Community fact-checking

[NorthernScrub]

Linus spoke at length during the last WAN show (https://www.youtube.com/watch?v=rnIeknursww&t=4413s) about options for improving factual accuracy in LTT videos, with some speculation on utilising the strength of the watching community in the process. Particularly, there was some discussion about how such a scheme would work, and what rewards would be available to fact checkers. There is an already established precedent for such a community-oriented programme, although not from a source one might expect.

 

Here in the UK, we have a society programme called YouGov, funded in part (or perhaps in whole) by the government itself. This organisation seeks to gain some insight into society's general opinion on current events, in various categories that include politics, social media, celebrity status, film and television, and a vast array of other subjects. Each survey completed returns an amount of points, very much like other survey systems. When a user has gathered a total of five thousand points, those points can be redeemed for a cash transfer of £50, or a giftcard of the same value. It can take anything from four months to a year to accumulate points, but they are legitimate - and the rewards also, as I have successfully received a bank transfer from this program.

 

A similar concept could be applied to a fact-checking programme. Users sign up to the programme, register their details, and use these credentials for submitting corrections to the LTT team. Such a submission would need to be in a standardised format, submitted via webform, and contain at least a primary source, or two secondary sources, alongside the video URL and a timestamp. A brief text field would, in my opinion, be necessary, since a correction might require a contextual explanation - even though a text field will increase the amount of time required to evaluate each submission.

 

Speaking of submissions, this is where the points system comes in handy. Without such a system, LTT would be required to make very frequent payouts of minimal value - something that payment processors do not care for, and which might increase LTT's financial service costs. Additionally, since the number of reports for the same correction might be significant, it would be wise not only to use a points system, but also to limit the number of signups and expand only as necessary.

 

Finally, there comes the subject of the payouts themselves. I would strongly suggest a value of around £10, or $10, so as to limit the incentive. Each correction should be worth between one and five percent of the points required to receive a payout, which would then encourage consistent efforts by consistent members. Naturally, each correction submitted would need to be approved before being rewarded points - although perhaps a dual or tri-reward system might be efficable too. Such a reward system would reward a fact checker for a quality report, providing a bonus if the reward is correct. A larger bonus would be applied if the fact-checker happens to be the first to submit the report.

 

I would propose that altering the reward values for the actual correction be avoided - this just adds confusion and space for disagreement, which is unnecessary. To keep the process as streamlined as possible, as well as to ensure that fact-checkers are adequately rewarded for the effort they make, it is easier simply to set a standardised reward for all accepted submissions.

 

As for making the payouts, a bank or chequeing transfer is preferable to most users. However, where this is not available, it also makes some sense to provide alternatives, just as YouGov do. Gift cards, virtual prepaid credit cards, and store credit should also be offered where possible. In this format, I would expect it to be viable to run this programme not only in Canada and North America, but in the UK and most of the EU as well. It may even be possible to run the programme in Oceania and parts of Asia, since the reward structure is designed to be as flexible as possible with regards to reward payouts themselves.

Edit:

Some refinements:

- It would be useful also to have an additional public forum of sorts. When a correction can only be sourced from secondary sources or principle knowledge, placing a correction up for discussion would be a useful tool for exanding and confirming corrections. However, the reward system cannot necessarily be applied to these discussions, at least not in the same fashion as a nominally accepted correction. Rather, a fact-checker might submit an addendum to a correction (be it their own correction or otherwise), providing alternative sources if possible. Addendums would need to be rewarded, on a two-part scale denoting whether or not the addendum refines the correction, or disagrees with it entirely.

- The limits on the number of signups would need to be profoundly strict. Although payouts would not be an immediate expense, they would most definitely be an expense in the future. Limiting the number of signups limits this expense, whilst still looking to the diverse skillsets of the community.

Edited August 5, 2023 by NorthernScrub

[Brolil]

This issue kinda screams Lab to me. I feel like if he wants a team to do this, but is worried he can’t hire an expert for everything, maybe see how the Labs team feels about adding “fact checking” to their job description, and then hiring a few more Labs employees to offset the extra work added. That way you can take the experts you already have employed and get their input while providing them help with their existing workflow to not overburden them. 

[NorthernScrub]

5 minutes ago, Brolil said:

...Labs team feels about adding “fact checking” to their job description, and then hiring a few more Labs employees to offset the extra work added...

Linus mentions the breadth of expertise that is covered in LTT's media. He's not wrong - and as a business operator myself, it is very difficult to find the balance of skillsets that allows an organisation to function whilst being financially viable. Hiring a handful of Labs employees is unlikely to cover the scale of expertise that would be required for this to make any meaningful impact, which is why a crowdsourced approach is such an attractive option.

[Erioch]

28 minutes ago, NorthernScrub said:

which is why a crowdsourced approach is such an attractive option.

Plus they don't have to pay anyone.

[Brolil]

12 minutes ago, Erioch said:

Plus they don't have to pay anyone.

I think Linus kinda wanted to pay people for this on a morality basis. Whether that was just him trying to keep a good look in the public eye is another story, but he definitely expressed a desire to compensate in some aspect. 

[NorthernScrub]

2 hours ago, Erioch said:

Plus they don't have to pay anyone.

No, LTT still has to pay their crowdsourced factcheckers here. The idea is to limit and organise the number of reports coming in, and reward in bulk rather than for each individual report. Doing it via a select group prevents staffers behind the programme from being deluged with mass reports every time a presenter mistakes a minor detail.

[Zodiark1593]

Sounds like LTT is looking for QA, with fairly established policies. 

[Erioch]

2 hours ago, NorthernScrub said:

No, LTT still has to pay their crowdsourced factcheckers here. The idea is to limit and organise the number of reports coming in, and reward in bulk rather than for each individual report. Doing it via a select group prevents staffers behind the programme from being deluged with mass reports every time a presenter mistakes a minor detail.

I wasn't paying super close attention but I only remember them talking about "rewarding" select people.  I remember Linus saying specifically that his wife was not going to be writing 5 cent checks.

[Erioch]

4 hours ago, Brolil said:

I think Linus kinda wanted to pay people for this on a morality basis. Whether that was just him trying to keep a good look in the public eye is another story, but he definitely expressed a desire to compensate in some aspect. 

Good for them if they get people to do it for free or nearly free.  I mean, they had volunteers at LTX and no one seemed to care.

[LogicalDrm]

Haters are gonna hate them regardless. Imo this is pointless and should be resolved with internal look at procedures and policies before any external sourcing is done. Or if external people are brought in, then use coaches to train people first. I feel like even moving from daily videos to 3+1 until they improve in quality when it comes to accuracy.

 

This would require making some verification and database with personal information. So same issues why Linus DOES NOT want to have user reviews on Labs site.

[leadeater]

21 hours ago, Brolil said:

This issue kinda screams Lab to me.

Lots of errors come up in videos that Labs couldn't help with, like incorrectly labelled graphs or what is being spoken not matching what is being shown. Fact checking is one thing but other types of issues exist as well that make it through to published videos. Basically it's an all encompassing quality control aspect.

 

Personally I'm on the fence about the idea, well intentioned idea but may not work out. Don't know 

[Needfuldoer]

I agree that fact checking is sorely needed (just look at all the asterisks and corrections the editors add), but it needs to be done before the video is released, not by the audience after the fact.

[leadeater]

15 minutes ago, Needfuldoer said:

I agree that fact checking is sorely needed (just look at all the asterisks and corrections the editors add), but it needs to be done before the video is released, not by the audience after the fact.

The intention is to be done before public release, after or during parts of LTT internal review. Wouldn't make much sense otherwise if you think about it, how would it be any different than now 

[HappyPie]

Hi first time poster, hopefully I found the correct subforum otherwise I do apologize for it.

 

About the last WAN show (Aug 8) where the topic was about fact checking the video scripts, how to approach it, pro and cons of different suggestion.

In short, I have an idea and basically it is:

• Let the current error rate stay, collect the feedback from viewer comments and then make a "correction video" content correcting the information and if feasible also having a deep dive into that subject.
• The video style format could be something like "Correct Quickie", a general LTT video or WAN show 2+ person panel style format.

 

 

The reason, philosophy and concept behind this presented idea and about the topic regarding failure/mistakes in general:

Generally instead of fearing mistakes I tend to embrace it rather than "avoiding it at all cost". However if it is easy to avoid the mistake and that there is nothing to be gained from it then I do try avoiding it.

Usually the 80/20 rule apply to many things and I believe that it might as well here where the last 20% of the fact checking could cost 80% of the resources.

 

About the mistakes, I would categorize the errors in two where one is "Nerdy in depth knowledge" and the other one is basically "Slip ups, typo and brain farts". The "Slip ups" should be mostly avoided as much as possible similar to when editing text for typo's grammatically; While the "Nerdy in depth" category (which is also both resource heavy and costly to both research and fact check) is the one that I suggest to spend only a reasonable amount of resources on instead of "avoid it at all/high cost".

When the mistakes do eventually pass by with your "reasonable fact checking procedure" you have an opportunity instead of making content out or them, creating interaction and discussion along with it.

 

 

Future suggestion using AI:

When possible, when the implementation technology of GPT has matured a bit you could have a "fact checker" of the script's content similar to having an "autocorrect" for the script's grammar.

You can basically do this today, somewhat manually though, with ChatGPT using Wikipedia and other resources to fact check the content HOWEVER generally GPT based AI should still not be regarded as a 100% correct tool; Although it can be used in early stages to flag for potential errors lowering the errors and become more efficient while doing so.

Might also be a LABS thing perhaps, an endeavor to make an inhouse "fact check helper" AI tool and I bet both that you could make content out of it similar to your automatic benchmark system and that Luke would love it, at least love the idea of it but perhaps not from a business standpoint and its feasibility (at least not today perhaps)

 

 

Lastly, love your WAN show

[LogicalDrm]

6 hours ago, HappyPie said:

Hi first time poster, hopefully I found the correct subforum otherwise I do apologize for it.

Merged to previous topic about same subject. This was posted on Sunday and was still visible on first page. Please look at least couple of first pages of the subforum when making threads about current/hot events/subjects.

[HappiePlant]

I feel like something similar to Twitters Community Notes could work. People can make corrections in the draft and in videos and other people can rate them as helpful or unhelpful. If a correction gets enough votes, it will be shown to the script writer or someone at LTT.

[LAwLz]

I am very glad that I watched that clip. 

As someone who has been very critical of LTT's seemingly large quantity of errors in their videos, it is very good to see them thinking about solutions to fix those issues.

 

As someone who has also from time to time gotten the impression that Linus sometimes takes advantage of his position of power (perhaps unknowingly), I also found the comments about not just using fans, interns or junior members as free or cheap labor was really refreshing in a time where influencers take advantage of parasocial relationships left and right with no shame in their body.

I absolutely agree with Linus that asking fans to do free labor for them would be exploitative, even if the fans are willingly doing it. Especially if you are such a big and successful company as LMG is.

The bigger you are, the more responsibility you have to do the right thing, and I think Linus' brainstorming here is hitting the nail on the head over and over again.

 

While I think the system is a good idea, and doing it through Google doc and comments, I am not sure paying money is a good idea. I am worried that it will result in people rushing to find nitpicks and it could potentially cause a lot of negativity as you said, especially if you take the bonuses away from your ordinary staff. Punishing a writer for saying "HDMI" instead of "HDMI 2.0" for example is probably not a good idea. Not sure what the solution is, but I really like that you are thinking about it. Maybe monetary payments aren't the answer. Not sure what the alternative would be, but money might cause people to join the program for the wrong reasons which in turn would cause issues. For example, people not really correcting things, but rather adding tangents like some kind of Wikipedia-reply-bot just because they want some contributions and thus money.

Or who knows, maybe money is the answer?

 

 

Linus, if you are reading this, I know we have had our differences but in my eyes, this is exactly the type of thing I wish you would do, and it's potentially a very good solution to some of the issues I have with the channel.

You might have forgotten about it, but I haven't forgotten the conversation we had back in 2014. As I said back then, am more than willing to help you with this if you want my help. 

If it ends up being through submissions then I hope you post it somewhere I will find it. I think a lot of people who may be willing to help might not watch your videos, so that might be a good idea to keep in mind.

[Kisai]

On 8/5/2023 at 9:57 AM, NorthernScrub said:

 

Finally, there comes the subject of the payouts themselves. I would strongly suggest a value of around £10, or $10, so as to limit the incentive. Each correction should be worth between one and five percent of the points required to receive a payout, which would then encourage consistent efforts by consistent members. Naturally, each correction submitted would need to be approved before being rewarded points - although perhaps a dual or tri-reward system might be efficable too. Such a reward system would reward a fact checker for a quality report, providing a bonus if the reward is correct. A larger bonus would be applied if the fact-checker happens to be the first to submit the report.

 

 

Never offer payouts to "correcting" data, otherwise you just incentivize gaming it. If something is factually incorrect, you'd just be rewarding the first person who notices it instead of the 6000 people who might recognize it, but perhaps see it as a comedic bit, or not important enough to report on.

 

If you want to offer a payout to correcting misinformation, it has to be good-faith, experience-driven data. Not simply "well this textbook and wikipedia say this", because wikipedia has often been wrong when two or more people fight over the data quality, and the person who gives up first (right or wrong) just walks away from it. Bad-faith, conjecture-driven data just leads to people all claiming they found the same bit of misinformation, even if the actual information is correct but grammatically badly formed.

 

As it is, when it comes to video content, what are you going to do, film a video and then beta-test it with some viewers? No. You're going to release the video right or wrong, and if something is extremely wrong or objectionable, you're just going to leave it up anyway. Nobody is going to watch it again, even if it's corrected.

 

No, to me a better strategy is to "defer to expertise", bring on people, either in-person at LMG or on-site at the guest's location with the right tools and have them actually demonstrate how something that Linus maybe shouldn't be trying demonstrate, or even handle it.

 

If something then slips past that, and into a video, a subsequent video with an expert in the field should be brought on to explain why Linus was incorrect about THING in VIDEO, and demonstrate why it's wrong. That solves the problem of too many nitpickers sending low-quality fact-checks. Put your own neck on the line.

 

Let's say "Linus said the Sky is blue, a color you can find #87eeb with your web browser"

Now, technically, that is correct, but a nitpicker would then say

"Well actually, the sky is a range of colors depending on the particulate matter in the air, the position of the sun, and the latitude of the person observing it."

 

It's that kind of nitpicking I find extremely aggravating, and it happens a lot on this forum as well.  Someone can not be satisfied with a generalization with maybe a jokey reference to something barely on topic, they have to nitpick something so they can be more correct than the person who said it originally, despite it further derailing the topic.

 

[Needfuldoer]

5 hours ago, Kisai said:

As it is, when it comes to video content, what are you going to do, film a video and then beta-test it with some viewers? No. You're going to release the video right or wrong, and if something is extremely wrong or objectionable, you're just going to leave it up anyway. Nobody is going to watch it again, even if it's corrected.

Exactly.

 

Videos that are meant to be factual need to be QA-ed by someone* internal before they're released to the general public, because you can't put the toothpaste back in the tube. Half the audience has already seen a video full of on-screen corrections, and maybe noticed more corrections in the description or a pinned comment (if they're watching on a platform where they can see those).

 

Videos full of afterthought corrections look sloppy, and they don't reflect well on the data Labs is producing (even when it's rigorous and accurate).  "Whoops that was wrong, oh well, on to the next one!" isn't a solution to the problem either because it erodes credibility in the long term.

 

If the sets and production schedule are so packed that reshoots aren't possible, can the host just record a VO in a tracking booth?

 

*Ideally, they should be QA-ed by someone who has familiarity with the subject matter, but wasn't involved in the video's production to that point. We all have some degree of blindness to the problems with our work, so a fresh set of eyes would cover those gaps.

 

[7h3-3ng1n33r]

How about just flip it on its head, setup some community chat groups with Subject Matter Experts (SME's), have the writers (and/or LTT internal QA team) ping sections of scripts where they are unsure on the facts or where they feel they need more clarity.

This prevents a free for all on comments in scripts, plus the writers are SMEs in their own areas so not everything they've written needs to be fact checked, otherwise can open yourself up to so very pedandtic comments.

Also the comment Linus said is he wants the writers to do the job they are paid for and what their bonuses are paid for, this is just him giving them all the possible tools to ensure their scripts are correct and to be successful.

Plus if an edit is made through the SME process, then if during script review he questions it he can be shown the chat section to give him the confidence (note have to have a list of the people that are SMEs and their credentials).

As a reward give the SME's a floatplane subscription and/or special LTT store discount (keep it simple)

There will have to be a process for the LTT team to select and deselect SME's, start off with one big group with a few select people, and then expand out with different groups for different specialisms i.e. Linux Group, Car Group, Electronics Group, Mechanical Engineers Group, Data Centre Group, etc.


It would take time but its at least doing something to be more successful.

[NorthernScrub]

On 8/6/2023 at 4:33 PM, leadeater said:

The intention is to be done before public release, after or during parts of LTT internal review. Wouldn't make much sense otherwise if you think about it, how would it be any different than now 

 

I'd argue that this isn't really possible, for a multitude of reasons. Not least, the security aspect - you'd be relying on people not to leak videos pre-release, which could be tricky and potentially damaging. There's already a torrent of LTT's unreleased media, which could have been a great deal worse had it been more recent stuff.

 

 

On 8/7/2023 at 9:22 PM, HappiePlant said:

I feel like something similar to Twitters Community Notes could work. People can make corrections in the draft and in videos and other people can rate them as helpful or unhelpful. If a correction gets enough votes, it will be shown to the script writer or someone at LTT.

 

 

This is a fantastic idea, and perhaps one that LTT and other youtube creators can suggest to youtube. A community context bar would be much easier to implement than editing a video in-place.

 

 

 

On 8/8/2023 at 7:03 AM, Kisai said:

1)

Never offer payouts to "correcting" data, otherwise you just incentivize gaming it...

2)

... As it is, when it comes to video content, what are you going to do, film a video and then beta-test it with some viewers? No. You're going to release the video right or wrong, and if something is extremely wrong or objectionable, you're just going to leave it up anyway. Nobody is going to watch it again, even if it's corrected...

3)

...a better strategy is to "defer to expertise", bring on people, either in-person at LMG or on-site at the guest's location with the right tools and have them actually demonstrate how something that Linus maybe shouldn't be trying demonstrate, or even handle it...

4)

...It's that kind of nitpicking...

 

 

1) Statistically yes, in reality no. This is why survey sites work - the amount of work that people need to put in to get a reward should be significant - this means that the odd eedjit who wants to play hanky with it is easily weeded out. It's also why corrections need to be backed up with evidence, which is where the moderation comes in.

 

2) Except this isn't true. Repeat viewership on Youtube is *huge* - and not only this, but archival and third-party editing/clipping are also extremely common on the platform. Any single video can be cut down for humourous content. Just look at Tom Scott when he tried to "vape" and practically hacked up a lung - the sheer number of meme edits of that single clip alone are uncountable. LTT videos are the same - they are often watched many times over, downloaded, edited, reuploaded in different formats, you name it.

 

3) It's been said quite a few times already that this simply isn't possible. LTT cannot hire a subject matter expert in every single topic covered on the various LMG channels. It's not financially possible.

 

4) To some degree yes - although in a manner of speaking, LTT also *wants* this sort of nitpicking. Not quite to the extent of "well akshually", but the actual clip I posted in the OP includes Linus mentioning a desire for smaller issues, like slightly miscontextualised statements, to also be addressed. It's a very delicate balancing act.

 

On 8/8/2023 at 11:43 AM, Needfuldoer said:

Exactly.

 

Videos that are meant to be factual need to be QA-ed by someone* internal before they're released to the general public, because you can't put the toothpaste back in the tube. Half the audience has already seen a video full of on-screen corrections, and maybe noticed more corrections in the description or a pinned comment (if they're watching on a platform where they can see those).

 

Videos full of afterthought corrections look sloppy, and they don't reflect well on the data Labs is producing (even when it's rigorous and accurate).  "Whoops that was wrong, oh well, on to the next one!" isn't a solution to the problem either because it erodes credibility in the long term.

 

If the sets and production schedule are so packed that reshoots aren't possible, can the host just record a VO in a tracking booth?

 

*Ideally, they should be QA-ed by someone who has familiarity with the subject matter, but wasn't involved in the video's production to that point. We all have some degree of blindness to the problems with our work, so a fresh set of eyes would cover those gaps.

 

 

I don't think they look sloppy at all. Maybe it's my generation, or perhaps just the autism (no, seriously, not a joke), but a video filled with corrections speaks volumes to me about the effort that the producer of that video is willing to go to to ensure that the information they are passing along is accurate.

How many times have you found a guide on an old forum somewhere, only for it to be slightly incorrect or subjected to link rot? I have - even on large, active forums like XDA. It would be nice if the modern form of internet entertainment was similarly scrupulous in maintaining the accuracy of its information.

 

 

On 8/10/2023 at 5:35 PM, 7h3-3ng1n33r said:

How about just flip it on its head, setup some community chat groups with Subject Matter Experts (SME's), have the writers (and/or LTT internal QA team) ping sections of scripts where they are unsure on the facts or where they feel they need more clarity.

This prevents a free for all on comments in scripts, plus the writers are SMEs in their own areas so not everything they've written needs to be fact checked, otherwise can open yourself up to so very pedandtic comments.

Also the comment Linus said is he wants the writers to do the job they are paid for and what their bonuses are paid for, this is just him giving them all the possible tools to ensure their scripts are correct and to be successful.

Plus if an edit is made through the SME process, then if during script review he questions it he can be shown the chat section to give him the confidence (note have to have a list of the people that are SMEs and their credentials).

As a reward give the SME's a floatplane subscription and/or special LTT store discount (keep it simple)

There will have to be a process for the LTT team to select and deselect SME's, start off with one big group with a few select people, and then expand out with different groups for different specialisms i.e. Linux Group, Car Group, Electronics Group, Mechanical Engineers Group, Data Centre Group, etc.


It would take time but its at least doing something to be more successful.

Not a bad idea overall - large communities over several organisations are not a new idea - look at mailing lists. I'm still a member of LOPSA and a few others. LTT taking the initiative and creating a new form of closed, dedicated group catering specifically for SMEs in various fields would be fantastic.

[leadeater]

1 hour ago, NorthernScrub said:

I'd argue that this isn't really possible, for a multitude of reasons. Not least, the security aspect - you'd be relying on people not to leak videos pre-release, which could be tricky and potentially damaging. There's already a torrent of LTT's unreleased media, which could have been a great deal worse had it been more recent stuff.

That torrent isn't really the same thing here, came about for very different reasons. Trust is a concern obviously but that is why only trusted people are going to be chosen but I can assure you it's a process that will be happening before main public video release.

 

Community based review to improve QA of videos cannot happen any other way otherwise it's not different from now, that is why it's being proposed and trialed.

[Kisai]

13 hours ago, NorthernScrub said:

 

1) Statistically yes, in reality no. This is why survey sites work - the amount of work that people need to put in to get a reward should be significant - this means that the odd eedjit who wants to play hanky with it is easily weeded out. It's also why corrections need to be backed up with evidence, which is where the moderation comes in.

You do realize that people get practically assualted with "do you want to take a survey" junk, and you know how much it pays? 

 

Nothing. You can spend 20 minutes on a damned survey and get a few pennies out of it. Forget it. That's why this stuff get's gamed. People sign up for surveys, answer "c" for everything get get their 10 cents.

 

 

13 hours ago, NorthernScrub said:

2) Except this isn't true. Repeat viewership on Youtube is *huge* - and not only this, but archival and third-party editing/clipping are also extremely common on the platform. Any single video can be cut down for humourous content. Just look at Tom Scott when he tried to "vape" and practically hacked up a lung - the sheer number of meme edits of that single clip alone are uncountable. LTT videos are the same - they are often watched many times over, downloaded, edited, reuploaded in different formats, you name it.

Repeat viewership by the same people? Nope. Never.

 

And re-uploading the video will get it penalized by youtube for "Duplicate content"

 

13 hours ago, NorthernScrub said:

3) It's been said quite a few times already that this simply isn't possible. LTT cannot hire a subject matter expert in every single topic covered on the various LMG channels. It's not financially possible.

 

Then they shouldn't jump into subjects they can not get a SME for. Period. It's not a question of "not financially possible" it's a question of who wants to put their neck out for being wrong. There are people who are geniuses in their field, and would love to just be bloody asked if something is correct before some e-celeb rubs their butt all over it on it trying to sound smart.

 

Like nothing annoys me more on Youtube videos, is someone speaking out of their butt about a subject they clearly did not research. Bad information from a trusted source, lowers the trust of that source. You know how many darn youtube videos are just plagarised Wikipedia? LOTS. There are people mass-producing videos with stock artwork/video, who are pretty much just reading Wikipedia, word-for-word. 

 

Occasionally my favorite youtubers "didn't do the research/didn't understand the topic" and I'll give them a pass if it's completely NOT their subject. EG I watch at least 5 different lawyer youtubes, and occasionally they make tech/history mistakes because that's not the subject they are trying to speak on. They are speaking of the liability of certain things, which is why I was watching the video, and I don't really care if they incorrectly call Linux "Lin-ox"

 

LTT, I want the tech information to be accurate, and not just "I stole it from wikipedia", If I wanted wikipedia, I'd go to wikipedia.

 

13 hours ago, NorthernScrub said:

4) To some degree yes - although in a manner of speaking, LTT also *wants* this sort of nitpicking. Not quite to the extent of "well akshually", but the actual clip I posted in the OP includes Linus mentioning a desire for smaller issues, like slightly miscontextualised statements, to also be addressed. It's a very delicate balancing act.

 

LTT videos, worst crimes are Linus doing things he should not be doing, either for a safety reason, or for a liability reason. If he wants to break stuff for fun, while explaining how electronics work he should bring on Electroboom. Otherwise, I didn't come to LTT for Electroboom-style lectures and fires.

 

 

13 hours ago, NorthernScrub said:

 

I don't think they look sloppy at all. Maybe it's my generation, or perhaps just the autism (no, seriously, not a joke), but a video filled with corrections speaks volumes to me about the effort that the producer of that video is willing to go to to ensure that the information they are passing along is accurate.

No, if a video has an entire thread about it, be it the comments, ltt forums, reddit, etc about it being bad, the video should come down, fixed and then re-uploaded with the thumbnail and title having "Corrected on DATE" . But you only get to do one of these a year, otherwise people will just go "oh there goes Linus making Buzz Feed look good."

 

If it's nitpicky nonsense, I don't care. Either it was right the first time, or the nitpicky nonsense doesn't add value to the video. If every video had to sit there and answer every nitpicky point, every video would be a 4 hour video essay instead of the sane 15-30minutes.

 

13 hours ago, NorthernScrub said:

How many times have you found a guide on an old forum somewhere, only for it to be slightly incorrect or subjected to link rot? I have - even on large, active forums like XDA. It would be nice if the modern form of internet entertainment was similarly scrupulous in maintaining the accuracy of its information.

 

Videos are worse for information rot. At least text you can ctrl-f and find the thing you're looking for. A video you have to sit through, or try to scrub through it to find when the person does exactly X thing you're looking for. And maybe they played a different version and exploited a bug to get there.

 

Case in point, I streamed a game the other day, and the save game was nuked at the end because of a bug, not a feature. So if someone watches it and wonders why the game did a thing that they couldn't do, it's because I played the release day version, not the version that got patched 3 hours into the stream.

 

But I digress, I would rather LTT not gamify fact checking, because it'll only end up with people doing things for money, not because they care. Find actual SME's who are not financially incentivized to help, and instead give them a writing credit on the video and a link to what they are working on. If the video could not be produced at all without them, then consider paying them as a guest in the video.

 

That, in many cases is more valuable when it comes to people who explicitly want to be known as a SME. There are many people who are experts in their field and just groan every time they see tech "garbage" in television shows, and wish to be consulted so it stops taking people out of their immersion in the story. People learn stuff from TV and then repeat it to their friends and are humiliated later when someone tries to sound smart from it.

 

[Uttamattamakin]

Something similar to Twitters community notes might be a good idea.  Allow trusted floatplane and forum posters, and/or employees, to propose a "community note" and /or vote on these community notes.    Twitter reviews how one has tweeted and categorizes peoples views on various hot topics.  For a note to get accepted people from all sides of a controversial issue must agree that it is valid.  

Note this is all done with pseudonyms.  My Twitter name and my Twitter Notes name are not the same (and I do not get to pick the name either).  

 

The reward is not money but getting to see your note have an impact.   Seeing the record corrected.  One form of mental reward won't apply to an LTT video though since we all like LTT (right?).  That being seeing someone get BIG MAD when you inform them, with authority, that the world is in fact round... either with a note you wrote or a note you voted for. 

[hammerquill]

I know this topic has been preempted by this week's controversy, but I was just listening to this particular WAN show and wanted to weigh in. I hope LTT people are still reading this thread and thinking about this issue. Among other things, it is the kind of thing that will potentially help avoid major missteps like this week's trainwreck.

 

I suspect a lot of people here are also space nerds, so some may already be aware of this, but I think the model currently used by EverydayAstronaut is an excellent one. In his case, he releases scripts to his patrons on Patreon ahead of time, for them to fact check. Then he releases a mostly-edited video for a final check, again just for a certain level of patron. The comments are in google docs comment style, so anyone can see them, but it's not really working in real time, so you avoid the stan aspect that Linus was concerned about in the editing space. You'd have to ask him how well it works, but he seems to do well with it.

 

I suspect you could do the same with simply releasing the video to floatplane and soliciting corrections, then editing the video for corrections before release to the public. And as for rewards, I think the name in the credits would be plenty for most people to give them an actual reward - if you're a techie who cares about getting correct information out to the community, then getting your name displayed as a thank you for correcting LTT is something you could potentially even put on your resume.

 

If you really want to reward them more, you could also put them on lists that allow them to buy samples and the like that are being sold. Sorry, didn't mean that in a too soon way, just struck me as a good idea when done right.