19 ms·

It's only scary at the beginning. Then you get used to it. Every single social media site - including HN - has uninformed people agreeing that a correct-soundin

by speff 3y ago

It's only scary at the beginning. Then you get used to it. Every single social media site - including HN - has uninformed people agreeing that a correct-sounding answer must be right. My friend the tax accountant gets downvoted for clarifying how taxes actually work. My wife the linguist gets downvotes for explaining no that's not how language works. It's not scary - it's typical.

jodrellblank 3y ago
[flagged]
- speff 3y ago
  To be honest, I was debating on even posting my message initially. It was off topic but I figured it would've been ignored. If I knew it was going to derail the discussion about Stratoscope's analysis that much, I wouldn't've posted it. Edit: Also grammar nazi'ing has little to do with linguistics and more to do with being a jerk...usually.
tcgv 3y ago
It is the Gell-Mann Amnesia effect, but on social media: > The phenomenon of people trusting newspapers for topics which they are not knowledgeable about, despite recognizing them to be extremely inaccurate on certain topics which they are knowledgeable about.
spicyusername 3y ago
Is it not possible to be both scary and typical?
0ckpuppet 3y ago
whoe political movements are built on this kind of momentum
deadbabe 3y ago
Hackernews is similar to ChatGPT in that regard. Lots of correct sounding answers that are really just a word salad.
MenhirMike 3y ago
> my friend the tax accountant gets downvoted for clarifying how taxes actually work. Let me guess: Tax brackets? That's the one thing that most regular workers in the US just don't seem to understand (and arguably, many people knowingly spread falsehoods to further some agenda).
- speff 3y ago
  Decent guess, but nah. Something to do with corporate tax accounting. Can't remember the details because that's out of my element.
  - tpmoney 3y ago
    Tax write offs would be my guess. Every employee of charities that partner with retail locations for PoS donations must die a little inside each time some fool confidently asserts that they never give to those charities because it's all a scam. The money, they will assert with the confidence only someone so wrong can muster, is just used for write offs so the executives can have a big bonus and the company gets to claim they donated all the money. Bonus points if they assert that the charity doesn't even get the donated amount.
- PopAlongKid 3y ago
  I think the basic thing about taxes that is least understood is the difference between gross income and taxable income (the latter is the amount that tax brackets apply to). A close second is the difference between tax liability, and refund/balance due on the tax return.
  - sgerenser 3y ago
    Just try to convince the average person that “getting a big refund” is a bad thing, since it means you gave the U.S. government an interest free loan.
    - rvnx 3y ago
      (Not from the US) Why is it a good thing to lend for free to the US gov ? Because the regional banks aren’t that stable ?
nsagent 3y ago
I think this also partly explains the LLM hype — people can be as confidently incorrect as LLMs, or maybe LLMs are as confidently incorrect as humans since they are trained on text from social media.
- wolverine876 3y ago
  The LLM output matches what the crowd expects.
- okamiueru 3y ago
  I hope we reach a collective maturity on this. LLMs have so far I've noticed, left a trail of mediocrity. I'll of course not notice the parts that are good, so there is some confirmation bias here. And, I hate it. If not used appropriately, it's an automated output from left side on the Dunning-Kruger. Bullshit asymmetry has gotten more skewed, and it's tiresome.
  - yieldcrv 3y ago
    Thats just not my experience I’m using Miqu 70B Q4 and it immediately replaced Mixstral 8x7B Q5 for me I screen almost all of my responses in relationships through it, with deep context on why its modifying things, total paradigm shifts like a therapist is showing me more effective conversation styles. I wasn't seen as having low emotional intelligence to begin with and the results have been great. Translations Coding syntax Entire code bases Nuanced legal aspects of industries (deep conversations about obscure drug and treatment pricing by region and billing method, which matched reality) More stuff about different kinds of insurance and how to navigate insurance brokers, to great effect Whenever a contractor or professional outside of my knowledge domain gives me a word salad, I make a note of what they said verbatim and have the LLM translate that for me. Then I come back to them with informed responses they cant bullshit around. I got my HVAC fixed by pointing out what they are probably missing, and they were previously too prideful to notice or consider or admit. Got a payment coming from my landlord for this because they caused my energy bill to be higher. Large document analyses, which I thought was the final boss. I’m only giving these things an 8,000 token context window and things have been great and coherent
city41 3y ago
I've noticed whenever a topic comes up that I have a lot of knowledge in, people almost always chime in with incorrect or just flat out made up stuff. I always remain suspicious of anything I read in any comment section. Including here on HN.
handsclean 3y ago
The way I internalize it: public voting selects for layman plausibility, not correctness. Because laymen massively outnumber experts, the layman vote always overwhelms the informed one, so the reaction of people who don’t know the subject is the only thing that matters. Truth only seems to matter because most subjects either can be somewhat intuited by non-experts, or are in a niche that you’re not, so “layman plausibility” means your reaction, too. But the true nature of the dialog reveals itself as soon as people talk about something you’re an expert on. Answers like this aren’t a bug in a truth machine, they’re a plausibility machine working as designed.
- glitchc 3y ago
  As we know in the age of the internet, truth doesn't matter, only popularity does.
  - dspillett 3y ago
    The internet has taught me how many brilliant people there are out there. And how massively outnumbered they are by the rest of us!
- waveBidder 3y ago
  there's another reason for some optimism about a voting-truth connection: wisdom of the crowds. As long as there isn't a strong bias to people's estimate, the average will converge on the truth.
  - lupire 3y ago
    That only works when people bet that their guess is correct.
  - tripleSex 3y ago
    I am quite unsure as to the veracity of the claim that "the average will converge [upon] the truth". I recall cases being made (as asides) for the opposite conclusion. Intuitively even, this idea of equating truth with convergance towards the average opinion appears contradictory, counterfactual, and ahistorical. Excuse my being brass, but a "wisdom of crowds" seems to me oxymoronic on its face. I'd love to be persuaded otherwise though; mainly due to my perception of a lack of credence towards your view. Perhaps I have misunderstood your qualifier: "As long as there isn't a strong truth bias to people's estimate . . . "? Off the top of my head, I can't imagine any scenario in which a mixed population of laypeople and academics/experts would converge towards the same (vote average) findings as a sample of a handful of experts/academics. For example, would The Average converge towards correct mathematics or physics answers? Besides trivial, non-technical questions that do not require complex analysis, I think not. (See: False Memory: Mandela Effect. [0] [note]) [0]: https://en.m.wikipedia.org/wiki/False_memory#Mandela_effect https://en.m.wikipedia.org/wiki/False_memory#Mandela_effect [1]: https://en.m.wikipedia.org/wiki/Information_cascade https://en.m.wikipedia.org/wiki/Information_cascade [Note]: My point is that groups' thinking is liable to be compromised. (After all, what has been more important to a human — evolutionarily: the truth or social access?) Also see: Information Cascade. [1] {Post-Scriptum: My position is that if averages for answers to questions were taken, from the 'crowd' of the whole Earth, then these would diverge significantly and routinely from The Truth. If there are cases in which you feel this to not be the case I would inquisitively consider such scenarios waveBidder.} <Edit: Deletion: " . . . ~difficulty in lending~ . . . ">
whoswho 3y ago
This is why I ask for qualifications when someone has an authoritative tone.
isleyaardvark 3y ago
It's amazing how far that can take you. I saw a post on another social media site about something being wrong, and a comment said it's not wrong, it was just missing a "not". Which was the exact reason it was entirely wrong. So people can state absolute absurdities and have people agree.
- datavirtue 3y ago
  "So people can state absolute absurdities and have people agree" That Reddit mission statement.
  - BadHumans 3y ago
    Happens on HN all the time too.
- lupire 3y ago
  Some people are able to correct typos when reading.
  - isleyaardvark 3y ago
    Not the case there. The context of this was definitely not a typo.
wonderfulcloud 3y ago
But Reddit is exceptionally bad at this though. It's basically about what sounds the most positive for the upvoter's way of thinking rather than anything else.
- Auracle 3y ago
  Reddit is a place where you get downvoted for linking something that proves what someone was saying is wrong just because it goes against the site’s overall narrative. Lies are encouraged if they’re the correct lies.
  - redsoundbanner 3y ago
    The exact same can be said for academia
  - graemep 3y ago
    Depends on the sub. Political ones are the worst.
graemep 3y ago
Most often it is because people are too lazy to take the time to understand the explanation they are being given. Sometimes that is because experts just say "I am an expert so trust me" without proper explanations or links to explanations or evidence.
poulsbohemian 3y ago
The difference is that (subjectively) there used to be less of this on HN. The herd moves much more aggressively now rather than granting for debate. The generational differences are much more pronounced; the politics not a match of Democrat vs. Republican, but something equally vindictive. It isn’t particularly pleasant, but what are the alternatives?
scotty79 3y ago
> It's not scary - it's typical. It's really a great strength of human species. We may not exceed animals in any other quality, except for persistence hunting, but we are exceptional at copying the behavior of other individuals of our species without considering whether it's sensible or not. Even monkeys don't do this as much as humans.
yieldcrv 3y ago
truth is stranger than fiction in my experience thats true, its less familiar
fennecfoxy 3y ago
Yeah this behaviour is pretty normal for humans/tribal animals, I don't know why people are surprised really. I mean just look to politics, that already explains enough!
indymike 3y ago
> It's not scary - it's typical. It's no surprise that people who lack expertise will downvote and actual expert. I see it happen every day when people try to be the smartest person in the meeting or just simply cannot allow anyone to think they are wrong.

by speff 3y ago

jodrellblank 3y ago

[flagged]

speff 3y ago
To be honest, I was debating on even posting my message initially. It was off topic but I figured it would've been ignored. If I knew it was going to derail the discussion about Stratoscope's analysis that much, I wouldn't've posted it. Edit: Also grammar nazi'ing has little to do with linguistics and more to do with being a jerk...usually.

tcgv 3y ago

It is the Gell-Mann Amnesia effect, but on social media: > The phenomenon of people trusting newspapers for topics which they are not knowledgeable about, despite recognizing them to be extremely inaccurate on certain topics which they are knowledgeable about.

spicyusername 3y ago

Is it not possible to be both scary and typical?

0ckpuppet 3y ago

whoe political movements are built on this kind of momentum

deadbabe 3y ago

Hackernews is similar to ChatGPT in that regard. Lots of correct sounding answers that are really just a word salad.

MenhirMike 3y ago

> my friend the tax accountant gets downvoted for clarifying how taxes actually work. Let me guess: Tax brackets? That's the one thing that most regular workers in the US just don't seem to understand (and arguably, many people knowingly spread falsehoods to further some agenda).

speff 3y ago
Decent guess, but nah. Something to do with corporate tax accounting. Can't remember the details because that's out of my element.
- tpmoney 3y ago
  Tax write offs would be my guess. Every employee of charities that partner with retail locations for PoS donations must die a little inside each time some fool confidently asserts that they never give to those charities because it's all a scam. The money, they will assert with the confidence only someone so wrong can muster, is just used for write offs so the executives can have a big bonus and the company gets to claim they donated all the money. Bonus points if they assert that the charity doesn't even get the donated amount.
PopAlongKid 3y ago
I think the basic thing about taxes that is least understood is the difference between gross income and taxable income (the latter is the amount that tax brackets apply to). A close second is the difference between tax liability, and refund/balance due on the tax return.
- sgerenser 3y ago
  Just try to convince the average person that “getting a big refund” is a bad thing, since it means you gave the U.S. government an interest free loan.
  - rvnx 3y ago
    (Not from the US) Why is it a good thing to lend for free to the US gov ? Because the regional banks aren’t that stable ?

nsagent 3y ago

I think this also partly explains the LLM hype — people can be as confidently incorrect as LLMs, or maybe LLMs are as confidently incorrect as humans since they are trained on text from social media.

wolverine876 3y ago
The LLM output matches what the crowd expects.
okamiueru 3y ago
I hope we reach a collective maturity on this. LLMs have so far I've noticed, left a trail of mediocrity. I'll of course not notice the parts that are good, so there is some confirmation bias here. And, I hate it. If not used appropriately, it's an automated output from left side on the Dunning-Kruger. Bullshit asymmetry has gotten more skewed, and it's tiresome.
- yieldcrv 3y ago
  Thats just not my experience I’m using Miqu 70B Q4 and it immediately replaced Mixstral 8x7B Q5 for me I screen almost all of my responses in relationships through it, with deep context on why its modifying things, total paradigm shifts like a therapist is showing me more effective conversation styles. I wasn't seen as having low emotional intelligence to begin with and the results have been great. Translations Coding syntax Entire code bases Nuanced legal aspects of industries (deep conversations about obscure drug and treatment pricing by region and billing method, which matched reality) More stuff about different kinds of insurance and how to navigate insurance brokers, to great effect Whenever a contractor or professional outside of my knowledge domain gives me a word salad, I make a note of what they said verbatim and have the LLM translate that for me. Then I come back to them with informed responses they cant bullshit around. I got my HVAC fixed by pointing out what they are probably missing, and they were previously too prideful to notice or consider or admit. Got a payment coming from my landlord for this because they caused my energy bill to be higher. Large document analyses, which I thought was the final boss. I’m only giving these things an 8,000 token context window and things have been great and coherent

city41 3y ago

I've noticed whenever a topic comes up that I have a lot of knowledge in, people almost always chime in with incorrect or just flat out made up stuff. I always remain suspicious of anything I read in any comment section. Including here on HN.

handsclean 3y ago

The way I internalize it: public voting selects for layman plausibility, not correctness. Because laymen massively outnumber experts, the layman vote always overwhelms the informed one, so the reaction of people who don’t know the subject is the only thing that matters. Truth only seems to matter because most subjects either can be somewhat intuited by non-experts, or are in a niche that you’re not, so “layman plausibility” means your reaction, too. But the true nature of the dialog reveals itself as soon as people talk about something you’re an expert on. Answers like this aren’t a bug in a truth machine, they’re a plausibility machine working as designed.

glitchc 3y ago
As we know in the age of the internet, truth doesn't matter, only popularity does.
- dspillett 3y ago
  The internet has taught me how many brilliant people there are out there. And how massively outnumbered they are by the rest of us!
waveBidder 3y ago
there's another reason for some optimism about a voting-truth connection: wisdom of the crowds. As long as there isn't a strong bias to people's estimate, the average will converge on the truth.
- lupire 3y ago
  That only works when people bet that their guess is correct.
- tripleSex 3y ago
  I am quite unsure as to the veracity of the claim that "the average will converge [upon] the truth". I recall cases being made (as asides) for the opposite conclusion. Intuitively even, this idea of equating truth with convergance towards the average opinion appears contradictory, counterfactual, and ahistorical. Excuse my being brass, but a "wisdom of crowds" seems to me oxymoronic on its face. I'd love to be persuaded otherwise though; mainly due to my perception of a lack of credence towards your view. Perhaps I have misunderstood your qualifier: "As long as there isn't a strong truth bias to people's estimate . . . "? Off the top of my head, I can't imagine any scenario in which a mixed population of laypeople and academics/experts would converge towards the same (vote average) findings as a sample of a handful of experts/academics. For example, would The Average converge towards correct mathematics or physics answers? Besides trivial, non-technical questions that do not require complex analysis, I think not. (See: False Memory: Mandela Effect. [0] [note]) [0]: https://en.m.wikipedia.org/wiki/False_memory#Mandela_effect https://en.m.wikipedia.org/wiki/False_memory#Mandela_effect [1]: https://en.m.wikipedia.org/wiki/Information_cascade https://en.m.wikipedia.org/wiki/Information_cascade [Note]: My point is that groups' thinking is liable to be compromised. (After all, what has been more important to a human — evolutionarily: the truth or social access?) Also see: Information Cascade. [1] {Post-Scriptum: My position is that if averages for answers to questions were taken, from the 'crowd' of the whole Earth, then these would diverge significantly and routinely from The Truth. If there are cases in which you feel this to not be the case I would inquisitively consider such scenarios waveBidder.} <Edit: Deletion: " . . . ~difficulty in lending~ . . . ">

whoswho 3y ago

This is why I ask for qualifications when someone has an authoritative tone.

isleyaardvark 3y ago

It's amazing how far that can take you. I saw a post on another social media site about something being wrong, and a comment said it's not wrong, it was just missing a "not". Which was the exact reason it was entirely wrong. So people can state absolute absurdities and have people agree.

datavirtue 3y ago
"So people can state absolute absurdities and have people agree" That Reddit mission statement.
- BadHumans 3y ago
  Happens on HN all the time too.
lupire 3y ago
Some people are able to correct typos when reading.
- isleyaardvark 3y ago
  Not the case there. The context of this was definitely not a typo.

wonderfulcloud 3y ago

But Reddit is exceptionally bad at this though. It's basically about what sounds the most positive for the upvoter's way of thinking rather than anything else.

Auracle 3y ago
Reddit is a place where you get downvoted for linking something that proves what someone was saying is wrong just because it goes against the site’s overall narrative. Lies are encouraged if they’re the correct lies.
- redsoundbanner 3y ago
  The exact same can be said for academia
- graemep 3y ago
  Depends on the sub. Political ones are the worst.

graemep 3y ago

Most often it is because people are too lazy to take the time to understand the explanation they are being given. Sometimes that is because experts just say "I am an expert so trust me" without proper explanations or links to explanations or evidence.

poulsbohemian 3y ago

The difference is that (subjectively) there used to be less of this on HN. The herd moves much more aggressively now rather than granting for debate. The generational differences are much more pronounced; the politics not a match of Democrat vs. Republican, but something equally vindictive. It isn’t particularly pleasant, but what are the alternatives?

scotty79 3y ago

> It's not scary - it's typical. It's really a great strength of human species. We may not exceed animals in any other quality, except for persistence hunting, but we are exceptional at copying the behavior of other individuals of our species without considering whether it's sensible or not. Even monkeys don't do this as much as humans.

yieldcrv 3y ago

truth is stranger than fiction in my experience thats true, its less familiar

fennecfoxy 3y ago

Yeah this behaviour is pretty normal for humans/tribal animals, I don't know why people are surprised really. I mean just look to politics, that already explains enough!

indymike 3y ago

> It's not scary - it's typical. It's no surprise that people who lack expertise will downvote and actual expert. I see it happen every day when people try to be the smartest person in the meeting or just simply cannot allow anyone to think they are wrong.