responding to?!
Because it is technically an option present to a jury even though it is rarely exercised.If it never happens and is a moot point, why did you bring it up?
It's not a strawman. Just like you were talking about 'specific' people, I'm talking about specific people when I wrote:Actually, let's run through that whole quote chain.
Okay, this establishes that the subject of the conversation is the "hot coffee" case.
And I'm complaining that the specific people who tend to criticize this case usually fail to have any respect whatsoever for the intelligence of the jury.
My specific complaint about specific people treating a specific jury like they were idiots is immediately strawmanned into "a jury's decision is somehow sacrosanct."
Fair enough. I'll grant that I agree with you on this position now. I have changed my mind a bit more on the role of the jury - I was wrong to blame them when they're statutorily limited in what they can do.And then whatabouted into a discussion about not being so deferential to juries in cases police brutality. Fine, I'm game.
My point here is, "acquittals (or no-bills) are not necessarily attributable to the jury." It's a recurring theme in this thread!
Okay, you know what? Fair. I'll concede the point: acquittals in police brutality cases are not just a grand jury thing.
You offer examples. I dig down on both of them in subsequent posts, but the reply chain we're talking about focuses on Castile, so we'll do the same here.
Your inital premise is that the "Same people who argue that a jury's decision is somehow sacrosanct will [...] question the juries when they rule to exonerate police of police brutality and defend qualified immunity."
So, since I'm someone who (in your strawman) is treating the jury of the McDonald's case as "sacrosanct," your contention is that I should be critical of the jury acquitting Philando Castle.
Except, contrary to your expectations, I'm not critical of the jury. I'm critical of the prosecution, because so was the jury.
You basically reiterate that I should be critical of the jury, because a juror who thought the case was provable but not proven didn't vote to convict and force a mistrial, which the juror regrets.
I point out that's not the jury's job and not what they swore to do. If they are not given enough evidence to convict, they acquit. And now we get into the "null verdict" thing. Reminder: Shelby was acquitted of Castile's murder.
You can, but you would be wrong. As a result of this trial, McDonald's changed its extensively documented practice of serving scalding hot coffee in order to save money, which had injured multiple people in addition to this specific plaintiff. "The system" worked just fine to encourage a stonewalling corporation who had refused to simply cover medical bills to settle with the injured party and adjust its unsafe practice.But I do think we can criticise the outcome without laying the blame on the jury - i.e. the system failed to mete out justice.
Not sure what you're on about. I don't have any problem with McD decision.You can, but you would be wrong. As a result of this trial, McDonald's changed its extensively documented practice of serving scalding hot coffee in order to save money, which had injured multiple people in addition to this specific plaintiff. "The system" worked just fine to encourage a stonewalling corporation who had refused to simply cover medical bills to settle with the injured party and adjust its unsafe practice.
Because you and many others chose to accept the anti-tort pro-business PR spin is more an indictment of your lack of critical thinking than it is of the legal system. Even when you realize maybe your original thinking was misguided, you can't just admit you didn't understand the case at all.
It does not matter this was said by a member of a unique subgenre of ambulance-chasing-like attorneys: this is the exactly correct question
Internet fuckwads, though, incite rage, and rage boosts engagement. A lot more than friendly, agreeable chatter. So there's going to be a heavy bias toward rage inducement, because money.There is a question as to how much of this sycophancy was intentionally built in; I think there is an additional effect that might be hard for them to remove which is that it is learning from human conversations on the web and those tend to be sycophantic by nature. People are trying to be pleasant with each other (sometimes!) and empathizing with each other and so I think the natural training bias will be towards that kind of behavior. But then the question is why or how they prevent it from trolling and "greater internet fuckwad" behaviour, which is also common on the internet.
It matters if we feel compelled to prove gross negligence rather than simple. Strict liability means it doesn't matter--the defective product is defective and its manufacturer is liable. First, let's agree on its defective nature. There isn't much productive reason to debate Thing 2--intent--without that.There is a question as to how much of this sycophancy was intentionally built in; I think there is an additional effect that might be hard for them to remove which is that it is learning from human conversations on the web and those tend to be sycophantic by nature. People are trying to be pleasant with each other (sometimes!) and empathizing with each other and so I think the natural training bias will be towards that kind of behavior. But then the question is why or how they prevent it from trolling and "greater internet fuckwad" behaviour, which is also common on the internet.