@scitodor's banner p

scitodor


				

				

				
0 followers   follows 0 users  
joined 2022 September 05 17:32:33 UTC
Verified Email

				

User ID: 651

scitodor


				
				
				

				
0 followers   follows 0 users   joined 2022 September 05 17:32:33 UTC

					

No bio...


					

User ID: 651

Verified Email

I was thinking in % as accurate as a human, but didn't say that. I'll correct my original post.

I was also assuming it wouldn't be as good as a human, because of course a general model wouldn't accidentally be better than a specialist... Or would it?

I don't think it would be that hard to devise an experiment to get at least a rough idea of it's capabilities. Get some doctors, maybe professors, to devise questions of the sort OP was generating, present them to LLMs and real doctors (and maybe non-doctors with Google for an extra point of comparison), then have the professors grade the answers blind. I recall people giving LLMs math problems in this way, but I don't know if experiments have been performed with any rigor.

To your final point: what is the purpose of a doctor? Is it to heal at all costs, or is it to make people feel better? Many people go to doctors seeking specific prescription medication as a goal, when they could cure their ailment in a better but more laborious way. Some of them don't even have an ailment and only desire the effect of the meds. Many people also have procedures performed which physically make them less healthy, but fulfill their desires.

So, what should an LLM say to these people? "Don't take the pills, exercise" or "don't have surgery, improve your social life?"

I suppose it's not a new problem, but it does move the power and responsibility. Who is it moving to though?

This is the nature of taking risks, no? You can always say in hindsight that it was a bad idea, but when you succeed it's a triumph. They can seem more or less sensible in the prior analysis, but you're only really going to know if your assumptions are correct once you try it.

I actually think there is a good reason for doctors to wear masks: spit. Quite often, saliva just gets ejected from the mouth, whether it's from coughing or merely speaking. It doesn't really matter if masks prevents spread of disease via spit, it's just gross to get spat on, and this is a situation where I'm comfortable saying that it "just makes sense" that masks will block ejecta from the mouth. It goes both directions, people don't want to be spat on by doctors, and doctors don't want to get their patients' spit in their mouths.

Anecdote: my mother noticed a cashier at her local grocery store had briefly stopped masking when the mandates lifted, but went back to wearing one. When asked, the cashier said she wasn't worried about COVID, but she had realized that the mask prevented her from getting customers' spit in her mouth. And cashiers don't have to get right up in their customers' faces.

Not a technological barrier then but a tactical one. Agree on all counts.

How far is "that far"? Have you ever worked in robotics or automation? Outside of tasks that are purely software I mean.

I don't think the CIA was telling people they had psychics, but they spent quite a bit of time researching it https://www.cia.gov/stories/story/ask-molly-did-cia-really-study-psychic-powers/

It'll get him some good PR with the UFO believers.

That's all just a matter of opinion though. You might not find any interest in a "hunk of metal at the bottom of the ocean" but many people clearly do. If the CEO was being truthful about his desire to inspire people, then his submarine could have been a stepping stone to letting the average person view the Titanic with their own eyes, and further beyond, opening the depths of the ocean to occupation and exploitation. History has shown there is plenty of glory in colonization!

What really determines if your risk was stupid? An old saying goes, "if it's stupid and it works, it ain't stupid." Likewise, if this Titanic exploration venture worked, would we be calling it stupid?

I do agree there is plenty of adventure to be had today, and as someone who finds "rocks in space" pretty interesting I am participating in the greater efforts to explore and exploit them. Consider my original post a bit of nostalgia.

wow, Karl did videos with one of the murderers in the CHAZ militia, that's way worse than I was expecting.

Are you repeating a statement from the ARFCOM thread/other socials or is that your assessment? Genuinely curious.

I ask because the escalation of rhetoric throughout this ordeal has been really bothering me. The first one that stuck out at me was how Karl and all of his supporters are now "groomers." It's easy to see how his detractors arrived here: he allegedly made comments supporting "drag queen story hour," and drag queens have been reported making sexual comments to children. But is implicitly, nay, explicitly accusing Karl of grooming children helping? As far as the "antifa CHAZ murder" claim goes, to my knowledge this has been leveled at Tacticool Girlfriend, an openly trans gun YouTuber. I have never seen this claim made with evidence, though I haven't made any effort to confirm or deny it myself. But again, it's now more often than not that I see this person referred to as "the CHAZ killer" as if it is hard fact.

I'm shooting from the hip here, but I think produce in America has gotten a lot better in the last 20 ish years. When I was young, my family was an early adopter of Natural and Organic foods, and the only places to get it were Whole Foods and local "health food" boutique grocery stores. We immediately noticed that, while processed foods didn't have the heightened artificial flavors, the produce and simpler items (such as seedy whole grain bread) were richer and more complex. There was plenty of coverage pooh-poohing the movement (heard plenty of "duh all food is natural and organic dummy" back in the day, and if you were anti-GMO you were letting people in Africa starve). Now, even the chain grocery store in my grandma's podunk town carries some Organic produce. Perhaps the popularity has encouraged "conventional" produce to pull back a bit on prioritizing looks at the cost of taste in order to compete.

Don't underestimate the stupidity of supposedly intelligent people. Remember that the CIA (or some element of it) seriously studied psychic powers after believing that the Soviets had them. If 9/10 Russian science advisors say "no UFOs" but one says "yes absolutely UFOs we need to study this it could be the end of us!", I'd say the chance of Putin listening to the UFO guy is greater than 1/10 thanks to how human psychology works. Those odds might be worth it.

For the record, I don't believe there's a concerted UFO psyop, but I think it is likely at least some of the US intelligence apparatus is happy to let the believers run with it a bit for the counterintel effect. There are enough wacky people already, you don't need to plant evidence.

At first this comment annoyed me - you can't reframe it this way, it completely changes the question! But upon further thought, it helped me see why the original poll might have turned out the way it did. To a rational person, like most of you here, the pills can be treated as machines. Pull lever A, you might die. Pull lever B, you are guaranteed to not die. Obviously you pull B. However, to a socially driven person, like those on Twitter, nothing is merely an object. Everything has character, and which you choose reflects on you. You don't drive a car: you're a Toyota driver. You don't enjoy a movie: you're a Star Wars fan. You don't vote for a politician: you're a Burnie supporter. I don't think the blue pillers are literally imagining their choice as an election, but the framing was the kick I needed to see things in the society centric way. If you don't choose the blue pill, you're signaling that you're anti-social. This is the essence of culture war: when everything is viewed from the tribal perspective, all choices are made in judgement of others.

I've seen other signs of hope too, like LaCroix seltzers becoming so popular that Coca-Cola and PepsiCo were forced to respond. Just today I read this article from ArsTechnica, and was surprised by its relative frankness: https://arstechnica.com/health/2023/09/what-can-we-do-about-ultraprocessed-foods/

I generally expect Ars to be solidly lefty, and while they pay a bit of lip service to mention "exposure to racism and weight bias" as potential causes of poor health (eye roll) the message is pretty clear: obesity is unhealthy, it's widespread because people are eating too much, and our modern diet of ultra processed foods is the culprit.

As the author of the first thread I feel a need to comment, though I don't have strong feelings on the matter. My impression from my original research was that Ballard is a weirdly motivated guy who actions were probably on net good. My guess here is he, like many men with a little power, simply couldn't keep it in his pants. Disappointing, but not surprising. Of course now the media is feeling vindicated and taking this opportunity to do victory donuts over the conservatives who rallied around the film. I doubt it will have any large effect since the public has already moved on from the spectacle of the "weird conservative Christian Qanon etc. movie" hitting it big. People who disliked the movie will feel smug, and the people who liked it will ignore the scandal and make excuses, or simply move on.

Ugh, the "unknown elements" quote made me facepalm when I first heard it - because I can imagine how a perfectly reasonable reality got twisted into "woah aliens!" Firstly, there is absolutely research being performed on wreckage of UFOs; literally, flying objects that are unidentified. Zero implication of extra terrestrials or advanced technology. Say US Intel dredges up some scrap from the ocean floor in the vicinity of a North Korean missile test, or collects shards of metal from a Middle Eastern desert where they detected what could be an Iranian drone crash - those could easily (and correctly) be classified as "UFO wreckage". They're going to perform tests to identify what they collected, which could involve classifying what they're made out of - not because it's some mysterious wonder element, but because they want to know if NK is being provided Chinese metals, or they want to estimate how far Iran is in developing composites. Some of the material is inevitably going to be unidentifiable because tests aren't perfect - there's always a chance of false readings and the samples could be damaged beyond recognition or contaminated or many other reasons. Now imagine some bookkeeper with no context reads the report. They're going to see "TOP SECRET", "UFO", and "test inconclusive, composition unknown". They leak that to the whistleblower with aliens on the mind, and of course it's going to turn into "alien vehicles with unknown elements."

Even in a helical mag 90 rounds is going to be... Well maybe you could come up with a clever way to multi stack Trounds and make them telescoped, but that is still going to be a chunky mag. Sounds badass though.

If only it were so simple. Even the high IQ people with previously successful ventures come up with real stinkers. Many such cases.

I would appreciate if you could find that post, I don't think I've read it.

I do have to wonder, does the US military produce more procurement boondoggles than other nations on average/per capita?

My gut feeling is that it is partially true. There are almost certainly more boondoggles in other countries that we don't hear about because they aren't as open as the USA and they don't get the same media attention.

However, the US military is terrified not of their adversaries technologically surpassing them, but of them merely closing the gap. All of US strategy relies on overmatch, the theory that you don't just outperform your enemies, you can totally crush them if desired. In the situation where your advancement is plateauing, the bad guys are rapidly catching up, and you're at a severe disadvantage in manpower but you're flush with money and brain power, why not throw everything at the wall and see what sticks?

I have followed Karl for quite a while and I still don't know what he was thinking with that post. He is far enough left where it is not beyond the pale that those are his true feelings, though more extreme than anything he'd publicly expressed before. However, he is also a troll, and an explicit part of his mission statement is to "slay sacred cows." He even made an emblem of a silhouetted cow with a sword through it to represent this idea for his channel.

Personally, I have decided to walk away from his content. Regardless of his true beliefs I cannot support someone who makes such agitating statements. It's not a decision I made lightly.

I purposely kept my focus limited for this piece but I would be curious to know more about OUR. The critical articles throw out some accusations of questionable tactics and effectiveness, but I don't recall anything claiming it's a scam. The credible thing to do, if you wanted to undermine the film, would be to publish a scathing expose of OUR, but that would take effort.

I find this idea very scary - I'm sure you know about the LLM phenomenon of what's been dubbed hallucination, so how can you trust the model with something as impactful as medicine? It might be better than a human doctor* 99% of the time, or even 99.99%, but like self driving cars this is exactly the kind of case where being good but not quite as good as a human is really dangerous. Remember that the script of House, MD is in the training set right along side WebMD with nothing to distinguish the truth from fiction - are you going to catch when the model probabilities switch from being driven by Gray's Anatomy to Grey's Anatomy? There's no way to know mathematically, because the whole thing is a non-deterministic black box.

If the process is as simple as to pattern match symptoms to diagnosis from a list of previous diagnoses with their associated symptoms, why can't this be done with a good old stupid database? Something like this has to exist. Does the LLM have an advantage beyond writing the answer in nice prose?

My warning to anyone trying to evaluate LLMs: your ability to judge it is inevitably compromised by the fact that the only optimization parameter the model has is how good the output sounds to a human, and you are not immune to being human!

*Edited, original was overall accuracy instead of relative to a human

The steelman: if the occupants of the submarine are saved (which is nigh impossible at this point) he would be made an example of: his company discredited, he himself fined and possibly jailed. As it is, the sub will either not be found, or it will be found filled with corpses, which also serves as an example of what not to do but is arguably not as effective as having a person to haul in front of a courtroom.

Agree on all accounts. There is so much we still don't know, especially about the cartridges. The Army could be hiding this information because it's embarrassing... or that could be SOP. I don't have the experience to say, and I doubt many of the detractors do.

I wonder, if one looked at everything the Army ever adopted and counted up the number of items with performance claims that turned out to be false, how many times it actually happened. People point to failed programs saying "look at all the stupid stuff they tried to make!" but the vast majority failed before adoption... which is how the system is supposed to work! And don't even mention The Pentagon Wars, that book is full of stupidity and lies.

I remember reading discussions on the plastic ammunition before the NGSW was awarded. Some (most?) of the commentators had the Sig entry in distant third because it was barely any different than existing platforms and we're talking up the supply chain benefits of the GD and Textron entries. Honestly, the seethe when Sig won was pretty funny. I take it to mean that the other systems just didn't work very well. At least True Velocity is continuing to develop their ammunition, so maybe even.277 Fury will come in plastic someday.

Nice reference on the Trounds too, the Dardick makes a lot more sense when you look at it as a platform for the Trounds instead of a serious pistol design. I think they also had an auto cannon design that used the Trounds to achieve some ungodly fire rate.

My point is that there is nothing notable about the content of the movie to warrant controversy - it's all about how it was made and is watched by the wrong type of people. Skim through that linked Rolling Stone article - it's pretty obvious culture war. And understand that this isn't an isolated event; there are many more articles like that one, and most of the discussions of this movie on Reddit quickly devolve into claims of conservative histeria over supposedly non-existent child trafficking.

We also made heavy use of fire bombs in Germany and Japan, and I find it hard to believe only the Allied forces had that capability. From what I have read about the effects they sound as terrifying as chemical weapons but for some reason the ethics of using them are never discussed, especially in comparison to chemical and nuclear weapons. What is the difference? There was the lingering memory of The Great War where chemical weapons were deployed and condemned, so that certainly contributed. Fire has also been a component of war since mankind first sought to wage it, so perhaps packaging it in bombs did not seem like much of an escalation (though these were napalm bombs, which have uniquely nasty properties). From a strategic perspective, chemical weapons do not destroy materiel, and ostensibly the goal of strategic bombing is to destroy your opponents industry. Fire bombs do not share that issue, and American strategists liked them in Japan especially because of how effective they were at destroying (mostly wooden) Japanese structures - never mind the fact that they tested the bombs on a replica of a Japanese family home.