Last active
August 17, 2023 10:39
-
-
Save Suzhou65/a76e9121aba8271860bdd8e2a1195f20 to your computer and use it in GitHub Desktop.
[ YouTube Subtitles ] The Problem with Linus Tech Tips Accuracy Ethics Responsibility
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Speakers | Content | |
|---|---|---|
| GN | a few days ago this clip from an LTT lab tour got a lot of attention online | |
| LTT | the difference between us and somebody like Gamers Nexus or Hardware unboxed is we test new components new tests every time | |
| GN | and then Hardware unboxed tweeted dissatisfaction with the statement made in that tour | |
| GN | we didn't say anything at all anywhere even though it's not actually true they don't collect new data and new tests every single time and it's not always necessary anyway | |
| GN | so when Linus went on the win show we hadn't gotten involved Linus responded to hard run box on the WAN show and then said the following | |
| Linus | there's a fair number of people that are talking about the the whole trust me bro situation uh there where some creators that .... it seems like you guys are .... look ...(sigh | |
| GN | just say it everyone knows you're talking about us | |
| Linus | it's pretty clear that not everyone in the Creator space handled that super professionally | |
| Linus | I don't obviously agree with some of the takes that were out there I | |
| Linus | I don't think that it's particularly um journalistic for example to ignore facts to ignore personal relationships | |
| GN | it is my job to ignore personal relationships | |
| GN | if lmg is going to make false accusations and wants to try and use personal relationships to gain favor so that we don't cover something we think is not okay | |
| GN | it's not going to work with us | |
| GN | we don't play that game | |
| Transitions | ||
| GN | the first clip from an LTT lab tour caught our attention | |
| GN | but it itself isn't worth commenting on | |
| GN | we have something wider in scope we want to comment on | |
| GN | and it's simply that we believe LTT or LMG is rushing too much to hit arbitrary and largely self-imposed deadlines | |
| GN | that are causing frequent and sometimes significant data errors and information | |
| GN | and this conflicts with the ramping marketing surrounding accuracy and data-driven objective reviews that they're pushing for from LTT Labs | |
| GN | so we've decided to turn off all YouTube monetization here | |
| GN | we will not be placing a sponsored ad in this video | |
| GN | we were just going to report on serious concerns that we have with LTT | |
| GN | and bring some awareness to these critical issues we've been thinking on this for a few months | |
| GN | now we have a few main points of discussion for today | |
| GN | they span from raw data that potential actually misleads consumers | |
| GN | even if unintentional accidental that counts too | |
| GN | and we have some ethical concerns and corporate connections that we want to talk about | |
| GN | including great ethical concerns regarding ltt's inaccurate testing and misrepresentation of a small company's prototype | |
| GN | followed by the auctioning off of that prototype | |
| GN | despite the company wanting it back for other media and not granting permission for the auction | |
| GN | also we're concerned about ethics issues regarding lmg's recent review of the pwnage Mouse | |
| GN | where they failed to notice that there was protective tape over the Teflon feet to knock the friction coefficient of the mouse | |
| GN | followed up by a pseudo correction where an lmg fails to take full ownership of its mistake | |
| GN | and instead blames the product for not hand holding them through the process of removing said tape | |
| GN | but we'll start with the simpler stuff and then come back to this | |
| GN | so the marketing is strong but exiting just entertainment and entering tactical data-driven reviews | |
| GN | and marketing on the back of that means that there needs to be proper work to back all that up | |
| GN | because people are now putting a different level of faith in the material coming out | |
| GN | and the expectations are different | |
| Transitions | ||
| GN | and we know these issues at LTT in terms of the the rushing the pushing employees and staff to hit deadlines despite concerns with the data | |
| GN | or the collection of the data | |
| GN | we know that these come from top down as shown in a recent LTT interview that it conducted with its own employees | |
| GN | where many of them stated concerns with restrictive time constraints | |
| LTT STUFF | let's publish less videos | |
| LTT STUFF | I wish we could back off the amount and focus on quality for a bit | |
| LTT STUFF | uh... more time for projects would be good | |
| LTT STUFF | just allow us more time to work on a video | |
| Transitions | ||
| Linus | we never have time for retrospection | |
| Linus | it's always just that's out what's the next thing | |
| LTT STUFF | would lessened a little bit have time to like debris if you look in the rear of your mirror though how things went | |
| LTT STUFF | I don't love that we have to release so many videos | |
| LTT STUFF | I don't know what our current count is weekly | |
| LTT STUFF | but I know it's very high very rarely am I particularly proud of a video | |
| GN | I've worked on and Linus's own comments recently reinforced that the company is still weighing carefully the impact | |
| GN | of spending another half day of work on something to get it right | |
| GN | rather than just shoving it out the door incomplete or potentially wrong | |
| Linus | but it was designed for a 3090 Ti | |
| Linus | well we went for gold and we tried it with the 4090. | |
| Linus | the temperature results were bad it was clear that it wasn't making perfect contact | |
| Linus | and and so Adam was the writer on that project | |
| Linus | what Adam asked is if we could spend more time put the 3090Ti on it and get real results | |
| Linus | and what I told him was no | |
| Linus | because and guys ... come on ... like | |
| Linus | for real okay it's 800 US dollars | |
| Linus | I don't know guys | |
| Linus | I'm not sure if I can apologize for not spending another 100 200 300 500 of various people's time | |
| GN | and that's an important indicator of mindset that comes from ownership of the company | |
| GN | we'll come back to that clip later | |
| GN | though they are rushing and this Rush is causing significant and frequent testing errors | |
| GN | but more importantly it is clouding their ethical judgment | |
| Linus | and what I told him was no | |
| GN | they are choosing to publish sometimes known bad or known incorrectly acquired data | |
| GN | that's different from a mistake which all of us do in fact make | |
| GN | yes | |
| GN | but that's part of this topic too | |
| GN | and | |
| GN | so given the knowledge of a lack of qualifications when known unqualified | |
| GN | or the intentional forfeiture of quality and accuracy for quantity instead | |
| GN | it calls into question the motivation editions of the corporation | |
| GN | and does in fact make it look to be bottom line Centric | |
| GN | this is to hit arbitrary self-imposed deadlines and uploading scheduling | |
| GN | whose end result is misleading for consumers | |
| GN | and at some point it is no longer an unintentional mistake to mislead those consumers | |
| GN | such as with known shortcuts that reduce the accuracy of the content | |
| GN | rather it becomes intentional to do so because of the knowledge of that approach | |
| Transitions | ||
| GN | we know some people won't like this video and those people will leave comments enraged citing drama as some discourage this desperately needed conversation | |
| GN | it's their way of emoting | |
| GN | I'm uncomfortable | |
| GN | and I agree with you this is a very uncomfortable video for me to make | |
| GN | I haven't enjoyed the process of it | |
| GN | but this is something we feel is desperately needed otherwise we wouldn't be doing it | |
| GN | and it is not simply drama to talk about errors and factual testing issues in a way that hopefully brings to light the problems | |
| GN | so that the media organization LMG can attempt to rectify them | |
| GN | if it in fact wishes to do so | |
| GN | we stand to lose a lot here we risk pissing off people with direct and very valuable and useful to us connections to YouTube | |
| GN | if we ever happen to need them | |
| GN | I don't have a connection at YouTube | |
| GN | really who can help me in any way | |
| GN | likewise we risk pissing off people who are connected to LMG that would include who operate in a class that we can't even begin to approach | |
| GN | but most of all we risk some of our viewers getting upset | |
| GN | but we are approaching this as objectively as possible and we ask that you listen to the concerns we are raising and the evidence that we are presenting | |
| GN | because ultimately if Linus Media Group wants to be consumer first | |
| GN | then it too should want this level of transparent currency and should want this knowledge of the issues we are presenting today | |
| GN | the company is now over 120 people it recently had a purchase or an acquisition offer at 100 million dollars | |
| GN | it's helmed by a new CEO whose past includes a career in director positions at companies like DELL and Corsair | |
| GN | its lab is led by a former marketing director for ASUS | |
| GN | which was recently the title sponsor for LTX following what we believe to be a rather soft-handed handling of asus's board and RMA issues by LTT | |
| GN | LMG is a big company now | |
| GN | it has the resources to do better and it should be treated as such | |
| Transitions | ||
| GN | okay | |
| GN | with that stance clear we're jumping into two key categories here | |
| GN | bad data and misinformation | |
| GN | this includes unintentional or intentional | |
| GN | there is a bit of a gradient | |
| GN | bad data is testing which is incorrect or erroneous specifications | |
| GN | for example doubling down on flawed testing and Reporting on conclusions as if it reflects profit testing when in fact it has been shown not to | |
| GN | these topics will later expand for far more important ethics discussion and concerns that we have regarding sort of the corporate positioning that LMG is in now | |
| GN | even with its heavily marketed debut of labs | |
| GN | LMG is still consistently making errors in nearly every single review that they publish | |
| GN | it is okay to make errors all of us do this in this line of work | |
| GN | but the concern we have is that these types of errors are in nearly every single LTT technical or review type of video | |
| GN | so there are enough errors that we can't fit them all into a normal length video | |
| GN | we'll cover a few of them as fast as we can to establish the premise and the concerns and everything is from the last year only | |
| Transitions | ||
| GN | we're starting with the missed published errors the more troubling ones are later in the video | |
| GN | let's start with the entire RTX40 series | |
| GN | LTT's RTX40 series reviews started with the RTX 4090 which debuted with multiple erroneous charts | |
| GN | but one in particular was heavily written into the conclusion of the card | |
| GN | that chart was for cyberpunk where LTT allowed to slip past them an outlier with the 4090 outperforming the 3090 Ti by 300 percent | |
| GN | in reality that should be closer to 72 to 75 percent | |
| GN | we explained this previously but that's because LTT parameters and double check that the graphic settings weren't changing due to other dependencies during test setup | |
| GN | this is a massive human error and it is indicative of poor process control | |
| GN | no amount of blaming the video game or blaming how it settings behave is going to get you off the hook for this | |
| GN | it's simply human error that anyone could then look at this chart and think it makes any sense further indicates a poor understanding of the data collected | |
| GN | most of the rest of their own data was relatively consistent so there should have been clear indicators that this was wrong | |
| GN | despite being the only one they originally corrected in their comment this wasn't ltt's only error | |
| GN | and that's a separate concern | |
| GN | because it comes across as burring in just how many errors there truly were by only acknowledging the obvious one | |
| GN | or not being aware of them | |
| GN | both are bad | |
| GN | shadow of the Tomb Raider from that same review was even outperforming the 3090 by 61 percent it should be closer to 13 maximally | |
| GN | once again there's a lack of basic attention to detail maybe even knowledge to identify such massive errors | |
| GN | moving to automation makes us even more dangerous | |
| GN | because now you need the skill to identify flaws through numbers alone without even looking at the screen | |
| GN | the trouble was just how long it took LTT to acknowledge this mistake | |
| GN | this should have triggered deletion of the video basically immediately when they confirmed it was erroneous probably a community post on YouTube a Twitter post | |
| GN | and a timeline for a fixed video with corrections | |
| GN | but LTT waited days to fix it and they eventually posted a comment | |
| GN | but they didn't even pin it until people in the community started getting mad about that | |
| GN | that comment that was posted also didn't disclose all the errors just the lowest hanging fruit of them | |
| GN | which is cyberpunk and eventually they replaced the entire video in place | |
| GN | they didn't delete it there's a YouTube feature that we don't have most YouTube creators don't have where they can swap it in place so they did that | |
| GN | but only after over a million people saw it and we don't believe there was enough active Outreach to get in front of this bad data | |
| GN | and alert those million plus people that they're purchasing decisions may need to be reevaluated if that data influenced those decisions | |
| GN | probably there should have been a recall the content | |
| GN | if you're curious how we handle these issues the point here isn't really to talk | |
| GN | about that we have a published set of guidelines that we follow | |
| GN | it is our process and our standard now that's on gamers.nexus | |
| GN | it's been there a little while | |
| GN | the 4060 review had enough evidence of being rushed | |
| GN | but one of the other pieces was this chart where an A750 is shown to be better than an RTX4060ti with Ray Tracy not | |
| GN | which no other public data seems to support including our own | |
| GN | they probably remade their RTX4090 mistake here with applying different settings or perhaps some type of upscaling | |
| GN | they didn't even catch this one this is another one of the ones that I caught | |
| GN | in their RTX4060 review we noticed that their GTX 1060 gains 20 percent more performance than shown in their 7600 review | |
| GN | that's definitely not from a driver change with a card this old and nothing else in those charts really methodological change either | |
| GN | this is clearly an error in one of the two results that no one noticed this is beyond us | |
| GN | the RTX4060Ti review they think that memory sharing between the GPU and system memory isn't a thing without misinformation | |
| GN | more recently in June LTT published its RTX 4060 review with an incorrect Lane count in the spec sheet | |
| GN | they listed it as by 16 when it's actually by 8 | |
| GN | it Maybe fore this LTT previously made the same exact mistake with its RTX4060ti review | |
| GN | this video on YouTube they replaced it in place again using the special feature that most people don't have | |
| GN | but if you look at their BiliBili Channel you can see what the original was which was the wrong data | |
| GN | in the month since they didn't roll forward that correction | |
| GN | LTT had already corrected this once before which indicates to us that they are once again lacking the proper processes to ensure learned lessons carry forward | |
| GN | so it is missing basic research which once again starts to become Reckless irresponsible and concerning for a company branding itself in this position | |
| Transitions | ||
| GN | moving to bad testing and coolers in May Linus Tech tips published a cooler Benchmark video it had a number of errors that take an entire video to explain | |
| GN | but among them this particular chart gives away one of the major problems because it illustrates inconsistency between LMG's own testing | |
| GN | in this chart LMG has the NH-D15 from Noctua in red and green | |
| GN | the top two bars | |
| GN | and then it has the Peerless assassin in blue and an orange | |
| GN | notice that for each result the D15 is consistently a few degrees better than the Peerless assassin in LMg's testing | |
| GN | this isn't the problem | |
| GN | the problem is that ltt's own data is inconsistent when you look at the 7700x | |
| GN | suddenly we go from a couple of degrees Advantage for the D15 over the Peerless assassin to 10 degrees Advantage for the Peerless assassin | |
| GN | there is absolutely no way with these CPUs that these results can deviate this much without error in the testing process | |
| GN | that this chart can get past someone and not trigger a removal of that data entry or a significant attempt to try and explain what happened | |
| GN | indicates to us that LMG presently lacks the qualifications to conduct proper cooler reviews | |
| GN | there is no reason for a sudden surprise 10 degree Delta to favor the cooler that was losing in every other test | |
| GN | LTT refers to the frequency reduction of 300 megahertz in one test as quote not enough to affect real world performance | |
| GN | when actually that's basically reducing a 3900k to a lesser CPU like a 13700 | |
| GN | that is a huge difference so the ability to judge the data itself once again is suspect and concerning | |
| GN | LTT shows a 280 millimeter liquid cooler has worse than a downdraft cooler then blames it on nzxt's software | |
| GN | we are not able to replicate this result | |
| GN | they stay in the XT runs in quote silent mode by default which still wouldn't reduce it to worse than a stock cooler | |
| GN | and additionally if it runs in silent mode at all that means there are methodological flaws where they are not properly controlling a fan and pump speeds for this testing | |
| GN | and it's not proper to just blame the product because you don't want to control your test environment and you don't want to use it the way it's designed | |
| GN | this point alone calls all data into all of these charts into question | |
| GN | and it doesn't even look like they're controlling the frequencies and voltages properly | |
| GN | the surface area calculation is incorrect I made that mistake many years ago so I get this one | |
| GN | calculating the total accident general area of the cooler does not count the surface area of the fins which is the area that actually matters because it's where it carries the heat | |
| GN | LTT says that Noctua D15 can take quote anything you throw at it and that quote just doesn't lose | |
| GN | then they say they use it on all of their internal test benches for reviews | |
| GN | then they show it thermal throttling and cinebench testing seemingly | |
| GN | without realizing it or at least without mentioning it that it can get this close to TJ Maxx at all should immediately | |
| GN | eliminate it from using test benches where you want to see how the base product performs without external variables | |
| GN | we're a bit baffled by how huge the endorsement was for the Noctua NH-D15 given their own data | |
| GN | but Noctua has close ties with LTT so that may subconsciously be influencing the interpretation | |
| GN | and even if it's not it's still unfair to Noctua because now it's calling into question a positive comment because of those ties | |
| GN | that's why it cuts both ways when you have a conflict this deep | |
| GN | it's not just advertising anymore | |
| GN | final Point here the Peerless assassin was tested backwards and LTT blamed the lack of instructions | |
| GN | but we also suspect a mounting issue given their 10 degree Delta | |
| GN | good lab methodology would have ever done in C is for redundancy is the Safeguard against human error or perceived instructionscerror | |
| GN | in April LTT reviewed the Amazon Basics cooler LTT shows data where in the veteran V5 is producing numbers which seem physically impossible as compared to other coolers | |
| GN | on the chart as you follow along note that they sorted the charge worst to best | |
| GN | one thing we noticed was the max temperature value which is a 13 degree Delta against the average in their cyberpunk chart | |
| GN | this indicates that either LTT is averaging data which is not yet at steady state or it may have a bad Mount or test processes | |
| GN | this data should have triggered additional retests | |
| GN | especially as there is plenty of publicly available data supporting the Venture V5 as minimally equivalent to the physically much smaller Amazon Basics cooler | |
| Transitions | ||
| GN | moving on from coolers last month LTT published a Labs video where they got the model of the power supply they were review in wrong | |
| GN | LTT mixed up the Thermaltake smart 80 plus and the thermaltake smart bm2 which are drastically different power supplies with different quality levels | |
| GN | they also mixed up the 80 plus 115 volt and 230 volt markers on the graphs | |
| GN | which have different requirements for the efficiency percent at a given load | |
| GN | this is something they caught after publishing | |
| GN | so that much is good pending a correction is the right move and they haven't always immediately done that like we saw with the RTX4090 review | |
| GN | so that's an improvement | |
| GN | but they need to at least get these basic attention to detail items right this isn't even the hard part | |
| GN | besides if you can't get this much right then as Linus as himself says | |
| Linus | and doing them halfway is basically meaningless | |
| Transitions | ||
| GN | while we're on the topic of basic errors let's do one last one | |
| GN | we randomly clicked on a short circuit video to check for accuracy | |
| GN | there was literally no effort in finding this one and we landed on a 5600X3D review | |
| GN | the video host States | |
| GN | quotes | |
| Jack | the lab put it through its Paces | |
| GN | indicating this is a piece that the lab specifically contributed to which is marketed on accuracy | |
| GN | within the review LMG initially States correct 5600X3D specs | |
| Jack | but it has 99 megabytes of cache | |
| GN | but incorrect 5600X cache specs | |
| Jack | the non-3DV cache version has 36 | |
| GN | then it displays incorrect 5600X3D specs on the screen later | |
| GN | and it it lists the 5600X3D is having four megabytes of L2 cache except the 5600X3D only has 512k of L2 per core totaling 3 megabytes not 4 megabytes | |
| GN | they also listed as having 99 megabytes of three except it has 99 megabytes total of L2 plus L3 cache at 96 plus 3 | |
| GN | they then state that it's one megabyte lower cache to the 5800x3d | |
| Jack | you only lose one megabyte of cache | |
| GN | despite showing a graphic that has the 5800x3s lower cache than the 5600x3d | |
| GN | just a total mass and it's complete chaos | |
| GN | in March LTT published its 7950X3D review | |
| GN | LTT blamed a quote bad CPU and noted poor performance for the delay is on the review | |
| GN | this entire video confused us because it seemed like they were hyper worried about getting the right answer as compared to AMD | |
| GN | but they should be more focused on producing their own numbers that is if they trust their own numbers | |
| Linus | AMD had every opportunity to contest our findings | |
| Linus | but when we showed them our results | |
| Linus | they basically went yeah seems okay | |
| GN | that's because it is okay if you're trying to nail the exact differences down to a single percent of a first party claim | |
| GN | then you aren't providing any value as a third party reviewer | |
| GN | AMD might as well publish its own review | |
| GN | it's not weird at all to have a few percentage points different or even upwards of 30 percent | |
| GN | they make the thing they're gonna test it in ways that benefit | |
| GN | them then they said | |
| Linus | this so if nothing else to go on if we had launched that review we would have been the only outlet with a broadly negative take on this chip | |
| GN | which is also just weird because they shouldn't be worried about popularity contest | |
| GN | they should be worried about reporting the truth of their own findings and then trying to explain them if they're different | |
| GN | LTT noted in its email to AMD which they show briefly on the screen | |
| GN | that part of their stability quote problem was enabling Expo without increasing V dim | |
| GN | except it is normal that higher clocks of memory won't be stable without higher V to MVD and vddq | |
| GN | that's not anything unexpected nor is it AMD's fault nor is it new | |
| GN | the AMD Madea jisa menu for OC and BIOS worked at this time and is what should have been used to avoid these pitfalls as it would Auto apply | |
| GN | that test processes do not account for this check for this early and just blindly trust whatever it is the motherboard manufacturer menu is doing is | |
| GN | once again indicative of poor experience and blaming the CPU is not the way out | |
| Transitions | ||
| GN | the next brief section is for asterisks errors which are added during Productions that's after filming and before publishing | |
| GN | you can't do text on the screen anymore in YouTube after you upload | |
| GN | so all of this is done before the upload | |
| GN | which means that when doing an asterisk correction a Creator is aware of that correction | |
| GN | however these have to be used carefully because most people won't read them they might just listen to the video or kind of tune it out | |
| GN | so these shouldn't be used for undoing criticism that was unfair | |
| GN | this past week Linus Tech tips showed an all Alienware setup with some asterisk errors in one | |
| GN | instance LTT states that the keys have quote stickers | |
| Linus | curiously cheap printing on some of the keycaps | |
| LTT STUFF | actually is double shot PBT | |
| Linus | yeah they're double shot except for they put stickers on them | |
| LTT STUFF | oh my gosh yeah | |
| Linus | yikes | |
| GN | and expresses disappointment | |
| GN | but the editors add a note that says not stickers immediately dismissing apparently the disappointment | |
| GN | in another asterisk correction LTT States | |
| Linus | and a 470Ti meaning we should be able to get some pretty good gaming performance out of this thing | |
| GN | but the editors had an asterisk that says not Ti | |
| GN | for hard criticisms that are stated by the host or the reviewer like the keycaps example | |
| Linus | yikes | |
| GN | serves no value or purpose to come and then put an asterisk to correct it later | |
| GN | you either cut it or you correct it in video | |
| GN | and going back to those interviews we think they're doing these Corrections because they're moving so fast | |
| GN | that they won't budge at the time to go back and properly correct something like with the new voiceover line | |
| LTT STUFF | I wish we could back off the amount and focus on quality | |
| GN | when it's clearly entertainment that's fine | |
| GN | but if you're going to start scrutinizing something it Entertainer or not | |
| GN | Last week LMG's Tech quickie Channel covered large language model hardware and within the four minute 45 second video | |
| GN | they had an on-screen tax reaction from 312 teraflops to 624 teraflops a big difference | |
| GN | then they had a related correction later | |
| GN | this is an extremely short video | |
| GN | three take two sentences just for audio would have taken a few minutes to send the host down to the studio | |
| GN | this reinforces our thesis statement that LTT is rushing its content and is pressured to maintain such high throughput that it can't slow down | |
| GN | even for a single sentence to be re-read properly or that it is unwilling to pay for content to be accurate | |
| GN | as long as it fulfills a good enough bar and don't be fooled by LMG they internally set the standard to upload at the frequency they do | |
| GN | it is therefore a willful decision to forego quality in order to hit the standards that they're setting | |
| GN | five more minutes is five minutes too many to get it completely right and that costs money | |
| GN | it will be one thing if they didn't realize this until after publishing | |
| GN | but realizing two mistakes within under five minutes of video prior to publishing should indicate a proper correction | |
| GN | in April LMG published a short circuit video that was paid for by Asus featuring the Asus 4070 Dual | |
| GN | the problem isn't the part where they refer to a 600 GPU as aquote budget card and then thank Asus for gracing us with the presence of a one dollar backplate | |
| Jack | it's got an aluminum backplate which is really nice to see on a budget card like this | |
| GN | or the clearly biased language where they give a manufacturer a blanket statement for are always being good | |
| Jack | it's an Asus card so it's going to perform great | |
| GN | but rather the problem is the simple correction for a misstatement on the amount of HDMI and DisplayPort available | |
| Jack | 3 DisplayPort and 2 HDMi on the Dual | |
| GN | they could have just cut around this | |
| GN | we're not sure why they left it in | |
| Transitions | ||
| GN | and all these post shoot pre-publishing Corrections bring us to the next set of concerns | |
| GN | which is an increasing amount of irresponsible actions and conflicts that we've seen emerging from Linus Tech tips | |
| GN | so we've been seeing an alarming amount of conflicts from Linus Tech tips as it relates to their corporate connections | |
| GN | their flow of money and the potential bias as a result of those things | |
| GN | that we've been looking into this maybe more of a future video thing | |
| GN | but some of the concerns include how Linus is personally invested two hundred thousand dollars into laptop maker framework | |
| GN | and is also covering their product and their competitors products | |
| GN | LMG also has a cross-branded again screwdriver with Noctua and that's been sold on their store while they're also reviewing Noctua products favorably | |
| GN | even when it's shown to be hidden by their own data 100 degrees Celsius | |
| GN | which becomes problematic for Brands like Noctua well | |
| GN | because now it is less clear if the praise is genuine or if it is biased whether intentional or unintentional | |
| GN | these issues combined with high-level corporate connections through the CEO and through the head of labs raise a lot of concerns about conflict | |
| GN | but we won't be getting into them in depth in this video instead we're going to focus on the consequences of irresponsible reporting and bad data and how that influences the purchasing decisions of millions of people | |
| GN | to start with let's discuss the implicit ethics of knowing a lack of qualifications yet continuing to operate on a certain type of content | |
| GN | or being unwilling to commit the time to do that content properly | |
| GN | especially if you know that there are our flaws and shortcomings already and choose to forge ahead | |
| GN | we don't really see any other options here it is either too rushed to the point of potentially producing harmful information to viewers | |
| GN | and to Consumers that they then act on and to manufacturers alike or the company is unqualified to produce that information | |
| GN | it's really one or the other | |
| Transitions | ||
| GN | this one talks about the ethics and power of a review outlet and therefore its responsibility to do due diligence on a product | |
| GN | on June 24th LTT published a video about a copper water block made by a small company called Billet Labs founded just last year | |
| GN | LTT took one of the company's first prototypes to test | |
| GN | valued in the thousands of dollars for manufacturing not counting opportunity value | |
| GN | in the video the LTT team stuck this special 3090Ti water block on an RTX4090 | |
| LTT STUFF | if this wasn't a 3900Ti then we don't have another 3090Ti that we can take apart | |
| Linus | how the hell did we not know what GPU this thing was | |
| LTT STUFF | I thought it was a 3090Ti | |
| Linus | so we just casually had a 4090 sitting on the shelf clearly labeled | |
| GN | stating that it should work despite it not being explicitly compatible | |
| GN | the company did not tell LMG that it would be 4090 compatible | |
| GN | they told LMG that they could try to use a 4090 but Billet Labs itself had not tested one as it didn't have a 4090 at that time | |
| GN | water blocks have very specific tolerances and clearances for components | |
| GN | so any misalignment would cause poor contact to the GPU core | |
| GN | LMG decided to try it and continued with the video | |
| GN | despite misrepresenting the product massively by using it on an incompatible device that it literally would not fit properly | |
| GN | LTT's presentation goes down the path of being irresponsible | |
| GN | presenting the product is much more difficult to work with than it would be if used properly | |
| GN | the conclusion is like | |
| Linus | it's a cool product | |
| Linus | but it's a bad product | |
| Luke | it looks great | |
| Linus | yeah it looks super cool | |
| Luke | are you just thinking | |
| Linus | today | |
| Luke | I don't think just purely | |
| Linus | it's bad because it makes absolutely no sense and nobody should buy it | |
| Linus | most of the communities actually amazing | |
| Linus | but there's a vocal minority that looks at something like that and goes lazy out of touch | |
| Linus | you know whatever else no I'm just really experienced at this | |
| GN | it found this conclusion on the basis of improper installation | |
| GN | this serves no one it needlessly harms a small company and this kind of behavior can kill someone's life work in a heartbeat | |
| GN | and hasbased on invalid testing and presentation | |
| GN | one of the company Founders posted below the video and stated this | |
| GN | quote | |
| GN | we love LTT but it was difficult to watch their criticisms when pretty much every design issue that Linus raised was done for a reason | |
| GN | which was all explained to Adam weeks ago none of which I guess was feec back to Linus | |
| GN | we also sent a full instruction manual that seemingly wasn't used | |
| GN | I think there's definitely been a major Communication Breakdown on the video pipeline | |
| GN | we assumed we'd have a chance to answer any problems that were raised and explain them before it being publicly misrepresented | |
| GN | LMG faced some controversy from its community at which point it replied on the following WANshow | |
| GN | doubling down despite a complete lack of accurate or even useful testing | |
| Linus | nothing would have changed | |
| Linus | nothing that obeys the laws of physics would have changed changed the conclusion | |
| GN | that statement is already a reckless and problematic doubling down considering the elimination of any possibility to change the conclusion | |
| GN | including a 20 degree reduction which should change the conclusion | |
| Linus | if the GPU had been five degrees cooler that would have been the conclusion | |
| Linus | if it had been 10 degrees cooler that would have been the conclude | |
| Linus | if it had been 20 degrees cooler that would have been the conclusion | |
| GN | and we're not alone and saying they should have just done it properly | |
| GN | at least then you have grounds to stand on for criticisms | |
| Luke | I also get not wanting to retest | |
| Luke | but I do think probably just testing it with the thing that they knew it worked with in the first place probably would have make sense | |
| Luke | because like because like you're saying the test result itself doesn't matter | |
| Luke | but if it's like if it's just supposed to look cool then it should be seated properly so it looks cool | |
| GN | about five minutes of rambling the conclusion effectively was | |
| GN | lol it still sucks though | |
| GN | Linus then after a publicly humiliating and trashing the work of two guys just getting their start in business | |
| GN | while relying on zero effort to examine the thing correctly complains | |
| GN | about how he'd have to spend a hundred dollars to 500 more on an employee's time to test it properly | |
| Linus | but I'm | |
| Linus | I'm not | |
| Linus | I don't know guys | |
| Linus | I'm not sure if I can apologize for not spending another 100 200 300 500 of various people's time | |
| Linus | sitting and Engineering a workaround to a product that no matter the result nobody should buy | |
| GN | now we don't have any connection to Billet labs and that comment was offensive even to us | |
| GN | so we can't imagine how heartbroken they must have been | |
| GN | Linus in order to protect his company's bottom line was willing to potentially inflict immeasurable harm to a startup | |
| GN | rather than spend by his own admission a couple hundred bucks in someone else's time out of the 100 million dollar company's coffers | |
| GN | making matters worse in conversations with Billet Labs | |
| GN | Gamers Nexus learned that the company asked for the Prototype back on June 28th | |
| GN | this would allow them to send it out to other Media or use it for further development | |
| GN | Linus Media Group agreed at least twice to return the product | |
| GN | the first time they agreed to send it back was June 30th | |
| GN | at LTX Linus media through put the one-of-a-kind water block up for auction at its extra life auction event | |
| GN | without the permission of Billet Labs | |
| GN | Billet did not learn of this until around August 11th at which point it was told the block is gone | |
| GN | again this is despite LMG agreeing in late June to send it back as LMG would have no further use for the block | |
| GN | Billet told us that it is now stalled as it no longer has its best prototype available for continued development | |
| GN | it's also missing one of its GPU | |
| GN | we voiced a Billet a concern of a potential competitor | |
| GN | many of whom were at LT export was auctioned | |
| GN | potentially getting a hold of the block to send it off to be cloned | |
| GN | whether or not this happened is irrelevant | |
| GN | the point is that there's a responsibility that was ignored | |
| GN | the question that you have to ask as a Reviewer is am I being fair when making this type of commentary | |
| GN | that's regardless of whether you are just an Entertainer or you have your Entertainer hat on for that video | |
| GN | but not the next one and this is particularly important for anyone marketing the future of their businesses success | |
| GN | on the back of objectivity fairness data-driven reviews | |
| Transitions | ||
| GN | another example of this just popped up recently as we were finishing this video | |
| GN | in the short circuit video where Linus Media Group reviewed | |
| GN | a mouse in there not a review the company made strong criticisms of the friction of the mouse in the video | |
| LTT STUFF | this is not a smooth Mouse and that's really disappointing | |
| LTT STUFF | especially on a premium gaming mouse that should be one of the primary focuses | |
| GN | Pwnage then contacted lmg to notify them that LMG had not removed the protective film from the feet of the mouse | |
| GN | Linus noted in the WANshow that the short circuit team then posted a pinned comment | |
| GN | saying that in fact there weren't any stickers on the mouse at the time of testing and that they stood by their commentary of the product | |
| Linus | the bigger issue here is that our initial correction which was posted | |
| Linus | said | |
| Linus | hey no it | |
| Linus | you know is high friction and we did test it and there were no films on the feet | |
| GN | Pwnage then tweeted publicly with an abrasive statement but one for which we can't fault them | |
| GN | they point out the irony of being chosen as a gift for ltx's whale and attendees despite the conclusion of the video being this | |
| LTT STUFF | the only question that remains is am Irecommending you buy this product | |
| LTT STUFF | no | |
| GN | this we don't think is about buying favor | |
| GN | but rather the dichotomy between being Chosen and then having such an egregious error unfairly undermine the Integrity of the product and its Chief marketing point | |
| GN | LMG then eventually began the process of replacing the video in place once again | |
| GN | we assume this time with YouTube's trim tool | |
| GN | that'll take a day or two to process they're cutting the section about friction | |
| GN | but this doesn't fix the problem they are choosing to leave the content live whose conclusions were at least partially based on massively erroneous work | |
| GN | because they don't want to hurt their video performance | |
| GN | people continued to see the original until it processed | |
| GN | that can take a day or two | |
| GN | but they're willing to throw Pwnage under the bus | |
| GN | One Last Time by providing feedback that the stickers should be more obvious | |
| GN | for an end user maybe but an end user doesn't get paid to test a product whether or not it's branded as a review | |
| Transitions | ||
| GN | all of this is ethically unacceptable | |
| GN | they like to Brand their short circuit stuff as unboxings | |
| GN | we're not other words | |
| GN | not reviews | |
| GN | but it's disingenuous to hide behind that whenever there's a mistake | |
| GN | this magnitude of error is what the mouse or in the case of the water block | |
| GN | the block is marketed on which is either Glide or thermals and to screw the strongest marketing Point | |
| GN | through such an egregious error as not taking the tape off of the product and still leaving the video up even | |
| GN | while processing a cut is callous to the impact that those choices have on these companies | |
| GN | Linus Media Group is a massive that has to be treated responsibly | |
| GN | there are correct actions it should result in some sort of immediate trigger | |
| GN | like a removal of the video with a follow-up saying we did it again | |
| GN | here's the new results | |
| GN | what's even more disturbing though is the community response every time someone has a criticism that's valid like | |
| GN | when the community beat down Hardware unboxed for voicing their concerns of LTT's testing accuracies issues which while not completely tactful | |
| GN | yes | |
| GN | we're valid then you look at the recent counts though like this | |
| GN | quote | |
| GN | if a professional Mouse reviewer couldn't figure it out what makes you think the average consumer will find the plastic wrapper cover | |
| GN | quote | |
| GN | short circuit is in a review channel it's an unboxing and first impressions Channel with some Labs testing | |
| GN | the conclusion was just David's opinion he had more issues with the mouse than just the feed | |
| GN | quote | |
| GN | and if your product can be unboxed by a tech reviewer who does this for a living | |
| GN | and that reviewer doesn't even realize there's packaging that needs to be removed before use that's bad user experience design of the product that should be called out as a negative | |
| GN | yes | |
| GN | but calling it out would require noticing it as a prerequisite | |
| GN | which a media Outlet taking its time would do | |
| GN | just because pwnage didn't phrase it in exactly the nicest way doesn't mean their point is invalid | |
| GN | the same went for a hardware unboxed with their concerns | |
| GN | the Pwnage was defending themselves they apparently according to Linus tried it the nice way first before they wrote that abrasive tweet | |
| GN | and and it's after getting humiliated publicly a second time through the pinned response the first one that basically said no our testing was right | |
| GN | and then people noticed that you could see the film reflecting in the b-roll OR at least we did they | |
| GN | then sent that tweet out and that it had no attack is irrelevant | |
| GN | so we're not going to try and pretty it up | |
| GN | it is unfair | |
| Transitions | ||
| GN | it's unfair to Consumers who get misled to bad purchases or not understanding the product | |
| GN | it's unfair to companies when there are major points that are erroneous and aren't properly addressed or handled | |
| GN | and it's unfair to other reviewers who actually are willing to invest the time | |
| GN | and even a hundred dollars or five hundred dollars or whatever it takes in some cases to get it done right | |
| GN | and it's sad to see | |
| GN | because it undermines all of us | |
| GN | they are prioritizing publishing on a fixed schedule which they have self-imposed | |
| GN | this isn't | |
| GN | the YouTube algorithm Rhythm isn't some magic thing | |
| GN | that actually needs daily uploads | |
| GN | does it help | |
| GN | yeah sure | |
| GN | but there's a lot of things you could do in business that would help that are maybe not the right thing to do | |
| GN | I'll be real here | |
| GN | it's getting difficult for those of us who try to do things right and we're not the only review Outlet here | |
| GN | we're not alone in this to weigh it into a product conversation especially after LTT | |
| GN | where we have to tiptoe around errors they make and write scripts around it to kind of try and subtly correct it | |
| GN | considering the most devout part of the community is propensity to lash out like you're seeing against these two companies defending themselves | |
| GN | and we don't care anymore | |
| GN | we're just going to say it like | |
| GN | it is being new to testing it doesn't responsibility of the content just because you're new to this certain branch of business | |
| GN | and they have to continue to respect the position that the company is in | |
| GN | especially when building the business's future on the back of labs data | |
| GN | as a leader you're responsible for what you say and whether or not you are branded as the CEO | |
| GN | you still have to feel that responsibility for saying the things that are said | |
| GN | Linus Media Group has grown into a massive Corporation | |
| GN | it's hired a seasoned CEO to replace Linus itcontrols over 120 employees | |
| GN | they affect the purchasing decisions of millions of people and they brand themselves as an unbiased for the people outlet | |
| GN | and this is a an admirable goal | |
| GN | but the reality right now is that Linus Tech tips and irresponsible way | |
| GN | with these bad data points and and conclusions and either laziness or an unwillingness to commit that time cheapness whatever is focused on the bottom line | |
| GN | all of these things | |
| GN | these irresponsible actions they can affect in the worst case as the life work of an individual say a smaller company or in the more common cases | |
| GN | it can just mislead consumers into spending money that they don't have to spend or on things they don't realize | |
| GN | or maybe not the right fit for them because of these problems | |
| GN | having deep and high-level corporate Connections in companies and going easy on manufacturers | |
| GN | while then briefly after taking their money for a corporate title sponsorship of a major convention | |
| GN | investing in laptop companies and covering their own products alongside competitors | |
| GN | these are actions that to us raise serious concerns about the objectives of the company | |
| GN | is it in fact consumer first or is it more looking at the bottom line | |
| GN | the fun Linus personality in front of all of it doesn't make it simply okay to do these things | |
| GN | to publish bad data so regularly and knowingly in some cases like with the block and not take the time to do it properly | |
| GN | no matter how funny and goofy the video may be it simply isn't okay | |
| GN | the data matters the consumer matters the perspective of the consumer also matters now viewing this company as a a sort of reforged not just entertainment company | |
| GN | and the truth matters | |
| GN | we think there's an eagerness in the less experienced employees coupled with a lack of disciplined and experienced management to steer them in the right direction | |
| GN | and check for mistakes | |
| GN | so some of this is fine if you are just an Entertainer | |
| GN | but if you start marketing your business on accuracy and data and objectivity like those of us who've occupied the space since the website days | |
| GN | then we should all be held to the same standards we should all play by the same rules and audience scrutiny and criticisms | |
| GN | and likewise we as reviewers have to hold ourselves accountable to the same standards | |
| GN | because again it protects all of us from manufacturers being | |
| GN | like these guys don't do anything useful | |
| GN | so that's kind of it this is my thoughts on this | |
| GN | and it's I want to get back to testing stuff | |
| GN | but this it was important to us and I hope that people listened before they just jumped to the OMG drama wall comments | |
| GN | but that's it | |
| GN | thanks for watching we'll see you all next time |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment