One thing a coworker said once that I think about a lot: ever read an article about a subject that you know a bit about, and invariably you come to the conclusion that the writer doesn't really have a good grasp on what they're talking about. Now think about all the articles you read about subjects that you don't know much about, why would the accuracy be any higher on those ones?
I've been thinking about this recently with internet comments on here, reddit, etc. There are very few topics on which I'd consider myself an expert, but whenever one comes up, the "top" comment (often something contrarian/snarky) is always significantly incorrect.
100%. It's extra disastrous when the topic is polarizing or charged. The topic of US politics for example will very often work off of some false dichotomy or fundamentally flawed premise and twist into some horribly deformed conversation.
Try telling someone that "the process of voting doesn't work that way" or try to clear up some common misconception about the economy or crime and it just doesn't go anywhere. People have a lot of false assurance to back up their mode of thinking and it is nearly impossible to break that. In fact, there is an entire economy in vindicating people's beliefs which makes people even more assured.
I'm convinced the only proper stance on complex ideas, concepts and topics is that we shouldn't have the hubris to think that we completely understand something or even have a solid grasp.
Keep in mind that people don't talk about that which they believe they completely understand. When someone reaches a satisfiable conclusion in understanding, they lose interest in the subject and move on to something else.
The political discourse around polarizing topics isn't disastrous. It takes place exactly because people realize that they are complex issues not fully understood, and are talking about it in hopes that more information will come to light to help them reach a greater understanding. This is why the more complex the issue, the more it will keep coming up over and over.
human beings are bad at gathering information, inferring the right things from it, and responsibly passing it on to others. It is incredible what we’ve achieved in spite of this
All was achieved when it was widely not so, and by people not affected by this.
Thank you. This is a refreshing take on much of human interaction.
As a teacher in middle school I see this demonstrated throughout my classes. I have even started collecting data points on how many times a week a topic or topic adjacent to another topic crops up as discussions in my class. There are almost always heated discussions, but in the end we typically come to an understanding about facts or how we can’t truly know fully situations. There are of course some holds outs, but for the majority we find understanding or compromise.
With regards to politics, wasn't that found to be a Russian active measures tactic where they'd get low quality or inaccurate comments up voted to the top, down vote actual informative comments, and flood their up voted comments with low quality replies and threads to push the informative stuff so low that users likely will give up scrolling before they see it?
They just enhance it, the division and polarisation already exist, what Russia has learned is that if they press those buttons hard enough they make our democracies ungovernable.
They aren't causing it, just furthering our issues for their own goals, and unfortunately it's working pretty well.
I think this is a significant contribution to imposter syndrome. It took me a long time in life to realize that the confidence some people have in their opinions rarely comes from true expertise; rather, it comes from their personality.
Are you claiming expertise in computing or non-computing topics? HN tends to be - as we might expect - quite good at computers and bad/average [0] at everything else.
I'd assume it is the same logic as comparative advantages in economics; it doesn't make sense for communities to become experts in everything. The only caveat is that people in the chattering communities (eg, journalists, influencers, celebrities) are some of the last to look to for informed opinions on reality since their speciality is attracting attention and telling stories rather than anything linked to success in the physical or academic worlds. They're often clever, just not involved in complex topics.
[0] average = bad. Goes to show how catastrophic dictatorships are that a democracy can consistently outperform one.
Democracies function on the notion that most people have no idea what they're talking about, and vote randomly. The few who do know what they're talking about will all vote the same way, so most of the time the right answer will prevail.
There are a lot of problems with that assumption. Thus far it seems to work out better than the alternatives, but we'll have to see if that continues to hold.
I don’t think that’s true. I’m pretty sure the hope is that people are reasonably well informed. And anyway, we’re voting on representatives, not specific issues, for the most part. (I mean clearly referenda exist, but they aren’t the main thing). Or we’re voting on matters of preference, in which case the populace is essentially correct whatever their decision.
Yes, but false reporting isn't coming from just one side is it, it's something all sides can participate in. The way the education system is right now along with societal expectations, people are going to be misinformed rather than admitting to be uninformed.
In the absolute sense yes, but this thread was framed in comparison to undemocratic systems. The media aren't biased in the direction of making their host country worse off, so their biases aren't negative in that sense. We'd get better outcomes if people in the public discourse held themselves to higher standards; but their low standards don't stop the tendency of marginal voters to bias on rational decision making.
The media aren't biased in the direction of making their host country worse off
i don't believe that. at least western media are biased towards majority and conservative views and for profit entities, ignoring or even suppressing minorities and that is making us worse off.
Conservative views are things that worked in the past though. For profit entities are all dedicated to satisfying the needs and wants of people and are part of society too. >90% of all the gains since the industrial revolution came from for-profit entities so it is a stretch to say they are biased in favour of making things worse. Minorities are, by definition, not a group that includes most people so things can get really brutal for them before it starts hurting the greater society.
These are all biases that may be politically undesirable to you (the corporate media are basically the vanguard of class warfare, so they should be undesirable to a bunch of people), but they aren't biased in the direction of making things worse at the highest level of abstraction. A rising tide benefits all ships.
That is indeed one of the problems I mentioned in the last paragraph. We are stressing the limits of how much misinformation democracy can handle. We might well be over it.
> HN tends to be - as we might expect - quite good at computers and bad/average [0] at everything else.
Who's we? I don't expect that at all. "computers" is such a broad subject, with countless areas of specialty, and in my experience HN commenters are often, all too often, quite ignorant of my particular areas of specialty, though that doesn't stop them from overconfidentally asserting falsehoods about those subjects.
Knoll's Law of Media Accuracy: "Everything you read in the newspapers is absolutely true except for the rare story of which you happen to have firsthand knowledge."[0]
This is why I'm rather bemused when people "freak out" about LLMs hallucination rates. The number of comments I've seen on Hacker news (ostensibly the intellectual sister of reddit) that make extraordinarily confident outlandish claims about various subjects with zero citations, sources, etc across what would otherwise be considered highly scientific disciplines - neuroscience is a particularly egregious offender - is far more worrisome to me than the latest hallucination from Gemini.
My experience working at Apple on both Music and Maps: anything that I had worked on that was reported on was usually wildly inaccurate. Mainstream news, bloggers, HN comments... so much assumption and usually way off.
Covid was the litmus test for me. I happened to know a lot about a lot of topics that became wildly reported on, from virology to immunology, and I had to watch in horror as every single media presence and expert weighed in with at best non-existent data to support major changes to existing wisdom, and at worst (and most common) obvious lies that contradicted not just available science but the very nature of how viruses operate.
For me, it’s often a case of semantics. Like, if I watch a YouTube car repair video, and the presenter is diagnosing an issue and says something like “ok, so we can see here the battery is draining far too quickly so I better check the alternator first,” in my experience, it’s rarely the alternator and often either a bad battery or a short somewhere. My first check is either of those two things. I might eventually arrive at the alternator but it’s all just semantics and interpretation at a certain level. I don’t feel like they’re breathtakingly _wrong_, they just probably have had a different experience than me, leading them to a different conclusion.
Caveat: I often work on old cars where a short is far more likely.
The accuracy of other articles can be higher or lower because other people researched and wrote them. This is why people tend to have favorite reporters and favorite news outlets. They do not trust every thing they read, they look for people who seem to reliably report the truth as they know it, and those people/outlets they _tend_ to trust. People are naturally skeptical. Often they just read to entertain themselves or to gather topics to discuss with friends or coworkers later. That doesn't mean they believe it all. You can read all the UFO articles but that doesn't mean you believe in UFO's. You might instead find them interesting as a sign that the military ignoring UFO's mean enemy spy drones can easily sneak in and spy on maneuvers, a big vulnerability that the military is _supposed_ to be cognizant of, but is apparently ignoring.
I never really knew about, or thought about, this phenomenon until I started trying code generated by ChatGPT. Then I started to realize ChatGPT's answers seemed really good for programming languages I knew little about, but pretty bad for programming languages I knew a lot about.
Then I realized this probably applies to many different subjects.
> Now think about all the articles you read about subjects that you don't know much about, why would the accuracy be any higher on those ones?
Why wouldn't the accuracy be higher? Extrapolating that way from limited experience is not something I would personally do. I would read everything, learn gradually, and continue learning and testing information with experiments and more information.
In the context of articles the answer is simple: because the journalist isnt the specialist, they likely just interviewed a few people and wrapped it up the article.
For comments it's a different story. You do occasionally get the real experts, but more often then not, you just get another mediocre human like myself that just voiced their opinion. Furthermore, there is a strong correlation with experts not commenting as they've got better things to do vs the average Joe that is commenting with little effort, essentially outputting significantly more comments.
Another detrimental factor is that mentally unwell people will often output orders of magnitude more comments then everyone else, complicating everything even more.
OP has a point. It depends on your frame. Some areas require far more specialized knowledge than others. If your specialization were something incredibly complex like quantum physics, you'd probably find ridiculous mistakes in 99% of mainstream sources. OTOH, articles about business or tax law may have far fewer lay errors, so you would be mistaken to project the popular information error rate in your own field onto popular reporting in every other field.
At least in theory, while the journalist isn't a subject matter expert, they have spent time reporting on the same domain. They know the people who are experts, and work under an editor who has worked in that field for decades. They should be able to validate any facts and be able to put them into context.
As newspapers have gotten hammered there are fewer and fewer people who actually do that, and more and more bloggers with loud opinions and few facts. Many good sources have gone defunct or given up trying. But there do remain a few sources of decent journalism, where the reporters and editor really are better informed than most laymen.
> You do occasionally get the real experts, but more often then not, you just get another mediocre human like myself that just voiced their opinion
Most of the comments experts make are just stating their own opinions as well. Take any credible expert in any field and you'll find equally credible experts who completely disagree with them. When experts are communicating with non-experts they're often really bad at explaining which points are just their opinions versus more widely accepted facts, and just as bad as explaining what other equally credible expert perspectives exist on a topic other than their own. They also tend to be pretty bad at explaining where the limits of knowledge in their field are as well.
I would suggest that the ego boost of being a highly regarded expert also tends to make these shortcomings worse with a lot of people, that increasing credibility and reputation can have the effect of increasing dogmatic-ness. Of the topics that I'm somewhat well informed about, I know very few "experts" who are good at explaining the perspectives of their field in a reasonably balanced way, I think Roger Penrose is quite good at doing this, but even he's not perfect and experts with his level of humility are not very common.
Sure. I mean, ask most coders what the best programming language or web framework is for a certain job, and they'll tell you it's whichever one they use. It's partly confirmation bias from their workplace, and partly "when all you have is a hammer..."
I have found people receptive to that observation though.
That is, it's not yet a widely noticed thing and I feel it helps to spread it around and people generally seem to welcome and understand the concept. It probably does not immediately change how they read - biases are very hard to fight. But it goes in the right direction.
Absolutely true, but also people who know about a subject often can't see the forest for the trees and an outside perspective brings new detail you haven't seen before.
> As someone that works in the crypto industry, I feel this every time I see the news or articles report on it, or comments here talking about it.
It was the same way during the 90's Dot-Com era, journalists would explain the World Wide Web with word salad, e.g., the "Graphical portion of the Internet".
With regard to Crypto, I decline to discuss it without mentioning the 3Blue1Brown video, sometimes the gambit works and triggers a better discussion: https://www.youtube.com/watch?v=bBC-nXj3Ng4
And I love the comments in that video about being "Confused, but on a higher level".
No I'm saying this is a frankly stupid argument to make. Crypto has almost a century of financial crimes, history, and policy to draw on. The fact that they're stepping on the same rakes despite them being labeled is a point against the entire sector. Anyone smart enough to build an alternative financial system is probably doing more legitimate work in the actual financial sector, leaving crypto to 29 year old geniuses who barely understand they're running ponzi schemes.
You're arguing that we should try out this new, unregulated, and terrible alternative financial system that seems to only be filled with such trustworthy people as scam artists, terrorists, drug dealers, and war criminals because a system designed before computers existed sucked ass. Your audacity is shocking.
A blockchain is a distributed public ledger of immutable information. The use cases for that seem clear and obvious: consider all repositories of public information that are either not distributed, and therfore subject to single points of failure, or allow extant data to be modified, and therefore subject to reliability and trust issues. Consider how blockchains might improve all such use cases.
You still have to trust whatever things you rely on upstreams, plus, if the consensus is against you, there's nothing you can do anyway? For example, you got scammed, its written to the ledger. The transaction is immutable and actually, actually did happen. What can you do now?
Large cryptocurrency networks like Bitcoin are extremely hard to execute a 51% attack upon due to their size. Though ultimately there's no perfect solution. As the saying goes, democracy is 2 wolves and 1 sheep deciding on what's for dinner.
But you can't use bitcoin or other blockchains with large cash flows as information storage at scale, since they use them to make transactions first and foremost and can't hold a lot of other data.
For many applications, it's possible to commit just a Merkle tree root hash to the blockchain and still get all the properties you care about, while only using O(1) blockchain storage.
I take the above comment as addressing a use case for blockchain that isn't currency.
The context I think they are proposing blockchain to be useful is in verifiable public information. For information to be verified or analyzed the blockchain can verify the source even offline.
Exactly. E.g. decentralized DNS, title transfers to real property, etc.
It's still extremely useful as currency as well, though, precisely because it actually is digital currency, and allows direct person-to-person transactions via the internet without requiring a proprietary middleman. The previous commenter's complaint about how it doesn't solve anything if you get scammed is a bit off-target, because dealing with frauds and scams isn't a feature inherent in the currency itself -- it applies equally to any payment solution, especially cash and checks -- and is something we deal with via the legal system regardless of how funds are transferred.
> A blockchain is a distributed public ledger of immutable information.
Is it immutable? Forks, 51% attacks, bugs. It it only immutable until consensus decides it isn't immutable, right? If something is immutable, but can be made immutable, is it ever really immutable?
Blockchains facilitate trustless transactions enabling truly decentralized currencies not under the control of any government. Many people find this quite valuable, as evident by recent trends in authoritarian overreach, fiscal repression, and inflation. Privacy optimized cryptocurrencies like Monero also deal with the problem of their blockchain 'providing evidence in court.'
A state doesn’t insure the value of a currency. In order, the largest influences on currency value are:
1. Speculation / belief
2. Approx 10x less influential is trade, most currency transactions both by volume and by value are in speculation which is underpinned and driven by beliefs in a particular currency. Some estimates put trading as low as 2.5% of all forex. Trading does impact the price of a currency, but in miniscule proportions to speculation
3. We can fairly confidently say tax collection is a lower influence than trade because of so many examples (and extraordinarily rare counter examples) but I don’t know of any reasonable way to quantify how much less the effect is. The threat of incarceration and / or asset seizure drives demand for currency which influences its value.
Places like Argentina are fascinating (said while acknowledging the extreme hardship being placed on people who have to use Argentinian currency!). In recent history all of the above have had visible influence on their currency along with pegging / dollarization which attempts to introduce a control on all the above factors with various degrees of success.
The value of Gold is based on scarcity dictated by the laws of Physics, and on people believing it has value. The fascination with that yellow gleam goes as far back as recorded time. The fact that it's easy to work with (malleability, ductility, etc.) probably played a part as well.
The value of Crypto products is based on scarcity dictated by the computationally intensive nature of certain algorithms, and on people believing those products have value.
Gold isn't a currency. Gold coins can be, but then you either have to trust the mint or you carry around a scale. The key is being fungible and it's the state that promises that a dollar bill is equal to any other dollar bill.
That's why I prefer to read books by scientists that are established in the field. Granted, I might miss some spectacular insights that some revolutionists might propose, but I can be relatively sure that most of what I read is actually true in the sense of representing reality relatively accurately.
I described a time where a friend summarized this effect and how it stuck with me since, which is the same effect described in the article? And saying that it's a bummer to think about? I didn't say it was original insight on anyone's part, just that it was a sticky thought when it was wrapped up in a small 1 sentence explanation. You're right, there's no great insight in my comment, all it said was "yea this happens to me too all the time" but for some reason it got upvoted.
Unbelievable how meta the original comment is. Seems like many of the subsequent replies, too. Skim a few paragraphs and that's plenty, time to share your thoughts with the internet...
This passage comes to mind:
> we want so badly to be believed, to be seen as someone who knows stuff.
> Skim a few paragraphs and that's plenty, time to share your thoughts with the internet...
Some articles posted to HN have poor enough quality that reading a few paragraphs is all they deserve. The false premise is usually evident from the first paragraphs, and there's no reason to read further (unless it's one of those articles where the author changes their mind through a long exposition).
> Please don't comment on whether someone read an article. "Did you even read the article? It mentions that" can be shortened to "The article mentions that".
Normally yes, but I think an exception is allowed in this case due to the irony of the person being oblivious to an article talking about people being oblivious.
For us old timers this is nothing new. For people that know what a 3 digit /. userid is, the meme of not reading the article is as old as internet news boards.
It's like when we see people "do programming" in TV shows/movies and laugh at how goofy it is. I suspect most technical professions are exactly like that, we just don't realize it most of the time.
Even worse when there are slightly creative solutions that haven't reached conventional wisdom status yet. They're out there, but almost impossible to find.
"Briefly stated, the Gell-Mann Amnesia effect is as follows. You open the newspaper to an article on some subject you know well. In Murray's case, physics. In mine, show business. You read the article and see the journalist has absolutely no understanding of either the facts or the issues. Often, the article is so wrong it actually presents the story backward—reversing cause and effect. I call these the "wet streets cause rain" stories. Paper's full of them.
In any case, you read with exasperation or amusement the multiple errors in a story, and then turn the page to national or international affairs, and read as if the rest of the newspaper was somehow more accurate about Palestine than the baloney you just read. You turn the page, and forget what you know."
– Michael Crichton (1942-2008)
Since I've started working in finance, I find all those conspiracies about banks and "wall street" a bit cringe: HSBC is actually a nice bank, they helped me a lot, hedge funds do lose as much money as they make in aggregate across all of them because end of the day it's all a random 0-sum with losers paying winners, inside an investment bank, risk is very managed and boring, deal makers spend lots of time trying to reject proposals, loans are rarely predatory and not always easy to get because nobody wants bad payers, bonuses are scarse, etc.
The headlines about some crisis with uneducated comments about how rich bankers exploit the weaks make me hesitate to trust journalism. Often incompetence is way more prevalent than nefariousness ...
Each time I say stuff like that I get comments saying "maybe but it's not a reason to defend banks", completely missing that it's exactly my point: this hatred of banks is a bit irrational, I think it's fine sometimes to defend them...
The flip side of this is that I’ve repeatedly seen people in some industry subject to new regulation claim that some effect of it is an unfortunate accident (when it’s actually the point, the behavior it’s stopping that you think is good is actually unethical and system-wide contributes to huge problems in ways that ought not be hard to understand) or that much-needed regulation is some terrible mistake that people will regret when we have strong evidence (like, real world examples) to the contrary.
Granted I’ve mostly seen that in the real estate world, so maybe other industries are less-shit about saying plainly dumb stuff based on “insider knowledge” (self-serving bullshit, actually)
See also: cops and all manner of things, like the relative riskiness of their own job (and which parts of it are risky—all that driving is a lot of the risk, and Covid-19 vs guns is a fun comparison for some years) or how dangerous it is to be within one meter of a small amount of fentanyl or whatever, or how we’ll all just super regret it if we keep them from trampling on civil liberties (common view, not making that up)
Sometimes the insiders know better and everyone else is wrong. Sometimes the insiders are all sorts of mixed up over some comically basic shit.
"loans are rarely predatory" misses the point why they are bad. They are bad for the economy because it's you (not the market) who decides who gets resources.
They weren't, they were likely invented in Ancient Rome, and soon recognized as harmful, and were illegal until recently, when they got allowed, and again recognized as harmful during the great depression (the Chicago plan that was never implemented).
The problem is that you only see the gains that those who take a loan get, but it only allows them to circumvent the market, and nothing is in fact created, only the resources get misallocated. Somebody else would use what got bought with the money.
Loans existed for at least as long as writing has, and probably for the entirety of humanity’s history. Code of Hammurabi had laws regulating loans, for example.
Why loans result in resources being misallocated? This doesn’t make sense to me. Say, I have some cash I want to save for retirement. Until then, I loan it to other people to run their businesses. Where is misallocation here?
They buy something for the money. That is the misallocation. They couldn't afford it, and your loan allowed it. Let's say they buy a tractor. The tractor would be bought by somebody else. Or perhaps different machinery would be manufactured, or perhaps something completely different would be done, or nothing at all would be done because people would prefer to have more free time. This is how you get massive market imbalances that lead to the great depression, or Evergrande.
No money is known that was used before classical antiquity, it must be a mistranslation.
Someone buying a tractor on a loan doesn’t take away someone else’s ability to buy a tractor. We have no shortage of tractors. Practically all new tractors today are bought on a loan, by the way.
Imagine what would happen if loans on tractors would not be allowed. Then, someone who needs a tractor would need to rent it, instead of buying it. In that world, pretty much the same people who use tractors now would also use tractors. Tractors use would be similarly allocated as today. The main difference is that instead of a bank earning income by renting you money, it would earn income by renting you tractors. The ultimate outcome would be effectively the same.
Loans is just renting, but instead of renting physical objects, you rent money. If you object to renting money, all your objections should apply to renting physical things. Have you thought of that?
Money isn't a resource that is objectively needed, its role is to regulate economy. You could do anything that you do just as well without it, if people could agree on it. It isn't like a tractor that is needed, and there is no problem with renting tractors. The problem is that you do use money in your example: you use it to store your savings. By lending it,it gets used twice - once to buy the tractor, the second time it's still used to store your savings.
It does take away somebody's ability to buy a tractor, as you buy one, so there is one less, or one extra needs to be manufactured. That can also be a problem - too many tractors can be created, so that food gets so cheap that it's basically free,and the farmers now can't make money. Which is what literally happened in the great deoression.
And now just think of all of the people who will be getting their knowledge from LLMs which are literally making up stuff through statistical linguistic inference on a grand scale from hearsay.
> LLMs are literally making up stuff through statistical linguistic inference on a grand scale from hearsay.
Hearsay being some personal "truth" it may be useful to know what the statistical average of that "truth" is. If we do it right perhaps we can get the various personal errors to cancel out.
Maybe you are referring to articles on the internet that are meant for SEO?
The internal white papers I've written and also help co-author (which are for discourse/dissertation on successful project implementations and to be used as reference for future work), or when on a stint with IBM doing ITSO Redbooks, were (are?) highly accurate, and involved subject matter experts (which, as the name would imply, were the SMEs that people would contact specifically as THE experts in that field; and, back then, we were really the experts).
OK Ok, you can stay on my lawn for a while longer... /j
Kind of a bummer to think about.