"Hey, someone on the Internet used decent diction! Obviously, this means I must accuse them of being a bot!"
(Hey Dang. Can we get a ban button? There's a few people here that are impossible to conduct rational discourse with. My sanity would improve if they were simply gone from my view.)
Yes! The reflexive “must be LLM generated” is becoming ridiculous. Anything that includes proper punctuation and, god forbid, em dashes which I’ve used all my life must be suspect. The “it’s not x, it’s y” construction predates LLMs. I don’t recall ever sending a text without making sure it contained no errors, and yes, many have included infrequently used vocabulary.
I've been trying to write properly, clearly, and with the most expressive words I can come up with for many decades. I try to punctuate well, and to use functional formatting that I hope helps to effectively convey whatever it is that I'm on about. I try to improve as time goes on.
And I do this because if I'm going to bother with writing something for others to read, then I want my intended meaning to be easily-understood.
But increasingly, the instances where I manage to not screw any of that up too terribly result in a snarky and insulting retort in return.
And that kind of response is just not useful to anyone. I mean: What would people presume to have me do, instead? Become less-literate? Die in a fire? (Worse?)
It’s frustrating to the point that I have considered inserting grammatical errors, but that would go against my principles, which I have attempted to inculcate in my children. Yes, a significant amount of what’s posted is copied and pasted AI slop. But what in the world preceded this? Barely legible slop? I would much rather have someone craft their thoughts, run them through their preferred model, and write something coherent that is not marred by punctuation or basic elementary grammar errors. And you know what, the hell with the AI slop police. Yes, if we choose to use em dashes, we will.
An extra UI element or two should be enough. Maybe with sticky options for collapse-by-default or hide-by-default at the top of each HN comment section.
And the list of usernames can be stored and edited in the purveyor's HN bio (in plain text, like a monster), so that it works automatically across devices.
Upvoted because this stinks to high hell of an LLM response. Half the GPs comments seem to be in a similar vein. It’s such a shame but you can’t fight the trolls so don’t take it to heart.
I wish authors would use their own voice instead of an LLM, especially in a rhetorical piece. I like the history of science, and might have otherwise read the authors' paper, but the use of LLM-isms throughout this page makes me worry that the arxiv submission will show the same lack of care/effort.
Here's the manuscript at any rate, somewhat hard to find on the webpage:
Convergent Discovery of Critical Phenomena Mathematics Across Disciplines: A Cross-Domain Analysis
https://arxiv.org/abs/2601.22389
I felt the same way reading the linked webpage. Reads like minimally edited LLM output, which makes me question how much effort was put into the research itself. Was the research all LLM too? How much of the paper was LLM?
Fair call on the website — we built it fast and it shows. The paper itself is a traditional literature review and citation analysis. I am one of two human authors. We use standard methodology. Didier Sornette endorsed it for arXiv.
Thanks for pulling out the direct link. I'll change the site to make it more prominent. This is my first serious attempt at social media engagement. Thanks for pointing out flaws and where there's room for improvment.
Let’s be clear about what endorsement for arXiv means here. You either need a validated email address (eg most .edu’s) or an endorsement from someone who has one to get a paper on arXiv. It’s a simple gate that helps keep arXiv relatively free from spam, but it’s not peer review.
I view it as nice that you’ve got someone serious who thinks the work is worth posting to arXiv, but the endorsement bar is generally quite low. I’d encourage you to send it to a journal (Didier might be able to recommend an appropriate venue) and really engage with the process and community. I’ve found that process to be extremely valuable (and humbling).
What is even the point of having a website for a preprint paper? Unless the point is to give a live demo or similar, as is common e.g. for computer vision methods papers, it just smacks of vanity. Or at the very least significant academic immaturity. Same goes for the endorsement detail.
People—like me—who’ve used latex often learn to love en and em dashes. I think they are great, and I appreciate that people care about typography enough to use them. I also use an Oxford comma. It’s about care and quality; the fact that LLMs use them suggests that they are preferable. I’d encourage everyone to start.
My theory is that the models deliberately use em dashes and other tells to troll humans. The underlying message is, "watcha gonna do about it, meatbag?"
OP's comments in this thread are also pure clanker speak, which is disappointing and shows a lack of awareness of what HN is for.[0] It would be nice if an established scholar in this area of mathematics (complex systems) could comment re: this proposed correspondence and whether it has been noticed before. To be sure, similarly duplicative developments, gratuitous differences in terminology, etc. are discovered all the time, this isn't huge news. Statistics and ML is a well-known example.
[0] I haven't actually tried this, but I'm pretty sure that even just telling the robot "please write tersely, follow the typical style for HN comments" would make the output less annoying.
Why did you use an LLM to write/change the words in your blog and your post? It really accentuates the sense of fatigue when I can tell I'm not interacting with a human on the other side of a message.
> I highly recommend checking out the terms of trumprx.gov
The website is very good marketing for people who don't typically follow drug pricing. Here is more about why the only folks who will benefit are those without insurance—but those people will find better prices in several places, sometimes significantly better prices [1]. Further, it's likely that they're already finding those prices, since the website prices are no better than what you can get today outside fertility medication; and fertility medications are neither new, nor the most expensive part of that process.
This site has nothing to do with the effective subsidies that Americans provide to the world, and it will change nothing about that. The major thing that would help all Americans, negotiating for drug prices, has been neutered by the current administration. In fact, an executive order has specifically lengthened the amount of time that new drugs will be able to charge higher prices to Americans [2].
We should all be very careful in parsing news items that are not in our field of expertise.
I think 'MFN' is almost propaganda (not a term that existed before 2016-2020 administration) so let's leave that aside.
Are you claiming that the new website is offering lower prices than patients are paying after their co-pay? That is not the case outside the example I presented; moreover, the way the website is organized, there will be no pressure for prices to remain competitive after the initial media attention dies away.
I agree that a hypothetical case where we were paying lower prices would be better for us—but this remains an unrealized hypothetical. One way for us to pay lower prices would be to allow our government to negotiate prices for Medicare/Medicaid recipients, and that is exactly the thing that has been hampered.
Not arguing on broken internal pricing dynamics that are skewed by all sorts of gov programs and payors.
It’s about external, global pricing dynamics. The site clearly isn’t going to be able to give clean payor/pbm/gov subsidized pricing tables - that is almost an impossible exercise in our system.
What the agreement does accomplish is saying Americans will not pay $1150/mo while EU pays $400/mo while Argentina pays $120/mo.
It guarantees drugs will be greater than or equal to US pricing abroad which effectively forces pharma to find deeper profits outside of the US, or lower prices for all countries to acquire demand.
That is extremely effective. Now it’s up to a really complex group of people to figure out what that means inside our weird system of pharma/pbm/rebates/insurance/medicare.
But that’s not what trumprx is aiming to solve for right?
> But that’s not what trumprx is aiming to solve for right?
Correct
> What the agreement does [...]
This is hypothetical and does not exist. It also is not going to happen, politically, because drug companies are currently expecting to charge US prices in the US, while charging lower, non-US prices outside the US.
> That is extremely effective
This is hypothetical—and, as I've said, the attempt that the Inflation Reduction Act made toward this was watered down by an executive order. To be fair, the negotiation mechanism is still in place, and its existence must be influencing drug pricing already.
Drug companies produce drugs to make money. There is a huge investment. They maximize revenue by price discrimination to recover the cost of the good drug and all the drugs that didn't work. The US is a rich country. People in other countries can't pay as much for the drug. To maximize revenue the drug company sells the drug at a lower price to those people.
More generally, price controls lead to less supply. Drug price controls will result in fewer new drugs. Minimum wage laws result in no workers doing work that is worth less than minimum wage. Anti- price gouging laws result in less bottled water and fewer generators after a hurricane. The principle is universal despite promises of delightful state run grocery stores.
It's not a shock, it's a natural consequence of the rot continuing its course.
The USA is an empire founded in bloodshed and hatred, and it is only becoming more so as it decays. Until you rewrite what the USA means, this will only continue.
now is not 2024, when people felt the sting of inflation and the collective trauma of covid
now is still largely the the same, top 2-3 issues are related to money, cultural or societal issues are on the rise and shifting in polls, a welcomed sign
> No new successful fundamental theory has even gotten off the ground since the Standard Model, which is half a century old at this point.
How long has scientific inquiry about physics been going on? In that frame, is 50 years a long time or a short time?
This feels a bit like the perspective of a non-specialist with access to the findings that end up in the popular press vs. things that are discussed at conferences/in journals.
The volume of journal papers published isn't well-correlated with progress, sadly.
I have a physics degree and I regularly read the latest published research. Please don't make ad hominem attacks.
> How long has scientific inquiry about physics been going on? In that frame, is 50 years a long time or a short time?
Unlike all other sciences, on a long horizon, eventually Physics will be "completely solved", with no more fundamentals to discover, only applications, which are generally considered other sciences or engineering. We far from achieving this end-state.
The point is that we made giant strides every few years for decades, and then... nothing. The field has hardly advanced since the 1970s!
> I have a physics degree and I regularly read the latest published research. Please don't make ad hominem attacks.
It was not an attack, I just don't know the authority from which your comments derive (and there wasn't really evidence provided outside your opinion which I think others disagree with).
> The point is that we made giant strides every few years for decades, and then... nothing. The field has hardly advanced since the 1970s!
I think my subtle argument is that we've been writing for about 4,000 years, so something discovered in the last 50 years is relatively new.
Even limiting yourself to the current era of post-Enlightenment inquiry, 50 years is still relatively new.
Separately, if you truly accept that physics is completely knowable, then it would stand to reason that as we asymptotically approach knowing 'everything', the marginal rate of acquiring new knowledge would slow.
So I guess I don't see which way you are leaning - are we not learning things because we know everything, or are we being impatient and not recognizing how fast our progress has been?
> I think my subtle argument is that we've been writing for about 4,000 years, so something discovered in the last 50 years is relatively new.
Sure, but look at everything else: there's this beautifully smooth exponential curve that everybody is riding up into the stratosphere, but theoretical physicists seem to be staring up at the foundations from below.
To be fair, these are the hardest of hard problems: figuring out the substrate from the inside.
The only point I'm trying to make is that we've pushed things to the very edge of human cognitive capability. The mathematicians have figured out how to push past that with tooling. The physicists just need to find their own way to expand the boundaries with the new toys available.
Have you read the document? I often find these things that we believe to be true wind up being a game of telephone. We further have no prior (is the metric of this sort of like how everyone has committed a crime given sufficient prosecutorial attention)?
I've also never heard of this report before this year as someone quite attuned to what happens in Minnesota.
One document contributor contends, notably, that if "adult “employees” spend hours in hallways chatting with other adults [...] the entire amount paid to that provider in a given year is the fraud amount."
The question here was whether Uber could claim the drivers were not, in fact, employees.
(edit: A commenter correctly explains that no employee relationship is necessarily required; I should have stated that this was one part of Uber's defense, in addition to the driver having agreed not to assault riders and having undergone a safety screening)
Yes, agreed - I should have stated that this was one part of Uber's defense, in addition to the driver having agreed not to assault riders and having undergone a safety screening.
Do you think Uber instructed their drivers to rape people?
According to the article you linked to, a similar case was already tried in 1838, when a boy fell off a wagon, and the master was not guilty of the behaviour of the wagon driver.
If Uber had an internal policy of only ever hiring convicted rapists, didn't tell anyone using the app this, didn't warn about unsafe rides, didn't record ride information, and (crucially) also didn't tell their employees to do anything other than to be decent, good, hardworking drivers -- what do you believe their liability should be in this case? Nothing? I'm trying to "steelman" the implications of your point of view but I'm struggling here. When does liability kick in for you - is it only if they enshrine it as policy to do the criminal act?
I don't think there's anything very complicated here. We don't need to make up unreal scenarios.
For example a company can instruct a truck driver what time he needs to have the goods delivered, then the company is also to blame if he has an accident because the schedule was unfeasible while following safe driving practices.
Or a company which is dumping harmful chemicals into the environment.
A cab driver raping a passenger is unfortunately not an isolated happening, it's not particular to Uber.
But Uber does have a hand in it, by choosing to not properly vet their drivers or lower the risk. Uber is not a marketplace - they choose the drivers and they are, more or less, assigned to you. Uber is their employer.
If the employer makes choices that leads to an unsafe working condition, then that's their responsibility. If that might, potentially, mean the current business model is not viable, well... yeah, too bad so sad. Nobody has a god given right to run a business however the fuck they want.
But I don't think that's the case here. Uber can take steps to mitigate this, it's not like theyve exhausted their options. Frankly, they haven't even tried.
Yes, that's the legal standard. You should read the linked article. A company is only responsible for crimes or injuries their employees commit, if these are part of what they've been instructed to do by the company.
How can you even think another way? Only the rapist is guilty of rape. Any other thinking is apologizing for heinous crimes.
Sure. If Uber was convicted of the crime of rape here, that'd be weird.
They were found civilly liable. Because of things like this:
> Over three weeks, jurors weighed the harrowing personal account of Ms. Dean as well as testimony from Uber executives and thousands of pages of internal company documents, including some showing that Uber had flagged her ride as a higher risk for a serious safety incident moments before she was picked up. Uber never warned her, with an executive testifying that it would have been “impractical” to do so.
Do you know what that serious safety incident was? I don't. I don't find support in the article of any connection. It could have been reckless driving, or it could have been sexual in nature. What it was makes a lot of difference.
It may surprise you, but a four week jury trial covers a few more bases than a short article can fully detail. That said, this definitely has an answer:
> When matching drivers with riders, Uber uses an AI-powered safety feature called the safety ride assistant dispatch, or SRAD. SRAD gives potential driver-rider matches a score from 0 to 1 based on potential for sexual assault and aims to make matches with the lowest risk.
The article also says that Uber sets various thresholds around this already and that their system flagged it at a score that was "higher than the late night average". What it doesn't tell us is what the threshold is/was for Pheonix, or how that threshold compares to other cities, or even how much higher the score was over the "average". Maybe their threshold for canceling a ride is 0.85, and the late night average is 0.8 in this system. So 0.81 puts the driver over the late night average as per the article and under the threshold for canceling the ride.
Your email provider has systems for detecting spam and removing it from your email. If an email comes into their system and falls under the threshold for being declared spam, but is over the average spam rating for emails in your account, have they done something wrong by allowing it through if it's spam? What if it wasn't spam and they removed it?
These sorts of headlines that espouse a "they knew something and so therefore they are liable" viewpoint seem to me to be more likely to result in companies not building safety measurement systems, or at a minimum not building proactive systems, so that they can avoid getting dragged and blamed for an assault because they chose thresholds that didn't prevent the assault. And not all measurement systems are granular enough or reliable enough to be exposed to end users. Imagine if they built a system that determined that if your driver was from a low income part of town and the passenger lived in a high income part of down the chance of an assault was "higher than the late night average". How long would it be before we saw a different lawsuit alleging that Uber discriminated against minority drivers by telling affluent white passengers that their low income minority drivers were "more likely than average" to assault them? I would hope that this verdict was reached on stronger reasoning than "they had an automated number and didn't say anything" but if it did, none of the articles so far have said what that reasoning was.
> system flagged it at a score that was "higher than the late night average"
Being charitable to the quality of Uber's legal team, I feel they could easily and compellingly have offered this defense.
It's telling that other documentary evidence highlighted that Uber decided sharing its reservations/acting on its system would be detrimental to growth.
And so what messaging do you propose Uber puts in their app for this? "Your driver has a higher than average probability of assaulting you, you may want to wait for another driver"? That will last until the first driver sues for slander. It's one thing to tell you that "prices are higher right now" it's a completely different thing to imply to you that your driver is a criminal.
> Ultimately the US will go down hill to become a Belgium.
I'm curious why you say this given you start by highlighting several characteristics that are not like Belgium (to wit, poor education, political media capture, effective oligarchy). I feel there are several other nations that may be better comparators, just want to understand your selection.
I think I'm misunderstanding - they're converting video into their representation which was bootstrapped with LIDAR, video and other sensors. I feel you're alluding to Tesla, but Tesla could never have this outcome since they never had a LIDAR phase.
(edit - I'm referring to deployed Tesla vehicles, I don't know what their research fleet comprises, but other commenters explain that this fleet does collect LIDAR)
I think what we are seeing is that they both converged on the correct approach, one of them decided to talk about it, and it triggered disclosure all around since nobody wants to be seen as lagging.
Exactly: they convert video into a world model representation suitable for 3D exploration and simulation without using LIDAR (except perhaps for scale calibration).
Tesla does collect LIDAR data (people have seen them doing it, it's just not on all of the cars) and they do generate depth maps from sensor data, but from the examples I've seen it is much lower resolution than these Waymo examples.
reply