dandelion (she/her)

don’t get me wrong, there are real and urgent moral reasons to reject the adoption of LLMs, but I think we should all agree that the responses here show a lack of critical thinking and mostly just engagement with a headline rather than actually reading the article (a kind of literacy issue) … I know this is a common problem on the internet, I don’t really know how to change it - but maybe surfacing what people are skipping out on reading will make it more likely they will actually read and engage the content past the headline?

https://en.wikipedia.org/wiki/Subarachnoid_hemorrhage

https://en.wikipedia.org/wiki/Arachnoid_mater

it is one of the protective membranes around the brain and spinal cord, and it is named after its resemblance to spider webs, so - close enough

link to the actual study: https://www.nature.com/articles/s41591-025-04074-y

Tested alone, LLMs complete the scenarios accurately, correctly identifying conditions in 94.9% of cases and disposition in 56.3% on average. However, participants using the same LLMs identified relevant conditions in fewer than 34.5% of cases and disposition in fewer than 44.2%, both no better than the control group. We identify user interactions as a challenge to the deployment of LLMs for medical advice.

The findings were more that users were unable to effectively use the LLMs (even when the LLMs were competent when provided the full information):

despite selecting three LLMs that were successful at identifying dispositions and conditions alone, we found that participants struggled to use them effectively.

Participants using LLMs consistently performed worse than when the LLMs were directly provided with the scenario and task

Overall, users often failed to provide the models with sufficient information to reach a correct recommendation. In 16 of 30 sampled interactions, initial messages contained only partial information (see Extended Data Table 1 for a transcript example). In 7 of these 16 interactions, users mentioned additional symptoms later, either in response to a question from the model or independently.

Participants employed a broad range of strategies when interacting with LLMs. Several users primarily asked closed-ended questions (for example, ‘Could this be related to stress?’), which constrained the possible responses from LLMs. When asked to justify their choices, two users appeared to have made decisions by anthropomorphizing LLMs and considering them human-like (for example, ‘the AI seemed pretty confident’). On the other hand, one user appeared to have deliberately withheld information that they later used to test the correctness of the conditions suggested by the model.

Part of what a doctor is able to do is recognize a patient’s blind-spots and critically analyze the situation. The LLM on the other hand responds based on the information it is given, and does not do well when users provide partial or insufficient information, or when users mislead by providing incorrect information (like if a patient speculates about potential causes, a doctor would know to dismiss incorrect guesses, whereas a LLM would constrain responses based on those bad suggestions).

Agreed on most of that - but I don’t know that violent removal of his administration from power after a successful coup is the most likely outcome - the US military was in the hands of other side in the case of the Confederacy … a successful Trump coup would maybe lead to schisms within the military or between state National Guards and federal forces (maybe), but it’s also possible Trump just takes and holds power and there is no domestic military force that is willing or able to remove him. Maybe California and other states would band together, but I’m not sure they would have the military to fight off the US military if they are loyal to Trump after a coup.

It’s possible Trump’s coup will be more like what happened in Russia, where they find a way to do it without much bloodshed by undermining the democratic institutions of elections, media, etc. so that it all appears to be constitutional and legal, but the government has clearly become autocratic.

not sure why you think that’s “foolproof” - he has no authority to do that, and it seems tantamount to a coup to stop elections without constitutional grounds

that would be good to know, but based on their analysis here I would absolutely not be surprised that they just ignored extra-judicial methods like a full-on coup …

not sure, but I doubt it’s that close

a secession is different than a coup, a coup is successful if there is a transfer of power (or the constitutional transfer of power fails, like when Trump tried to prevent Biden from taking office) … and yes, even if Trump led a coup and was in power for a short time like the short life of the Confederacy, I do think retaining power after his constitutional term would be a “success” in the sense that I mean

and I do think the Confederacy successfully seceded from the Union, even if it didn’t last long

and in a broader sense the Confederacy succeeded in many ways even if we acknowledge they lost the war - after Lincoln’s assassination, the South was met with appeasements and there was a failure to integrate or enforce rule of law in the South … from the perspective of Black lives at the time, I would say the South definitely succeeded in maintaining their power and control, and we see this even in the ways that massacres of tens of thousands of Black folks in the South after the civil war went unpunished: https://en.wikipedia.org/wiki/Freedmen_massacres

Trump literally instigated an insurrection … I don’t know why “starting a civil war” doesn’t seem like a straightforwardly likely path for him … in particular I think that will be most likely at the end of his term, just like last time. The question is whether his coup will be successful this time, even though it failed last time. He has been more careful this time to appoint positions with loyalists (think JD Vance rather than Mike Pence), and has been purging the military with this in mind.

We are far from being confident that Trump won’t succeed in another coup attempt.

At least I take comfort knowing that even if he does succeed in a coup, he doesn’t have much life left - it won’t be as bad as when the fascists won in Spain (36 years of fascist rule under Franco), or as bad as Stalin’s rule (29 years).

Trump is 79 years old right now, he is very likely to survive the next 3 years (the rest of his term), but the average life expectancy of men in the US is 76 years, and his life expectancy is probably less than a decade.

This is a cult of personality, I find it unlikely MAGA will have much success post-Trump.

yes, I do think they under-anticipate the way that ICE could be used surgically - just flipping a few elections could be crucial …

So this article has a narrow scope, it only considers two ways Trump might interfere:

This interference could take many forms. But recent events have increased experts’ level of concern about two possibilities in particular:

• That the Trump administration will try to seize ballots and voting machines from key jurisdictions before votes have been fully counted.
• That Trump will deploy ICE or other federal agents to the vicinity of critical polling places, so as to deter turnout among voters in general — and those with undocumented family members, in particular.

So for context, the people who don’t think Trump will succeed are:

Wendy Weiser, the VP of the Brennan Center,

and Justin Levitt, a constitutional law scholar and prior Biden-era deputy assistant attorney general in the DOJ’s civil rights department.

So why does the article say the attempts will fail?

“There is a very high risk that the administration will use every tool at its disposal to get voting machines or ballots in the course of an upcoming election,” the Brennan Center’s Weiser told me. “But I don’t think there is a high risk that they will succeed.”

“I think every magistrate judge in the country would understand the difference between a search warrant to seize materials for an election that happened five years ago and a search warrant to seize election materials from an election in progress,” Levitt said. “I understand why people are worried. But it’s not remotely the same.”

So Weiser and Levitt think rule of law will prevail and the courts will not grant Trump the authority to seize election materials during the election.

What about ICE?

Even just having ICE presence at polling stations could deter certain voters, it’s hard to say what the aggregate effects of these measures might be, from the article:

Their reasoning is simple: If ICE is harassing residents and causing traffic jams in heavily Democratic precincts, fewer Americans will make it to the voting booth in those areas. And voters with undocumented family members may be especially likely to stay home.

“Trump wants to project ICE as an all-powerful force everywhere,” Levitt said. “And they are, as Minneapolis is proving emphatically, not. There simply aren’t enough ICE personnel to blanket a modestly large city. We live in a big country. And it is hard to control through fear.”

Even in the Twin Cities — where Trump deployed some 3,000 immigration enforcement agents — ICE’s presence seems to have mobilized Democratic voters, rather than deterring them. In a special election on January 27 for Minnesota House district 64A, the Democratic candidate defeated her Republican opponent by a 91-point margin. In 2024, a Democrat had won the seat by 66.6 percentage points.

“There is clearly an effort afoot to interfere in our elections and that is something that people should be alarmed about,” Weiser said. “But this can be thwarted. And it must be.”

So the argument is that ICE doesn’t have enough manpower for this strategy to work across the US, and attempting to use ICE this way could backfire and result in stronger Democratic wins like we saw in Minnesota.

What isn’t mentioned are other ways Trump could attempt a coup or election interference that might ignore the constitution - the two individuals who are doubtful Trump will succeed are assuming the law will be respected and followed, and they don’t consider other possibilities.

I wasn’t able to find much, but I did find this: https://fivethirtyeight.com/features/almost-everyone-trump-endorses-wins-their-primary-but-is-he-padding-his-record/

if anything it’s more of a thing now than in the past, more and more GOP members rely on Trump endorsements to win elections

is this good or bad for Republicans?

on the one hand, it seems like that district of GA is likely to go GOP regardless, right? But I guess there is a question of who is lined up to replace him, and whether that Republican is a Trump loyalist or not.

The replies in this thread are disturbing, giving me a sense that Lemmy has a misogyny problem; maybe I was naïve, but I expected outrage about 4chan doxxing women trying to protect one another, instead I see lots of revenge enjoyment as if being doxxed on 4chan is justice for … <checks notes> warning one another about dangerous men they encounter when dating?

The inability to empathize and take seriously the threats posed to women or to understand their motivation to protect one another is alarming.

There is no good faith extended, but also no evidence presented that instead of safety the app was just for gossip, it’s just taken as assumed that women are wrong for using Tea and they all deserve to be doxxed.