id
int64 0
12.9M
| type
large_stringclasses 5
values | by
large_stringlengths 2
15
⌀ | time
timestamp[us] | title
large_stringlengths 0
198
⌀ | text
large_stringlengths 0
99.1k
⌀ | url
large_stringlengths 0
6.6k
⌀ | score
int64 -1
5.77k
⌀ | parent
int64 1
30.4M
⌀ | top_level_parent
int64 0
30.4M
| descendants
int64 -1
2.53k
⌀ | kids
large list | deleted
bool 1
class | dead
bool 1
class |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
41,812,400 | comment | epolanski | 2024-10-11T19:02:24 | null | Filling aisles of unhealthy and very tasty stuff then finding the cheat to keep doing this. | null | null | 41,812,163 | 41,811,263 | null | null | null | null |
41,812,401 | comment | supportengineer | 2024-10-11T19:02:33 | null | The benefits of fluoridation were amazing. I didn't have a single cavity until I was 25 and had moved away from that area to a non-fluoridated area. | null | null | 41,812,176 | 41,811,263 | null | [
41812707,
41812606
] | null | null |
41,812,402 | comment | NoRagrets | 2024-10-11T19:02:33 | null | This will make Norway one of the richest countries in addition to their already wealthy citizen pension fund.<p>This is very good news. Phosphate was deemed a finite source and we were running out of it. We are good for another 1000 years. At least. | null | null | 41,812,382 | 41,812,382 | null | null | null | null |
41,812,403 | comment | synergy20 | 2024-10-11T19:02:38 | null | the compilation database is essential for editors these days, cmake can generate it directly but not all code use cmake, in that case I just use compiledb, which is inactive (there is a newer compiledb-go though), then use bear. somehow bear did not work well for me comparing to compiledb so compiledb(now compiledb-go) is my main tool now. | null | null | 41,784,263 | 41,784,263 | null | null | null | null |
41,812,404 | story | null | 2024-10-11T19:02:41 | null | null | null | null | null | 41,812,404 | null | null | true | null |
41,812,405 | comment | acabal | 2024-10-11T19:02:47 | null | The more important point the comment you're replying to makes is not "what if people could diet and exercise" - i.e. accept the modern American lifestyle as given, plus force yourself to go to the gym and eat chicken and broccoli - but rather that the modern American lifestyle in fundamentally structured to lead to people being overweight.<p>Instead of being forced to drive everywhere for the most basic possible human needs - like getting groceries, going to the doctor, or dropping kids off at school - as is the case in 90% of America - what if you could walk to those places instead? You would get exercise as part of your every day life, with no extra effort!<p>What if instead of corn syrup being so heavily subsidized, we could use more filling sweeteners in a lower amount instead? What if people lived closer to agriculture, instead of in faraway suburban tract housing only accessible by car, so they had easier access to fresh meat and vegetables, instead of ultraprocessed package food?<p>These dreams are not "diet and exercise", they are a fundamental reshaping of American lifestyle that would directly lead to weight loss. We know this, because America used to look like this before, say, 1940. In old photos you see people in huge crowds in streets as they walked to their everyday errands, and menus and recipes of the era are mostly minimally processed food that is mostly local. Americans of the past were not overweight, because the way society arranged its physical existence didn't permit it! | null | null | 41,812,258 | 41,811,263 | null | [
41812567,
41812540
] | null | null |
41,812,406 | comment | oblio | 2024-10-11T19:02:48 | null | Musk has said many a things that bump his share price. I wish someone did one of those fact checkers for his Tesla statements, I doubt he's at 50% hit rate. | null | null | 41,810,178 | 41,805,706 | null | null | null | null |
41,812,407 | comment | BlueTemplar | 2024-10-11T19:02:50 | null | Right, so hardly anyone linking and so no presence on search engines yet.<p>(As a contributor for the wiki for PoE1 I really should have tried the obvious link... less obvious on the phone though !) | null | null | 41,811,230 | 41,797,719 | null | null | null | null |
41,812,408 | story | clairegiordano | 2024-10-11T19:02:54 | Talking Postgres podcast episode with Tom Lane: How I got started as a developer | null | https://talkingpostgres.com/episodes/how-i-got-started-as-a-developer-in-postgres-with-tom-lane | 1 | null | 41,812,408 | 0 | null | null | null |
41,812,409 | comment | anthk | 2024-10-11T19:02:55 | null | FB and Twitter do the same.<p>At least long ago Twitter had a Javascript-less login which worked everywhere and reduced the doomscrolling a lot.<p>Teens don't need dumb smartphonies, they need actual portable netbooks with a usable handbag. They need to produce actual media, not silly viral videos without no actual content. | null | null | 41,812,310 | 41,812,310 | null | null | null | null |
41,812,410 | comment | TheSalarian | 2024-10-11T19:03:04 | null | That's how Swiftkey works, yeah. If you put your cursor at the end of a word, hacker for example, pressing shift once gives Hacker as an autocorrect option, and a 2nd press gives HACKER. | null | null | 41,804,363 | 41,762,483 | null | null | null | null |
41,812,411 | story | amrrs | 2024-10-11T19:03:08 | Using GPT-O1 for Advanced Analysis in Osint Profiling | null | https://espysys.com/blog/gpt-o1-osint-profiling/ | 1 | null | 41,812,411 | 0 | null | null | null |
41,812,412 | story | null | 2024-10-11T19:03:11 | null | null | null | null | null | 41,812,412 | null | null | true | null |
41,812,413 | comment | sph | 2024-10-11T19:03:14 | null | You can see that in the lab-grown burger space already: it is a massive opportunity for companies to create an ultra-processed version of meat, and label it as healthy as the public opinion has gone blindly against meat and on the vegan bandwagon. They'll claim it's more eco-friendly, they can sell it 5x the price of beef and rake billions. | null | null | 41,812,056 | 41,811,263 | null | null | null | null |
41,812,414 | comment | whall6 | 2024-10-11T19:03:24 | null | Genuinely not trying to be a dick, but if you have time to peruse hn you might have time to go to the gym | null | null | 41,812,080 | 41,811,263 | null | [
41812596,
41812448
] | null | null |
41,812,415 | comment | mppm | 2024-10-11T19:03:36 | null | Ah... NATO expansion? Alleged discrimination against the Russian majority in Eastern Ukraine? Alleged foreign interference in the Maidan revolution? Not that I'm trying to start a discussion here, but dismissing the other side's arguments as "not even meant to be sensical" is <i>exactly</i> what I was arguing against. | null | null | 41,811,597 | 41,807,681 | null | [
41812617
] | null | null |
41,812,416 | comment | kuhsaft | 2024-10-11T19:03:39 | null | It still exists, but now “ad blockers” can’t use the blocking API to record and forward metrics on hits. Ad blockers don’t even need the webRequest and webRequestBlocking permissions anymore.<p>Now, if an ad blocker has webRequest permissions it’s a red flag.<p>For example <a href="https://developer.chrome.com/docs/extensions/develop/concepts/content-filtering#adapt_rules_based_on_observed_requests" rel="nofollow">https://developer.chrome.com/docs/extensions/develop/concept...</a> uses webRequest to send telemetry back to some remote server. | null | null | 41,812,305 | 41,809,698 | null | [
41812566
] | null | null |
41,812,417 | comment | lazide | 2024-10-11T19:03:46 | null | After this much time and that much work, how is it possible for a physics theory to <i>not have a single testable/falsifiable prediction</i> without it being intentional? It has been <i>over 80 years</i>. [<a href="https://en.m.wikipedia.org/wiki/History_of_string_theory" rel="nofollow">https://en.m.wikipedia.org/wiki/History_of_string_theory</a>].<p>The evidence is in the absence. | null | null | 41,811,377 | 41,808,127 | null | [
41812449
] | null | null |
41,812,418 | comment | al_borland | 2024-10-11T19:03:52 | null | I think the slider is mixing up kg and lbs for the calculations at the bottom.<p>Neat idea. I made one they dumps out a chart of the ranges based on a person’s height and tells them how far away they are from “normal”. Seeing it visually is nice.<p>Mobile has some issues as well, at least on Safari. I had to switch to landscape. | null | null | 41,812,216 | 41,812,216 | null | null | null | null |
41,812,419 | comment | rpmisms | 2024-10-11T19:03:52 | null | Because it was a halo vehicle, and it worked out very well. The Model Y is literally a lifted and heightened model 3 because it's not a halo vehicle. | null | null | 41,806,049 | 41,805,706 | null | null | null | null |
41,812,420 | comment | kimixa | 2024-10-11T19:03:52 | null | This reads more like a puff piece about Tesla's charging system than a real discussion on /why/ anything matters. And the prose is positive about everything.<p>I'd love to know what information is "missing" from the J1772 communication protocol, the trade offs on a digital vs analogue "protocol"? Does it require more control circuitry on each side? Are there safety issues with such high power being controlled by a relatively complex microcontroller and protocol? Is the claimed difference even <i>relevant</i>? J1772 has since 2012 allows digital information to be passed using a AC power-line communication standard. J1772 claims to support 240v, so why are they only using numbers showing the "advantage" of the (likely only available in non-residential supplies) 277v compared to 208v? And why aren't the comparing agains the CCS2 standard? Surely that's the most obvious competitor?<p>The article being so unfailingly positive yet incomplete triggers my "Advert" detector. | null | null | 41,811,646 | 41,811,646 | null | [
41812799
] | null | null |
41,812,421 | comment | ompogUe | 2024-10-11T19:03:54 | null | > For example, AR glasses could stop you from crossing the street when there is an oncoming car, regardless of what the traffic lights indicate.<p>Not sure this is the best example: Maybe turning off the AR and using your eyes and ears would work better.<p>Killer app for me is having the manual about an item I use embedded in a qrcode on the item itself, instead of in a paper book or online.<p>(Not anti-AR by any means - have been working with it/thinking about it since ~1998) | null | null | 41,808,955 | 41,808,955 | null | null | null | null |
41,812,422 | comment | bijection | 2024-10-11T19:04:08 | null | I've finally switched (back) to firefox today.<p>I switched from firefox to chrome for their superior devtools a few years back, but hopefully firefox has had time to catch up.<p>Everything old is new again! | null | null | 41,809,698 | 41,809,698 | null | [
41813042,
41812698
] | null | null |
41,812,423 | comment | mrweasel | 2024-10-11T19:04:13 | null | Exercise strengths your cardiovascular system, immune system, improves bone strength, builds muscle (which helps you as you age, preventing the risk of falling and fall related injury), improves mental health, reduces stress. If your in good physical shape, it's less of an issue if you carry around a little extra fat.<p>You can still be in bad health, even if your weight is spot on, but it's rare that you exercise a lot, but is overall unhealthy. | null | null | 41,812,084 | 41,811,263 | null | [
41813019
] | null | null |
41,812,424 | comment | connicpu | 2024-10-11T19:04:13 | null | Completely missing the point of GLP-1 agonists. The point is that it breaks the cycle by giving you the willpower to eat less. It doesn't magically make the calories you eat not contribute to your weight, it just makes it easier to eat less and still feel full. It also counteracts insulin resistance, which is another problem inherent to obesity. | null | null | 41,811,617 | 41,811,263 | null | null | null | null |
41,812,425 | comment | null | 2024-10-11T19:04:22 | null | null | null | null | 41,812,366 | 41,811,646 | null | null | true | null |
41,812,426 | comment | modzu | 2024-10-11T19:04:24 | null | chrome hasnt been cool for a while | null | null | 41,809,698 | 41,809,698 | null | null | null | null |
41,812,427 | comment | valval | 2024-10-11T19:04:35 | null | What? I have a son and a wife and I couldn’t care less about nuclear war or climate change or any abstract and distant catastrophe that could face humanity as a whole.<p>Living in the future is a silly affair. There’s only one moment and it’s the present. | null | null | 41,809,731 | 41,807,681 | null | null | null | null |
41,812,428 | comment | null | 2024-10-11T19:04:38 | null | null | null | null | 41,810,206 | 41,808,943 | null | null | true | null |
41,812,429 | comment | llamaimperative | 2024-10-11T19:04:39 | null | Those benefits <i>and</i> a billion more! Better state of nature, richer social fabric, probably less political polarization if people were bumping into their neighbors more frequently instead of going from drywall box to steel box back to drywall box day in and day out. | null | null | 41,812,380 | 41,811,263 | null | [
41812679
] | null | null |
41,812,430 | comment | gengelbro | 2024-10-11T19:04:40 | null | Someone compare the pollution to the average steel mill. The focus here is political and clickbait focused. | null | null | 41,811,032 | 41,811,032 | null | null | null | null |
41,812,431 | comment | acureau | 2024-10-11T19:04:42 | null | Or perhaps we as people just tend to idealize foreign cultures, because from our perspective they're different and interesting. You could view it as patronizing or endearing. You don't have to look far to find the exact opposite happening all over the world.<p><a href="https://en.wikipedia.org/wiki/Westernization" rel="nofollow">https://en.wikipedia.org/wiki/Westernization</a> | null | null | 41,812,107 | 41,811,309 | null | null | null | null |
41,812,432 | comment | rightbyte | 2024-10-11T19:04:43 | null | You should specify if you mean Israel, China or Russia. It is impossible to tell which you alure to. | null | null | 41,804,881 | 41,804,460 | null | null | null | null |
41,812,433 | comment | fragmede | 2024-10-11T19:04:44 | null | because ozempic reduces the food cravings, patients are able to implement and stick with a diet change. it's not like "put down that cheeseburger and have a salad" is something they haven't heard before and haven't internalized already, it's just their brain won't do it. ozempic gives them the space on their brain to actually do it. | null | null | 41,811,874 | 41,811,263 | null | null | null | null |
41,812,434 | comment | seanhunter | 2024-10-11T19:04:48 | null | I think the prediction you may be referring to is supersymmetry, which was apparently empirically disproved by the LHC, or at least the supersymmetric extension to the standard model was disproved.<p><a href="https://www.sciencenews.org/article/supersymmetrys-absence-lhc-puzzles-physicists" rel="nofollow">https://www.sciencenews.org/article/supersymmetrys-absence-l...</a> | null | null | 41,811,703 | 41,808,127 | null | null | null | null |
41,812,435 | comment | BrenBarn | 2024-10-11T19:04:51 | null | One that I came across a few months ago and was unexpectedly charmed by is /CannedSardines . A fun little niche interest, lots of good vibes, some useful stuff like recipes. One time a user even bought up a huge load of higher-end sardines that went on deep discount at a local store, then shipped some to others on the forum to share the bounty. | null | null | 41,807,371 | 41,807,371 | null | null | null | null |
41,812,436 | comment | cma | 2024-10-11T19:04:52 | null | Many people use tablets and foldables with keyboards | null | null | 41,811,406 | 41,809,698 | null | null | null | null |
41,812,437 | comment | gcr | 2024-10-11T19:04:56 | null | We have good food, but you won't find it on "The Easy Path."<p>The Easy Path is that gentle encouragement to hit up Chipotle for lunch, because it's "right there."<p>The Easy Path says dinner's hard and you've had a long day, so get something simple, like take-out or microwave.<p>The Easy Path is entropy. The Easy Path is self-care over struggle. The Easy Path is simple carbs shown on prominent display in store shelves. The Easy Path is advertising.<p>Hitting the gym isn't on The Easy Path, but forgetting to cancel your gym membership is.<p>These days, big food companies love "The Easy Path" because it's so easy to commoditize, it's the "Path that Americans are Expected to Take." For financial stewards, being on The Easy Path turns lack of willpower into your ally.<p>On the other hand, getting good food in the US requires passing the marshmallow test: you have to meal prep, or you have to shop around the sides, or you have to get something on the salad menu. You have to say no to advertising. You have to expend willpower, the most limited of resources to the average American. You have to Go Hungry or Suffer, or have An Upset Stomach. You frequently have to spend more money or time.<p>Semaglutides are not currently on The Easy Path. Maybe they will be someday. I personally doubt that, because putting GLP-1 on The Easy Path would require big food companies to rethink their entire portfolio.<p>But you're not wrong in that they could be Easy Path-ajdacent. The dialectic would shift: food companies would shift around to be Organic and Nutritious and Less Calories and find other ways to stay on The Easy Path. Sugar and fat's addictiveness is highly Easy Path-enabling, and that's a pretty big vacuum to fill. | null | null | 41,812,319 | 41,811,263 | null | [
41812510,
41812585,
41812706
] | null | null |
41,812,438 | comment | wyldfire | 2024-10-11T19:05:00 | null | The title is less clear than it could be IMO.<p>When I saw "no dependency" I thought maybe it could be no_std (llama.c is relatively lightweight in this regard). But it's definitely not `no_std` and in fact seems like it has several dependencies. Perhaps all of them are rust dependencies? | null | null | 41,811,078 | 41,811,078 | null | [
41812628,
41812665,
41812629
] | null | null |
41,812,439 | comment | sunshowers | 2024-10-11T19:05:05 | null | Thanks! To be fair, there are certain advanced scenarios which Rust's mutex model can't handle either -- sometimes you want to protect writes via a mutex, but allow reads without them (maybe you're okay with torn state on reads). This is a rare, expert use case with architecture-specific considerations that must be handled with care.<p>I do think Rust's mutexes handle almost every use case that can be thrown at them, though, and in a way where it's next to impossible to get it wrong. I think if you're writing a browser engine in the 21st century you should bake in parallelism and concurrency from the start, and Rust is the most suitable language to do that in. | null | null | 41,809,276 | 41,796,030 | null | null | null | null |
41,812,440 | comment | bozhark | 2024-10-11T19:05:08 | null | Why would you ever take a pill for weight loss?<p>That’s seems entirely fucked up | null | null | 41,811,263 | 41,811,263 | null | [
41812472
] | null | null |
41,812,441 | comment | btilly | 2024-10-11T19:05:10 | null | The problem is that we need an ego to be healthy. Attempts to destroy it can wind up compromising your mental health.<p>The first part of the solution is to be careful what's in your ego. See <a href="https://paulgraham.com/identity.html" rel="nofollow">https://paulgraham.com/identity.html</a> on this topic. See <a href="https://blog.codinghorror.com/the-ten-commandments-of-egoless-programming/" rel="nofollow">https://blog.codinghorror.com/the-ten-commandments-of-egoles...</a> for how careful choices in what we value in ourselves, can lead to thinking better.<p>This of course still leaves us with an identity. For that I've found that gratitude can help us deal with pain. And so targeted gratitude can help us avoid cognitive dissonance when we otherwise would be overrun by it.<p>Sadly, neither skill is widely taught in our society. | null | null | 41,812,297 | 41,808,127 | null | null | null | null |
41,812,442 | comment | kjs3 | 2024-10-11T19:05:16 | null | This is beyond true. Read and understand what your cloud SLAs <i>are</i>, not what you think they are or what you think they should be. There was significant consternation generated when I pointed out that the SLA for availability for an Azure storage blob was only 4 nines with zone redundancy.<p><a href="https://azure.microsoft.com/files/Features/Reliability/AzureResiliencyInfographic.pdf" rel="nofollow">https://azure.microsoft.com/files/Features/Reliability/Azure...</a> | null | null | 41,809,079 | 41,805,446 | null | null | null | null |
41,812,443 | comment | MattPalmer1086 | 2024-10-11T19:05:24 | null | Thanks! | null | null | 41,810,921 | 41,803,650 | null | null | null | null |
41,812,444 | comment | jasonhong | 2024-10-11T19:05:26 | null | This is a really great question, and well beyond my areas of expertise. What I can point you to is this excellent book by my colleague Bob Kraut and several of his colleagues, entitled Building Successful Online Communities: Evidence-Based Social Design. It summarizes a lot of empirical research into design claims, about how to socialize newcomers, increasing contributions, quality of contributions, and more.<p><a href="https://direct.mit.edu/books/monograph/2912/Building-Successful-Online-CommunitiesEvidence" rel="nofollow">https://direct.mit.edu/books/monograph/2912/Building-Success...</a><p>You might also look into research on pro-social behaviors.
<a href="https://en.wikipedia.org/wiki/Prosocial_behavior" rel="nofollow">https://en.wikipedia.org/wiki/Prosocial_behavior</a><p>One of my favorite books that I learned about from my colleagues is Influence by Robert Cialdini. It looks at how to use known social influence tactics to change people's behaviors. Ideally, these would be used for things that society widely regards as positive (e.g. less littering), though these have also been used for phishing attacks and other dark patterns.<p><a href="https://en.wikipedia.org/wiki/Robert_Cialdini" rel="nofollow">https://en.wikipedia.org/wiki/Robert_Cialdini</a> | null | null | 41,810,836 | 41,780,328 | null | null | null | null |
41,812,445 | comment | bee_rider | 2024-10-11T19:05:27 | null | Is that an opinion formed on something inside your domain of expertise? | null | null | 41,812,073 | 41,811,263 | null | [
41812838,
41812600
] | null | null |
41,812,446 | comment | zdragnar | 2024-10-11T19:05:29 | null | Meh, you see commercials for all kinds of medicines in the US. There's more being produced than your average general practitioner can keep up with, so the manufacturers appeal directly to consumers.<p>Given how ineffective it is once you stop, I'm personally expecting it to become a relatively short lived fad. Insurance companies won't cover it if it truly doesn't improve health outcomes long term (throwing money down the toilet) and people will learn to not pay out of pocket for it.<p>There's a reason insurance companies are loath to cover it for obesity now without prior authorization, which usually requires you seeing a specialist who has ruled out the usual suspects (nutrition and exercise changes). | null | null | 41,811,941 | 41,811,263 | null | null | null | null |
41,812,447 | comment | YeGoblynQueenne | 2024-10-11T19:05:33 | null | >> There’s no consensus in the literature on what these mean even if you make it more specific by talking about “mathematical reasoning”, so I don’t really understand what opinions like these are based on.<p>What literature is that? You can find plenty of very clear consensus on what reasoning is if you read e.g. the literature on automated reasoning. A brief taste:<p><i>Automated Reasoning</i><p><i>Reasoning is the ability to make inferences, and automated reasoning is concerned with the building of computing systems that automate this process. Although the overall goal is to mechanize different forms of reasoning, the term has largely been identified with valid deductive reasoning as practiced in mathematics and formal logic. In this respect, automated reasoning is akin to mechanical theorem proving. Building an automated reasoning program means providing an algorithmic description to a formal calculus so that it can be implemented on a computer to prove theorems of the calculus in an efficient manner. Important aspects of this exercise involve defining the class of problems the program will be required to solve, deciding what language will be used by the program to represent the information given to it as well as new information inferred by the program, specifying the mechanism that the program will use to conduct deductive inferences, and figuring out how to perform all these computations efficiently. While basic research work continues in order to provide the necessary theoretical framework, the field has reached a point where automated reasoning programs are being used by researchers to attack open questions in mathematics and logic, provide important applications in computing science, solve problems in engineering, and find novel approaches to questions in exact philosophy.</i><p><a href="https://plato.stanford.edu/entries/reasoning-automated/" rel="nofollow">https://plato.stanford.edu/entries/reasoning-automated/</a><p>After that you may want to look at the SEP articles on Analogical reasoning and Defeasible Reasoning:<p><a href="https://plato.stanford.edu/entries/reasoning-analogy/" rel="nofollow">https://plato.stanford.edu/entries/reasoning-analogy/</a><p><a href="https://seop.illc.uva.nl/entries/reasoning-defeasible/" rel="nofollow">https://seop.illc.uva.nl/entries/reasoning-defeasible/</a> | null | null | 41,810,807 | 41,808,683 | null | [
41812764
] | null | null |
41,812,448 | comment | llamaimperative | 2024-10-11T19:05:34 | null | You're not being a dick, just utterly unhelpful and non-contributory. | null | null | 41,812,414 | 41,811,263 | null | [
41812468
] | null | null |
41,812,449 | comment | seanhunter | 2024-10-11T19:05:36 | null | It predicted supersymmetry, which has been experimentally disproved. | null | null | 41,812,417 | 41,808,127 | null | null | null | null |
41,812,450 | comment | therein | 2024-10-11T19:05:44 | null | I think it comes down to how aggressive Chrome will be at changing the internal APIs that it uses. They could choose to make it a very expensive patch to maintain. But I think they would have to go out of their way to do that. | null | null | 41,812,032 | 41,809,698 | null | null | null | null |
41,812,451 | comment | wizzwizz4 | 2024-10-11T19:05:53 | null | The claim was about <i>promotion</i>, not effective promotion. (As a sibling comment points out, <i>effective</i> promotion is not unrealistic either. It won't happen on its own, but nothing does.) | null | null | 41,812,281 | 41,811,263 | null | [
41812560
] | null | null |
41,812,452 | comment | null | 2024-10-11T19:05:59 | null | null | null | null | 41,811,887 | 41,811,263 | null | null | true | null |
41,812,453 | comment | bbor | 2024-10-11T19:06:00 | null | Wow, fascinating stuff and "grokking" is news to me. Thanks for sharing! In typical HN fashion, I'd like to come in as an amateur and nitpick the terminology/philosophy choices of this nascent-yet-burgeoning subfield:<p><pre><code> We begin by examining the optimal generalizing solution, that indicates the network has properly learned the task... the network should put all points in Rd on the same side of the separating hyperplane, or in other words, push the decision boundary to infinity... Overfitting occurs when the hyperplane is only far enough from the data to correctly classify all the training samples.
</code></pre>
This is such a dumb idea on first glance, I'm so impressed that they pushed past that and used it for serious insights. It truly is a kind of atomic/fundamental/formalized/simplified way to explore overfitting on its own.<p>Ultimately their thesis, as I understand it from the top of page 5, is roughly these two steps (with some slight rewording):<p><pre><code> [I.] We call a training set separable if there exists a vector [that divides the data, like a 2D vector from the origin dividing two sets of 2D points]... The training set is almost surely separable [when there's twice as many dimensions as there are points, and almost surely inseparable otherwise]...
</code></pre>
Again, dumb observation that's obvious in hindsight, which makes it all the more impressive that they found a use for it. This is how paradigm shifts happen! An alternate title for the paper could've been "A Vector Is All You Need (to understand grokking)". Ok but assuming I understood the setup right, here's the actual finding:<p><pre><code> [II.] [Given infinite training time,] the model will always overfit for separable training sets[, and] for inseparable training sets the model will always generalize perfectly. However, when the training set is on the verge of separability... dynamics may take arbitrarily long times to reach the generalizing solution [rather than overfitting].
**This is the underlying mechanism of grokking in this setting**.
</code></pre>
Or, in other words from Appendix B:<p><pre><code> grokking occurs near critical points in which solutions exchange stability and dynamics are generically slow
</code></pre>
Assuming I understood that all correctly, this finally brings me to my philosophical critique of "grokking", which ends up being a complement to this paper: grokking is just a modal transition in algorithmic structure, which is exactly why it's seemingly related to topics as diverse as physical phase changes and the sudden appearance of large language models. I don't blame the statisticians for not recognizing it, but IMO they're capturing something far more fundamental than a behavioral quirk in some mathematical tool.<p>Non-human animals (and maybe some really smart plants) obviously are capable of "learning" in some human-like way, but it rarely surpasses the basics of Pavlovian conditioning: they delineate quantitative objects in their perceptive field (as do unconscious particles when they mechanically interact with each other), computationally attach qualitative symbols to them based on experience (as do plants), and then calculate relations/groups of that data based on some evolutionarily-tuned algorithms (again, a capability I believe to be unique to animals and weird plants). Humans, on the other hand, not only perform calculations about our immediate environment, but also freely engage in meta-calculations -- this is why our smartest primate relatives are still incapable of posing questions, yet humans pose them naturally from an extremely young age.<p>Details aside, my point is that different orders of cognition are different not just in some <i>quantitative</i> way, like an increase in linear efficiency, but rather in a <i>qualitative</i>--or, to use the hot lingo, <i>emergent</i>--way. In my non-credentialed opinion, this paper is a beautiful formalization of that phenomenon, even though it necessarily is stuck at the bottom of the stack so-to-speak, describing the switch in cognitive capacity from direct quantification to symbolic qualification.<p>It's very possible I'm clouded by the need to confirm my priors, but if not, I hope to see this paper see wide use among ML researchers as a clean, simplified exposition of what we're all really trying to do here on a fundamental level. A generalization of generalization, if you will!<p>Alon, Noam, and Yohai, if you're in here, congrats for devising such a dumb paper that is all the more useful & insightful because of it. I'd love to hear your hot takes on the connections between grokking, cognition, and physics too, if you have any that didn't make the cut! | null | null | 41,810,753 | 41,810,753 | null | [
41812809
] | null | null |
41,812,454 | comment | lucgagan | 2024-10-11T19:06:02 | null | Correct me if I am wrong, but these implementations are all CPU bound?, i.e. if I have a good GPU, I should look for alternatives. | null | null | 41,811,078 | 41,811,078 | null | [
41812550,
41812589,
41812495,
41812535
] | null | null |
41,812,455 | comment | sph | 2024-10-11T19:06:05 | null | What do you eat? I bet $10 either it's calorie-counted standard high carb diet or something that is low in protein and fats.<p>The physiology of dieting, and avoiding hunger, is pretty well understood at this point. Just don't ask your GP or they'll just tell you to stop eating red meat and eat more cereals as the "solution". | null | null | 41,812,384 | 41,811,263 | null | null | null | null |
41,812,456 | comment | detourdog | 2024-10-11T19:06:25 | null | I’m really afraid that everything you mention is pretty general and not my experience. | null | null | 41,810,915 | 41,776,631 | null | null | null | null |
41,812,457 | comment | nextos | 2024-10-11T19:06:31 | null | Exactly, good systems do not rely on willpower. They rather make obvious habits effortless.<p>Deviating from the mean is hard. Bad food and sedentarism are the norm. | null | null | 41,812,339 | 41,811,263 | null | [
41813288,
41812874,
41812568
] | null | null |
41,812,458 | comment | card_zero | 2024-10-11T19:06:50 | null | > Firstly nobody lives forever<p>Lose weight permanently through cremation? | null | null | 41,812,170 | 41,811,263 | null | [
41812513
] | null | null |
41,812,459 | comment | leni536 | 2024-10-11T19:06:52 | null | > But if you've got a separation of duties where a sysadmin sets up seccomp filtering generically across applications<p>Is this even possible, regardless of io_uring? | null | null | 41,788,426 | 41,788,426 | null | [
41813146
] | null | null |
41,812,460 | comment | rstupek | 2024-10-11T19:06:52 | null | Correction type-2 diabetes not type-1 | null | null | 41,812,106 | 41,811,263 | null | [
41812497
] | null | null |
41,812,461 | comment | vundercind | 2024-10-11T19:07:04 | null | My point is that skinny populations don’t seem to be skinnier than the US population due to greater genetically-backed willpower (better “configuration”). At least, if it’s a factor, it’s overwhelmed by other factors, it seems. | null | null | 41,812,361 | 41,811,263 | null | null | null | null |
41,812,462 | comment | greatgib | 2024-10-11T19:07:04 | null | It is really crazy that we are taken hostage/blackmail by whatever harmful decision Google takes in their own interest. | null | null | 41,809,698 | 41,809,698 | null | [
41812647
] | null | null |
41,812,463 | comment | myrmidon | 2024-10-11T19:07:05 | null | "Hardware limitations" are extremely unlikely in my view to establish useful limits.<p>Consider: Do hardware limitations establish useful limits on the kind of problems a computer can solve? The answer is a resounding NO in my view, because the limits of what can be expressed/solved grows so insanely quickly that it becomes a completely meaningless and unreachable limit even for super small computers (less capable than our brain).<p>As for learning time constraints: These are obviously reachable, but still useless in my view because they are too inconsistent- the kind of methods and insights that a human can acquire within a lifetime are completely different between persons, and highly dependent on <i>how</i> the learning happens... | null | null | 41,810,035 | 41,808,683 | null | null | null | null |
41,812,464 | comment | nomel | 2024-10-11T19:07:18 | null | I think it's simply that it's (mostly) black and white. | null | null | 41,812,294 | 41,798,369 | null | null | null | null |
41,812,465 | comment | Teever | 2024-10-11T19:07:18 | null | Are the 3/5ths of Americans who are not obese actually living a healthy lifestyle? Or are they just living a different unhealthy lifestyle?<p>Is a skinny homeless methhead with a great BMI healthy? Or the blue collar dude who has terrible blood pressure from chain smoking and pounding energy drinks on his way to the job site?<p>How many Americans who are not obese are on their way to becoming obese?<p>There are lots of paths to unhealthiness and overeating / eating poorly is only one of them | null | null | 41,812,312 | 41,811,263 | null | [
41812723,
41812614,
41812675
] | null | null |
41,812,466 | comment | tzs | 2024-10-11T19:07:23 | null | Around 500 people in the US are killed in police chases per year, and around 1/4 of them are bystanders. Most of those chases are over traffic violations or non-violent crimes like shoplifting. | null | null | 41,811,215 | 41,810,627 | null | [
41812666
] | null | null |
41,812,467 | comment | eastbound | 2024-10-11T19:07:37 | null | No really: Learn this UX to share your screen.<p>- The last window is always suggested FIRST when sharing,<p>- Learn what to look at: Look at the appearance of your browser window before you go click Share my screen.<p>The second will provide a visual confirmation that you’re clicking the right button. Don’t pass as a boomer when you can be a Millennial! | null | null | 41,811,392 | 41,800,602 | null | null | null | null |
41,812,468 | comment | whall6 | 2024-10-11T19:07:42 | null | Maybe. Or maybe pointing out the obvious | null | null | 41,812,448 | 41,811,263 | null | [
41812491
] | null | null |
41,812,469 | comment | bee_rider | 2024-10-11T19:07:44 | null | Medical FOMO? I wouldn’t worry too much about it, I mean, there are hypothetical upsides to countless decisions we haven’t made, right? We always miss some chances in life. | null | null | 41,811,539 | 41,811,263 | null | null | null | null |
41,812,470 | comment | crooked-v | 2024-10-11T19:07:48 | null | > As a default it is.<p>Why? | null | null | 41,812,166 | 41,811,263 | null | null | null | null |
41,812,471 | comment | bozhark | 2024-10-11T19:07:48 | null | This is absurd.<p>Caloric intake and outtake is just that. | null | null | 41,811,926 | 41,811,263 | null | [
41812594,
41812496
] | null | null |
41,812,472 | comment | Maxatar | 2024-10-11T19:07:53 | null | What is inherently wrong with taking a pill for weight loss?<p>I can understand being cautious about side-effects or efficacy, but if a pill existed that resulted in weight loss and was side-effect free, I can't imagine what would be inherently wrong about taking it.<p>Would it be preferable if it were a powder, or a cream? Is it the specific delivery format that has you concerned? | null | null | 41,812,440 | 41,811,263 | null | null | null | null |
41,812,473 | comment | Sophira | 2024-10-11T19:07:55 | null | Are you certain? The last I heard about it from Vivaldi[0], they were only going to keep the MV2 code around so long as it's in the upstream codebase:<p>> We will keep Manifest v2 for as long as it’s still available in Chromium. We expect to drop support in June 2025, but we may maintain it longer or be forced to drop support for it sooner, depending on the precise nature of the changes to the code.<p>Note that June 2025 is the same date Google plans to drop support completely[1].<p>[0] <a href="https://vivaldi.com/blog/manifest-v3-update-vivaldi-is-future-proofed-with-its-built-in-functionality/" rel="nofollow">https://vivaldi.com/blog/manifest-v3-update-vivaldi-is-futur...</a><p>[1] <a href="https://developer.chrome.com/docs/extensions/develop/migrate/mv2-deprecation-timeline" rel="nofollow">https://developer.chrome.com/docs/extensions/develop/migrate...</a> | null | null | 41,810,602 | 41,809,698 | null | [
41812748,
41812787
] | null | null |
41,812,474 | comment | mrtz | 2024-10-11T19:07:59 | null | Cross-compilling from MacOS to Linux. By default OCaml does dynamic linking of all C dependencies. Even with static linking there is a dependency on libc (or the MacOS equivalent) as far as I remember. I had some success so far with Rust, Nim and Go using a musl toolchain[1], but no luck for OCaml. At the moment I'm just using a docker container which mirrors the Debian distribution running on the Pi.<p>[1] <a href="https://github.com/messense/homebrew-macos-cross-toolchains">https://github.com/messense/homebrew-macos-cross-toolchains</a> | null | null | 41,808,175 | 41,775,641 | null | null | null | null |
41,812,475 | comment | rpmisms | 2024-10-11T19:07:59 | null | An emergency joystick would be pretty cool. Actually, a joystick to give the driving AI input would be great too. "Maybe squeeze a little farther left here". "Slow it up, the road is sketchy here". "It's safe, drive faster". | null | null | 41,805,821 | 41,805,706 | null | null | null | null |
41,812,476 | comment | lcnPylGDnU4H9OF | 2024-10-11T19:08:00 | null | > false claim<p>Just to add some more context. From the linked article:<p>> Experts contacted by NPR disagree with the company’s statement. The water is being used to cool the launch pad as Starship’s engines fire. While drinking water may be used in the system, after it comes into contact with the rocket exhaust, it contains high levels of dissolved solids and potentially toxic chemicals like zinc and hexavalent chromium, according to the license application submitted by SpaceX to Texas regulators.<p>From the SpaceX post[0]:<p>> Outflow water has been sampled after every use of the system and consistently shows negligible traces of any contaminants, and specifically, that all levels have remained below standards for all state permits that would authorize discharge.<p>There doesn't seem to be any reason to suspect SpaceX's claims about the contamination so it sounds like they're not running afoul of any state regulations. It's worth noting that they did pay a related fine to the EPA but that doesn't appear to be related to their practices:<p>> The subsequent fines levied on SpaceX by TCEQ and the EPA are entirely tied to disagreements over paperwork. We chose to settle so that we can focus our energy on completing the missions and commitments that we have made to the U.S. government, commercial customers, and ourselves. | null | null | 41,812,285 | 41,811,032 | null | [
41812981
] | null | null |
41,812,477 | comment | tredre3 | 2024-10-11T19:08:01 | null | uBlock lite can accomplish 2, 3, 4 completely. It's only #1 and #5 that are truly affected by v3. But those two were already pretty limited in chrominum browsers compared to firefox. | null | null | 41,810,478 | 41,809,698 | null | null | null | null |
41,812,478 | comment | bad_user | 2024-10-11T19:08:04 | null | Security and possibly performance, which is the selling point of MV3. | null | null | 41,812,158 | 41,809,698 | null | [
41812852
] | null | null |
41,812,479 | comment | borroka | 2024-10-11T19:08:10 | null | Given that a few decades ago obesity and overweight rates were nowhere near what they are today, this shows that a large part of the population is weak, fragile, and not very interested in their well-being.<p>I want to emphasize that a few decades ago, people were much thinner in the Western world and did not hate their lives because they could not eat a triple cheeseburger, go hungry constantly, or feel physically deprived. Those were my parents and my grandparents, I know them.<p>But if you show them hyper-caloric food that makes them feel like crap, they can't say no. It's disappointing. And the same can be said for addiction to social media, horrible TV series, and constant music everywhere. | null | null | 41,812,006 | 41,811,263 | null | [
41812833,
41813033
] | null | null |
41,812,480 | comment | BlueTemplar | 2024-10-11T19:08:14 | null | Right. Pretty sure it's illegal under EU law(s), and people were already condemned for it (but yes, in case ill intent was proven) - why wouldn't it be illegal under US law - it's basically akin to vandalism ?<p>(In other news, the Internet Archive got DDoSed today :( | null | null | 41,809,535 | 41,797,719 | null | null | null | null |
41,812,481 | comment | 4k93n2 | 2024-10-11T19:08:15 | null | less decision fatigue | null | null | 41,810,077 | 41,808,943 | null | null | null | null |
41,812,482 | comment | sweeter | 2024-10-11T19:08:19 | null | the UK has basically become the US by most metrics. This includes the increasing privatization of health services, transport, etc... and the excessive commodification of basic necessities like housing.<p>I find that saying that health initiatives don't work by vaguely gesturing at a country, is not a structurally sound argument. Its like the sentiment here is: "is the fact that we include Pizza as a vegetable in American schools part of the problem? Nooooo, that can't be it. it must be a moral issue!" and thats just one example.<p>The obesity problem in the US is tied directly to our relationship with highly processed (and CHEAP) food. Along with the stranglehold those companies have over state and federal institutions that allow them to directly sell these foods in schools and institutions, and heavily skirt FDA regulations via lobbying.<p>The US is uniquely bad when we have a ton of chemicals and ingredients in our foods that are banned in most other countries. It is largely a systemic problem and a problem that can easily be solved. Poorer people tend to eat cheap food, cheap processed food isn't well regulated and is directly tied obesity and a whole host of health problems. | null | null | 41,812,281 | 41,811,263 | null | null | null | null |
41,812,483 | comment | valval | 2024-10-11T19:08:24 | null | Thats not a thing though is it? Sanctions aren’t affecting Russia in any capacity. | null | null | 41,811,636 | 41,807,681 | null | null | null | null |
41,812,484 | comment | pfdietz | 2024-10-11T19:08:25 | null | It's the "git gud" of diet advice!<p>I wonder if GLP-1 agonists are going to break gaming addiction too. | null | null | 41,812,009 | 41,811,263 | null | null | null | null |
41,812,485 | comment | karlgkk | 2024-10-11T19:08:35 | null | > Cruise had driverless robotaxis on the streets while they had 2.5 to 5 miles per intervention. [1]<p>And they got (rightfully) pulled off the roads.<p>> there are still some rough edges in rare cases<p>Waymo has substantially lower interventions.<p>And, they have a huge fleet of humans running around the city attending to the cars. A few weeks ago I saw one stuck - whoever had last used it, managed to trap a seatbelt in the door. It was sitting for about 5 minutes when a guy pulled up and fixed it and sent it on its way. I'm not saying Tesla can't build that, but they're going to have to. | null | null | 41,805,885 | 41,805,706 | null | null | null | null |
41,812,486 | comment | kylehotchkiss | 2024-10-11T19:08:47 | null | > they enter homes and kill babies, attack children and adults and carry rabies epidemics<p>Are you talking about monkeys or dogs? I guess if you were talking about monkeys you forgot about how they raid the fridge and pretend to do laundry sometimes. | null | null | 41,795,953 | 41,795,218 | null | null | null | null |
41,812,487 | comment | clickzyn | 2024-10-11T19:08:48 | null | It's a good idea; it might be a good idea to put it on Khan Academy to obtain more visibility.<p>By the way, is it possible to be notified when the ebook will be available? I'm an indiepreneur, and I'm often overwhelmed. | null | null | 41,766,524 | 41,766,524 | null | null | null | null |
41,812,488 | comment | jasonhong | 2024-10-11T19:08:49 | null | I'm not an expert in AR and VR, but I can point you to papers by two of my colleagues who know a lot about this space.<p>David Lindlbauer is a faculty at CMU who applies a lot of perceptual psych to his research on VR.
<a href="https://scholar.google.com/scholar?hl=en&as_sdt=0%2C39&q=david+lindlbauer&btnG=&oq=david+lindl" rel="nofollow">https://scholar.google.com/scholar?hl=en&as_sdt=0%2C39&q=dav...</a><p>Roberta Klatzky is a perceptual psychologist that has done a lot of work on haptics. One of her ongoing projects is augmented cognition through wearables, e.g. giving people instructions in heads up displays based on the current state of things (e.g. it looks like you successfully removed the lug nuts, here's your next step in changing the car tire).
<a href="https://scholar.google.com/scholar?hl=en&as_sdt=0%2C39&q=roberta+klatzky&btnG=&oq=roberta+klat" rel="nofollow">https://scholar.google.com/scholar?hl=en&as_sdt=0%2C39&q=rob...</a> | null | null | 41,806,761 | 41,780,328 | null | null | null | null |
41,812,489 | comment | yapyap | 2024-10-11T19:08:51 | null | gonna be a good while | null | null | 41,811,263 | 41,811,263 | null | null | null | null |
41,812,490 | comment | crooked-v | 2024-10-11T19:09:01 | null | > that we should "all be on."<p>Who, exactly, has claimed that? | null | null | 41,812,273 | 41,811,263 | null | null | null | null |
41,812,491 | comment | llamaimperative | 2024-10-11T19:09:06 | null | Pointing out the obvious: the conversation isn't about GP and their own time allocation. It's about the empirically observable fact that we've built a civilization that turns people into lethargic, sedentary, chronically afflicted mush. | null | null | 41,812,468 | 41,811,263 | null | null | null | null |
41,812,492 | comment | opo | 2024-10-11T19:09:11 | null | >...Ozempic is for type-1 diabetes<p>At the present time, Ozempic is not approved for Type 1 diabetes:<p>>...Ozempic® is not for use in people with type 1 diabetes.<p><a href="https://www.ozempic.com" rel="nofollow">https://www.ozempic.com</a><p>Compare with Type 2, with Type 1 diabetes there are other risks that could occur:<p>>...While medications such as GLP-1 receptor agonists (Ozempic, Wegovy) and SGLT-2 inhibitors (Jardiance, Farxiga) demonstrated powerful benefits, they quickly were determined to pose too much of a liability for pharmaceutical companies or regulators due to concerns about safety. Specifically, GLP-1s can increase the risk of hypoglycemia (low blood sugar) and SGLT-2s can raise the risk of a serious, life-threatening complication called diabetic ketoacidosis (DKA).<p><a href="https://diatribe.org/diabetes-medications/why-diabetes-miracle-drugs-ozempic-are-limits-type-1s" rel="nofollow">https://diatribe.org/diabetes-medications/why-diabetes-mirac...</a> | null | null | 41,812,106 | 41,811,263 | null | null | null | null |
41,812,493 | comment | robohydrate | 2024-10-11T19:09:12 | null | I've been on tirzepatide (Mounjaro) for 4 months now. I'm down 13% of my body weight. I realized that frequent cannabis consumption interferes with the weight loss, so I've kicked the habit from daily to occasionally on weekends. I've started walking 2-3 miles a day, 2-3 days a week regularly, in addition to eating less and being more motivated to calorie count.<p>All this to say, this drug has been life changing for me. I spend more time doing things I want to do, depression and anxiety have less of a hold on me now. I feel that this drug has allowed me to be the best version of myself I have been in a long time. The only side effects so far have been positive. I do worry about what I will do once it's time to titrate off the weekly dose and the best I can think of is that the habits I'm forming in the time on the drug I will have the resolve to continue after cessation.<p>I say this because I have battled depression, anxiety and obesity issues my entire life. I've had many failed attempts at getting back to a healthy, productive and non-obese lifestyle. I don't know what is so different about having the drug help me, but I can tell you that it has been different. | null | null | 41,811,263 | 41,811,263 | null | [
41812565,
41813239,
41813078,
41812839,
41812641,
41813104,
41812939,
41812532,
41812793
] | null | null |
41,812,494 | comment | grahamj | 2024-10-11T19:09:16 | null | “There is now sufficient reason for greater optimism”<p>- Automated Personnel Unit 3947 | null | null | 41,809,358 | 41,799,068 | null | null | null | null |
41,812,495 | comment | null | 2024-10-11T19:09:22 | null | null | null | null | 41,812,454 | 41,811,078 | null | null | true | null |
41,812,496 | comment | rootusrootus | 2024-10-11T19:09:26 | null | That is a simplistic description which is superficially true. The body, however, is quite a bit more complicated than that. <i>Especially</i> when you get to that wrinkly pink lump in your skull. | null | null | 41,812,471 | 41,811,263 | null | null | null | null |
41,812,497 | comment | mrweasel | 2024-10-11T19:09:26 | null | Yes, sorry, I misread when I looked it up. You're correct, it's for type 2. | null | null | 41,812,460 | 41,811,263 | null | null | null | null |
41,812,498 | comment | detourdog | 2024-10-11T19:09:42 | null | We probably disagree that they are built differently. I’m willing to let you believe they are seperate and ask that you let
Me hold my ideals. If you want to discuss you would have to state what the differences are. Declaring they are different and expecting me to know why you think that is a real hindrance.<p>It is not obvious to me. | null | null | 41,811,420 | 41,776,631 | null | [
41813077
] | null | null |
41,812,499 | comment | kylehotchkiss | 2024-10-11T19:09:43 | null | For better and for worse this man left a massive impact on his country, and in many others surrounding.<p>Maybe my favorite is their JV with Starbucks, which had just enough Indian touch to be special. I'll take an Indian Starbucks location over an American one anyday. RIP. | null | null | 41,795,218 | 41,795,218 | null | null | null | null |
Subsets and Splits