Absolute gigachad. In the middle of a cheating investigation, gives everything asked for, continues to do runs with increased proof standards, crushes WR with a hand cam.
Yeah I almost cried when I found out he finally got his sub-5 among all of this absolute nonsense It's like he got so annoyed at Brood that it doubled his focus lmao
@@sadscientisthououinkyouma1867 giving nothing implies guilt, and there's no witch hunt, just one dude with chat gpt. Most others assumed innocence but wanted to do their due diligence.
@@AceWolf456 "giving nothing implies guilt" in what world? Prove to me you haven't murdered anybody, the format of which I demand uninterrupted stream of your entire lifespan such that we can verify for a fact you are not lying. If you refuse to meet my simple demands you are implying guilt (for not submitting). EDIT: also "most others assumed innocence but wanted to do their due diligence" is just a contradiction. You can't assume someone is innocent while also investigating them, because the suggestion they are innocent would mean they are not worth investigating. If he's innocent why investigate? If there's a chance he was not innocent and that justifies the investigation, then the assumption isn't that he is innocent but rather the assumption is he could potentially be guilty.
Blink was treated completely unfairly. I figured this video would end with Brood sheepishly admitting he was wrong and apologizing. But no, even after wasting everyone’s time with this nonsense, he still insists it’s cheated. Must have found a new ChatGPT model to learn from. Have some humility and give it up
I'm with ya. I can understand being convinced of something and then finding out you were wrong and apologizing. Not the end of the world! It happens!! But when you can't even admit you were wrong and say "yah looks like I messed up, sorry bros!", there's something seriously wrong with you. Just so, so, SO disrespectful of everyone's time and energy.
Yeah, this guy was a desperate, bad faith actor the entire time. Really wanted to be some cheat-busting legend, or really hated Blink for some reason, or both. Makes sense for someone who would use AI to try and learn something as in-depth as audio analysis. I hope he's banned from the community, but I doubt it.
@@idontwantahandlethough for that alone. this mf should be banned from accusing people. they aren't doing it for the good of the sport, but for clout. and unhonorably
Its been a problem in my university. It can be really useful when you already experienced in the field and can shift through the bullshit it gives you, but the amount of people submitting work entirerly based on it is way too high
I've gotten into political arguments with people online (poor choice, I know lol) and they used chatGPT as their "rebuttal". It was pretty clear they had no idea what they were talking about before that, but it _certainly_ was obvious afterwards! People are wild. What's even the point of life if you're not going to bother learning anything?! Learning new things is like the best part of life!!
once again the dangers of AI are apparent. In here, the stakes were relatively low (one person stood to lose their credibility in a niche community), but imagine if this perception that AI is a reliable source of information is used in law (we've seen this happen, actually - chatGPT simply invented a court case), or medicine, or other high-stakes environments where wrong information can lead to someone losing their freedom, or even their life. AI is NOT reliable. Stop using it.
@@NameHere2243 exactly, as long as you already kinda know the answer you expect from the LLM and are ready to "shift through the bullshit" (I really like that wording), it can be ok. But LLMs are very convincing in the false informations they output and will mislead users that lack basic knowledge on the topics they are asking them about. It is sadly also a problem in my university, with students failing to understand they LLMs are not magic, and just not questionning anything chatGPT tells them :/
Then imagine using your chatgpt expertise to tell people that actually do this for a living that they wrong and you just presented your argument wrong. Dude never had an argument
Deciding you are right and then looking for supporting evidence is NOT how this is done, especially not by using a technology that is well known for its ability to be painfully wrong, when you have a hypothesis, the first thing you should be doing is trying to prove it wrong.
Guilty until an LLM proves you innocent, then ignore the LLM and proceed to accuse the runner publicly anyway. Have every expert in the community prove you wrong, stand by your false accusation. ???? Profit
Maybe a hot take, but if I was a mod of this community, based on what's presented here, Brood would be getting a ban for at least a year for this BS. Tormenting a community member, wasting multiple people's time trying to research his allegations, blatantly ignoring mod's instructions to not publicly post the allegations before the mods do; all the while not being able to confidently stand behind what he's accusing without a chatbot there holding his hand. Absolutely baffling behavior.
Completely agree. If the allegation had any basis in reality, an apology would maybe be enough, but given how much time of everyone’s he’s wasted with his stupidity and the fact that he’s still not admitting he was wrong, there needs to be some serious reprimand
A bit of a hot take. Assuming lessons were learned and no malice, I think they should just move past it. Witch-hunting the guy for revenge isn't going to make anyone but uninvolved parties feel better by satisfying their bloodlust. Maybe Brood could make a public apology, though
@@displayer6023 I think giving them a time-out from the community might be for the best for both sides. It gives Brood some time off to think about things and also makes sure they aren't around for any potential witch-hunts. By the time they get back it would all have calmed down. Like, even for Blink themselves, imagine how frustrating it is to be accused of something you didn't do, and have to go through a full on investigation where doubts are cast upon you, and there's even supposed evidence, even though you know you never cheated. Honestly I wouldn't be surprised if this killed a bit of their passion for speedrunning this game. This sort of stuff doesn't go away with just "oopsie i guess i was wrong, anyways" The fact that Brood published their findings before mods could have a proper look is probably the worst part. This could all be avoided if they just listened to the people who actually know what they're talking about. I don't want to speculate why they did that, maybe just stupidity (Hanlon's Razor and all that) but the fact that this was very avoidable is still there.
The biggest sin of all this isn't the accusation, nor the ChatGPT usage... it's the attempt at public humiliation by going into his stream and calling him out.
In my opinion there’s nothing wrong with using chatGPT to become an amateur at something but please realize - you’re an amateur as much as someone who just did random tutorials.
Yeah that part really bugged me too. There’s nothing wrong with holding a run up for closer inspection, that keeps things fair for everyone. But the public call out with poor information before his “proof” was even out was really crappy.
I'd still say it was the ChatGPT usage, but yeah, what a damn troglodyte this guy was. He was so hungry for that big GOTCHA that he made himself look like a tool even before the GPT reveal.
they say IRL, the punishment is the process. If you got accused of murder and had to go through a whole trial, just because you were proved innocent doesnt mean its all ok
And even when someone's proven innocent, it doesn't always mean things will go back to how they were. Court of public opinion means some people will think "Well you were accused, so you must have done SOMETHING sketchy", or people will spread misinformation about how you were proven guilty, and people who know absolutely nothing about the situation will either be driven away, or jump on the mob bandwagon blindly.
I got accused of stream sniping after dropping in during a match twice; once at start to say glhf and again later because a player gave up during match and I wanted to see if I could tell they were a group. I replied to something (at this point just standing around cause I was going to let them all go) and suddenly streamer and chat started accusing me of sniping even though I had done nothing but message twice and my gameplay showed no hints of ss. It really hurt my feelings so I just decided to end the game instead of letting them go; I had left the stream again by that point. But because of how Twitch handles the Viewer list (not updating for sometimes 4 whole minutes even on refresh) they accused me of still being there. Even went on about me needing to "mute stream" to not hear callouts when 1; delay is so bad any info on your location in game is void, and 2; they had one callout that entire match and again delay. I didn't find out till later watching vod they had called me a clown and a whiner and kept accusing me of sniping, how disappointing it is when it's one of their own (vtuber), and banned me. It really really hurt and ate me up inside for WEEKS. I even had my own VOD as proof I didn't snipe since it was my first month streaming. I didn't even want them to admit they were wrong publicly I just wanted an apology for assuming and being so rude. It took forever to convince myself not to let it eat me up so much and that it wasn't worth all my energy. The accusations stick for a LONG time when you know you're innocent and it hurts because people won't just listen or take 2 seconds to look even just a minute longer at the proof.
@@Someguyhere111 Don't turn this one situation into this big thing lol. I'm pretty sure that once people find out that Chat EBT was used, they wouldn't take them seriously
This is an obvious attention seeker who publicly accuses someone of cheating on baseless grounds using arguments that he was absolutely unqualified to make. What's more offensive is the amount of effort that everyone had to take in order to disprove this nonsense, not to mention the mental toll it takes on you being accused. Disgusting behaviour.
Now take what you just said and apply it to medicine and vaccines. So much time and money has to be waisted proving people who should never have had a platform in the first place wrong.
26:58 >"If there is spliced footage, I think you should be permanently banned. If I somehow made an error and you are exonerated, you should be able to do whatever you like" huh???? And the guy who falsely accused a man using techniques he had no idea how to use and relied on ChatGPT for stays????? Please tell me that the guy got banned for falsely accusing him with flimsy and garbage evidence.
Oh my GOD I didn't even catch that part. He sounds so self-righteous. I would have left the community out of shame after that... but clearly he's built of a different type of IQ points...
I really wish this was pointed out in the video because it really shifts how Brood was approaching a possible cheating run. He doesn't even seem to think it was possible he was wrong in that same paragraph. That and the fact he doesn't seem to accept that the experts don't think there is any evidence of cheating and instead says he's getting better at his technique. I wouldn't be surprised if he accuses another innocent person. Definitely should get banned to some extent to prevent further occurrences, especially since he didn't listen to requests to not make the accusation public before the mods got a chance to look over the evidence.
I actually found what we saw of his logs with ChatGPT fascinating. Perhaps I'm reading too much into it, but it felt like a lot of his "questions" to ChatGPT were pointed towards finding the results he wanted, and not what was actually there. He had already decided Blink was guilty, and wanted to know how to make the "evidence" say that, instead of checking if the evidence actually showed he was guilty.
I was shocked that Brood wasn't banned at the end of the video. It's a big deal to accuse someone of something without knowing how to do it, let alone defying a moderator's request that it not be posted publicly.
I support a fact checkers existence to keep runs legitimate. It's good brood is around. He wasn't hostile and was professional about his approach. As someone with that investigative mind myself I want to be wrong so tell me why I am. If I am its a lesson for me that shouldn't lead to a reputation altering incident. Innocent before being proven guilty.
@@nightwintertooth9502 Fact checkers and validators are important, but I have to ask: are you high? Brood was SO far from professional, and made a mockery of the people who ACTUALLY do stuff like this by thinking he could supplant years of education and knowledge by using ChatGPT. He needs to be laughed out of the community, not extoled for being a "professional", because he wasn't. Maybe you just randomly clicked around the video, but the ACTUAL professionals are the kind of fact checkers we need, not a con artist like Brood who doesn't give a shit about the community and just wants to try and get internet famous for catching a cheater. There were plenty of ways to do this right, and Brood did none of them. Begone, shill.
@@nightwintertooth9502blink himself has talked about how extremely stressful it was to be falsely accused, even with flimsy evidence. Brood WAY jumped the gun and ignored mod directions to let them review his evidence privately. He could have had his evidence reviewed and achieved the same result without wasting a LOT of peoples' time and putting blink through all this.
For me the biggest indication that he did not cheat is the absolute cooperation and participation in the investigation. Anytime someone cheats and then people ask them questions about their setup or for help analyzing their gameplay, it's always "oh my setup is in another state, my brother took the console, so now I don't have it, and I'm using a different cart and cables and..." etc. Instead the dude described every part of it in detail and even loaded analysis tools onto a flashcart and provided every sample that was asked for. A cheater would never. He's as legit as it gets.
Honestly I can understand if a regular person does not feel the need to buy a flashcart so they can load it with analysis tools in the event that they are accused of cheating as it's not really a scenario a normal person would think of only a paranoid person. Like for some systems such as a GameCube you have to physically mod it in order for a flashcart to even work and most people don't have the know how or even technical skills to do that as improper modding can damage your console so a lot of people would even be hesitant to even attempt it
blink's cooperation with the investigation is a huge indication of his innocence imo. no cheater would record and share so much device information if they were actually splicing. can't imagine how it felt to be called out so publicly like this and then learn his accuser used chatgpt as his main source of learning about audio engineering, that's so wild
I agree, but only a little bit. An actual cheater might still be very capable at gaslighting people and blowing smoke up their asses by leveraging their popularity. My first thought immediately goes out to Dream, who (other than getting defensive) still offered various files and other bits of information to the people investigating his cheating. Same goes for actual criminals being interrogated by law enforcement: often the right to remain silent is skipped because they are super-eager to prove their own innocence in all the possible ways. Just because someone is helpful doesn't mean they aren't guilty.
@@Aviertjei dont think 'helpful' is the right word to describe a liar's attempt to weasel out of trouble, but i agree with your point. I think the word 'cooperative' makes more sense here.
@@Aviertje i get your point but even dream had his uncooperative moments, iirc he said something like "I can't give you my mods folder because we routinely delete everything in it for our minecraft videos" which *could* be possible but it just sounds like something he made up so he didn't have to show his mods folder. And in the end, turned out it was because he was indeed using a mod. Sooner or later you'll hit a sore spot which they can't give you without revealing themselves. Meanwhile with blink, there was not a single point in the video where he refused to provide something. He recorded like three baselines, gave device information, ran several tests... Idk it just feels too open for him to be suspicious. Though of course the real proof is in the data gathered, not his character.
LLMs can be useful if they point you to a resource. They have an enormous knowledge base and can find things for which you don't know the right words for. However, relying SOLELY on an LLM is very stupid, as LLMs are still nowhere near as good as humans in reasoning ability (they have a heavy bias toward what has the most text devoted to it as being the most important). LLMs are a great search tool, and can help find particular passages in long documents that talk about a concept - without being hidebound to specific words like a word-search is. But LLMs can't outthink humans, and in certain ways won't ever due to their focus being skewed through a language lens.
Brood should be punished for not only terrorising another speedrunner, but for also breaking the rules when it comes to publicizing his findings. It is ok to be skeptical, but not when you cause so much distress only to be wrong.
In honor of this man trying to bust a speedrunner with chatgpt, im going to prove abyssoft fakes all of his videos using reverse psychoanalysis that my grandfather learned selling nfts to a paralegal
something EVERYONE is missing. When confronted on the issue of getting good knight luck, he says he only watched 2 videos. All you had to do, is look at the last 10-20 knights on completed runs, and see how frequently they are lucky. All this sound engineer stuff is assuming their was inordinate luck in the first place. Ofc there will be a correlation, a good early is more likely to be completed, but you would expect a splicer to cheat their best runs, otherwise they wouldnt need to cheat. Zeals logic was...sound here.
@@Exodyr This exactly. Of course an early game boss will have better luck than others in the run, you're more likely to reset if the early game goes poorly, since you've committed less time to this run
There’s a reason why, during the Dream cheating debacle, his accusers didn’t just take the drop rates from one or two runs with suspicious results. They collected data from thousands of runs, even runs which ended prematurely, to prove that there was no conceivable way that random odds would produce such results.
@@jellywillreturn oh it was by no means thounsands of runs :) it was all the runs on a few streams, but mpre than enough to mathematically proof it. Stand up maths has a great video on the dream cheating scandal and its super interesting
I’m not even shocked at the use of ChatGPT. My students turn in AI slop all the time and usually can’t answer basic follow-up questions, because they are using it as a replacement for knowledge rather than as a tool for enhancing it. Shame on you Brood.
And before that they would just copy/paste wikipedia. I remember one presentation used the phrase "specific gravity" when talking about a battery. The teacher asked them what specific gravity is, and they had no answer.
Bro needs consequences for this. Wasting a tremendous amount of time of far more qualified people, and undoubtedly caused unnecessary stress on someone.
imo, asking a speedrunner in a long speedrun thats trying to shave a couple seconds off to break a milestone to "open a menu" is a dick move. Hes a few seconds shy of sub-5hour and you want him to waste a few seconds. I get that cheaters are pretty common in speedrunning...for some reason...but man i'd be annoyed in Blink's scenario at that request.
> ... for some reason... In the words of the (absolute legend) Karl Jobst, "Runners don't cheat to get a faster time. Runners cheat to get a time, faster."
Cheaters are common in anything that is competitive, from 'roids to riding buses during marathons to aimhacks on-stage during fps tourneys, splicing runs and having 3 people submit runs under 1 name, if it's a competition people will find a way to try and get an illegitimate edge.
Definitely a dick move. "Oh, you're in the middle of a highly complex, high stress performance? Stop and pay attention to me instead, or else you're cheating!"
"Hey chatgpt, tell me if this person is cheating and also give me insanely bad information about audio editing" Calling this a "scandal" is generous. This is like someone going into a 50 page document about how a Mario speedrun is cheated because he totally touched the piranha plant and I asked chatgpt if touching piranha plants was allowed in Mario and it told me no. Kudos to the runner for putting up with the inane BS of some schizo rando though.
"Guilty until prove innocent" should not exist. The accusing side needs to provide more solid evidence of foul play, not the accused side who needs to go out of their way to prove they are innocent.
This is why I prefer to use 'Innocent unless proven guilty', because 'until' in 'innocent until proven guilty' infers that guilt is there, it just needs to be proven.
@@sadscientisthououinkyouma1867 This kind of thing is prevalent in just about every online space due in no small part to the wide-spread use of social media as a primary social interaction outlet. Headlines get clicks, accusations are functionally the same as a verdict, and misinformation is heralded as gospel if your social circles demand it to be so. The whole of online society isn't interested in truth or evidence -- they want sacrifices.
@@OmegaSauruseasy mistake to make without knowing what’s being done. I’m more familiar with light so it’s like you have bucket of paint and want to know what percentage of the color yellow is (565-590 nm) in it. You could shine light on it, and pass the reflected light through a prism and then simply don’t measure the other colors. The removal of all other sounds to find the 50-60hz hum is similar to not measuring the other colors out of the prism
@@HexCopper The fact that brood CLEARLY didn't know what he was doing should be enough to be suspicious of the evidence this stuff should be handled by people who KNOW what they are doing so actual cheaters get caught and innocent people like blink aren't harassed by random people
That's the very definition of confirmation bias. He had a big pile of data he couldn't interpret, and he just stirred and poked at it until it became something that could be vaguely interpreted as evidence to his claims.
It's really wonderful to see a story of false accusations where the community (including the accused) comes together to make a show of proof that exonerates the accused. Excellent investigation and reporting. This is some class A effort on your part, Abbysoft. These kinds of investigative skills are professional grade.
He used... ChatGPT to get technical information. Oh, brother. In spite of the AI moniker, the glorified predictive text generator *is not actually intelligent.* And as you point out at the end, I can only imagine how Blink felt given it seems they were innocent and being accused of cheating when they didn't.
GPT 4o used to be smarter, and then it got replaced by a new model named chatgpt-4o-latest-11-20 on the 20th of November, which was a smaller and therefore dumber model. If he used gpt-4o-05-13 or o1-2024-12-17 it probably would've been better.
@@talkalexis no, alexis, you mistunderstand: NEITHER of them are smart. It did not "used to be smarter". NO AI is "smart" because they do not think. They don't know anything, they can't actually make connections between data points like humans can. The general public has VASTLY overestimated the "intelligence" of AI because of the shady way that AI companies have lied to consumers about their product. They shouldn't even be calling it "AI", in my opinion (I mean it technically is a type of AI, it's just not what people think of when they hear the word "artificial intelligence"). (it seems like you should probably know that if you know version #s, but the rest of your comment makes it sound like you don't understand that ChatGPT isn't "smart" in any real or measurable sense of the word.)
I was really surprised that the video didn't end with Brood getting banned. This is a huge thing to accuse someone of with no knowledge of how to actually do it, not to mention going against a mod's wishes to not post it publicly.
@@toumabyakuyaonly time I’ve found it useful was I was trying to navigate apples awful documentation I asked it how to get volume name from a disc drive; it made up an API that didn’t exist but it was in the correct library as the actual one you need so I looked up documentation for that and found the real one That said, if I was trying to do this a few years ago, when google actually worked, this would’ve found it on the first search “macOS get volume name of drive” or whatever
@@toumabyakuya It really doesn't. Requiring the use of Generative AI in any part of your "professional" process is all the proof anyone needs to know you are not actually a professional.
@@fumoffu_l Why? Isn't AI a tool like any other? Or are you arguing now that we should judge a construction worker based entirely on the brand of drill he uses? Are you saying that we should judge a teacher based on the brand of textbooks they use to teach their students? Because, if you aren't, then why should we judge somebody sorely because they used AI? Meanwhile the end product is of high quality, why should it matter?
So basically. A person with very little knowledge or experience in audio analysis, asks ChatGPT an AI chatbot for instructions on how to perform an audio analysis. And this person, with very little knowledge is somehow able to filter out 'bad' responses, and is somehow able to use the AI as a learning tool to perform an audio analysis. But instead of this being some fun, or a hobby on the side for fun. He uses the results of his 'audio analysis' to accuse another person of cheating, i.e antisocial behavior. Was he not tipped off, when he himself recognized that the AI gave 'bad' answers? Could it possibly mean that the chatbot was fundamentally incapable of teaching audio analysis to a person to a degree sufficient enough to be 'court admissible'. There is a very good reason why witnessing as an expert in a court has a HUGE barrier to entry.
"There is a very good reason why witnessing as an expert in a court has a HUGE barrier to entry." and honestly, still not as high as it should be in many places (like the U.S.). but that's a conversation for another time!
Yep pretty much. Honestly I think the experts kind of wasted their time studying the audio because all we really had to do was ask Brood about how he learned to analyze audio in the first place. The exact moment he said "I used Chat GPT" discredited him immediately and was a death sentence to his entire ""proof"" of cheating.
And how would he know which answers are "bad" when his knowledge of the subject is functionally zero, anyway? Were they "bad" responses because they didn't lead to the conclusion he already started with? He essentially used ChatGPT to give him the confirmation he was looking for. In other words, he was just demanding his confirmation bias be satisfied and that's why Brood still thinks the runs are cheated despite multiple experts saying they're not.
Conclusion: ChatGPT is NOT an expert and should not be relied upon as such. Brood really thought he had a gotcha there, really just seems like a case of dunning-kreuger
the thing is, it's not like ChatGPT was always that far off, but the dude doesn't even understand audio enough to know how to even prompt it? he's asking such general questions anyways, like ChatGPT doesn't have ANY of the info needed, such as what input devices are being used, is there a noise gate, are shielded cables being used, etc. If you gave this info to an expert without any other prompts, they would probably give similar responses (outside of using too much gain to detect the discrepancies, but there are several layers of issues with how this was asked before it got to that point) Point being, somebody well-versed in the topic could ask a more nuanced question to the AI and get an answer that might at least be a jumping off point to figure out the problem, but using ChatGPT alone without having even the most rudimentary knowledge of a subject is ridiculous.
and also he ignored what ChatGPT said anyways! his method of prompting got false positives and he just continued forward and ignored it instead of concluding that his methodology was flawed. ChatGPT literally gave him the answer he needed but he had a conclusion that he needed to follow.
@@SuperAmazingJared Thing is, at this stage chat bots are good at following instructions but that's all it can do really. Sometimes it seems that these bots are smarter than they should, but it's really only the result of having a good enough system prompt and/or knowledge base and/or can look up the information online. Which means if the bot is asked to do something that the bot haven't know how yet, then the bot can't do it correctly. The bot wouldn't become an expert until they were told how to become one. Also there's the concept of maximum token which is basically the "memory" of the bot. When a chat goes too long some of the chat histories just have to be ignored in order to make room for the more recent topic. And somehow most chat bot services just don't provide the information about how much memory capacity is already taken up and how exactly does this memory got cleaned up if it's full, and often times the bot can just malfunction without any warning, and a user without much knowledge might not be able to tell that the bot has already forgot how to do its job and can only pretend to still know things.
@@SuperAmazingJared I think so too. In my experience, ChatGPT says what you want to hear but not necessarily what is correct. When you try to "correct" GPT, it will just go along with it and rephrase the conclusion accordingly but it cannot understand context. People need to remember ChatGPT isn't a search engine, it is a text predictor and it's just outputting text not understanding what you or itself is even saying. And likewise I think Brood had already judged Blink of being guilty and was looking for evidence for that, and rather than assessing any evidence gathered and reevaluating his opinion, he just dismissed evidence that didn't go with his foregone conclusion or tried to reframe existing evidence to better suit his narrative.
It's not chat gpt really, it would be the same just google searching anwsers and looking at multiple sources. It's more when you're trying to learn from the ground up something new and have no knowledge to contest possibly wrong or info in wrong context is where you start to snowball a problem.
This feels like the original claim was not made in good faith. Like the intention was "I want to take this guy *down*" rather than "I want to keep the competition clean and fair."
@@SnackFoodCentral that is 50% true for the free version which brood used but you cant extrapolate that to chatgpt itself o3 was announced and got 25% on a maths benchmark that terrence tao (really good mathematician) could only do 1 question of
@@talkalexis IT. DOES. NOT. UNDERSTAND. ANYTHING. It might "look" like it is able to do stuff. In practice, it has pretty much all of the knowledge on the internet, way more than a human for sure, but it has 0 intelligence or wisdom. It literally predicts the next word that should follow. Like the options above your keyboard on Android that try to predict the next word, but a bit more sophisticated. It just does that over and over again. It can find patterns in words and phrases that humans might not pick up on, and extrapolate from those patterns, sometimes producing really convincing results that look intelligent. They might even be correct too, but not necessarily, because it isn't connected to any fact checker and is just a word prediction algorithm.
@Xnoob545 it is predicting the next word, thats true. its just so really really good at doing that, that it mimics actual understanding to the point that it can be "smarter" than the smartest mathematicians, which has already been achieved via that benchmark i mentioned
kinda pointless use case/selling point for o3 tbh, its still a bot of course its going to process math blazing fast. Its like usain bolt trying to outpace an f1 on a racetrack, never gonna happen
@@middyjohn thats not how llms work though consider how some llms fail at "is 9.11 or 9.9 bigger?" and how other state of the art llms only got 2% on those questions, llms arent calculators
the fucking double take i did at my phone at the chatgpt reveal. the trust people apparently have in a glorified chatbot with no true intelligence is astonishing. dude "talked" to chatgpt for pages and pages while its "memory" was going back probably a few hundred words at max and it never had a single clue about the context because its a glorified chatbot. fucking wow. glad the speedrunner was fully exonerated from such bs. what an unneeded mess.
Well, the memory of the bot is usually much more than that though, from a few thousand words to like 40K words. Definitely enough for solving a single question, and the issue is just that there's no indication if the memory has lapsed and a new chat need to be created to continue on the topic.
@FlameRat_YehLon today i learned then! i had futzed around with some llms in the past that had pretty limited memory retention, so i had assumed chatgpt was likely the same. thank you for the correction!
this whole thing seems so pointless. a guy deluded himself into thinking he noticed something, prompted chatgpt to give false positives and fuel confirmation bias, and then made a smug accusation despite having zero actual knowledge of what he was talking about. waste of everyone's time.
imagine involving multiple actual experts to go over stuff and then learn that you had to do it because of some moron who conjured some bullshit with chatgpt. what a waste of everyone's time.
Only to then go through multiple days, weeks, however long they took, doing the equivalent of genuine scientific analysis, to actually present findings, to have the amateur home-made armchair expert that caused this problem in the first place, basically say, "Yeah I'm gonna stick with the skills the internet and this predictive text generator have taught me, good luck losers!"
This is exactly how we ended up with anti vaxxers. Think about all of the time and money wasted fighting false information from these kinds of people. Think about how much real work could have been done with that time and money instead. Billions down the drain to appease these non contributors.
No, banning people for accusations just makes the community worse. Because, you are getting rid of those willing to put in the work to find cheaters. Preemptively you and anyone else are proven mindless sheep if you reply negatively.
@lanceknightmare The ban here isn't due to accusations, but rather due to a lack of proper procedure when evaluating the merit of said accusations, as well as unwillingness to recognize that the accusations have been disproven by people with actual qualifications to analyze the situation. If he just admitted he was in the wrong and apologized, it would be fine, imho. The fact he is pushing back after being proven wrong is the main problem.
Blink helped me big time when he gave me the strat for easy auto-encounters on the Mideel coastline for my PS1 FF7 pure 100% run. I saw his name and insta-clicked this vid thinking “ain’t no way in hell.” Blinks a legend, this Brood dude apparently, is not.
A disappointing case of ego. I was really surprised one of the key points was 'responding too soon'. Imagine playing a game so many times, it becomes second nature to be apprehensive about an area that can cause issues for your run. Sorry to hear Blink had to deal with something so poorly thought out and executed.
Yeah it was clearly a "oh i remember there's traps in here" and thenblink falls into a trap Done it so many times it's not nearly enough expecally since its a reaction
This is something I brought up early in the investigation. In addition, what stream viewers are seeing is delayed by the upscaling and capture chain but he’s seeing it without that on a CRT.
Like if the traps positions don't move it's more than possible for them to have moved in a direction only to remember the trap in the position a bit too late and react to the trap they knew was there from likely countless runs.
@@mogalixir Except as evidenced by this very video it can't even give you good tutorials... It's blind faith in a glorified autocomplete that was the real issue here. LLM's are garbage that cannot currently be trusted for tutorials, for information, for anything.
@@kezia8027the only thing you can really use ChatGPT for is just to have fun playing around with it and like nothing serious at all, and even then it’s limited becuase it’ll always go “as a large language model i cannot…” which is no fun rather than just bullshitting an answer which actually would be nicer in this case lol
I agree. The worst part is that these people seem to think Karl and Abyssoft only took off when they started talking about cheaters. They both have had successful YT careers long before Silly Bitchell became such a big thing. It's hard to believe they didn't think this was their fast track to fame. And that assumption is just wrong. Karl became huge after completing the Dam Agent speedrun untied WR, not by accusing people.
I asked him a *lot* of tricky questions and he patiently supplied answers, video, and more. I had to work very hard to control bias. As in, my own, as I immediately suspected he was innocent based on how open he was.
I've just finished the video and honestly i'm kinda pissed at the accuser and feel so bad for blink.... He's cooperated the entire time it seems without any sorta pushback too to clear his name, also one of the final messages looks like the accuser still thinks he's a cheater which is WILD
This isn't a social deduction game, it doesn't matter how someone acts. The only thing that matters is what an analysis of the gameplay itself shows or doesn't show. I'm not saying anything one way or the other
I think he's innocent too, but you'd be surprised: statistically speaking, it's actually GUILTY people who are more likely to go along with the investigation, despite the reality that it very well may get them caught. On the contrary, people who are innocent [and improperly accused] are more likely to get (correctly) angry and lash out, _refusing_ to cooperate. That's just not what happened this time, obviously. Human psychology is weird! :D
Where I know very little about what's going on and I mean next to nothing outside of this video. I agree with you because it's most likely the guilty person thinks they're never gonna get caught, so they're willing to go through with anything. They can even if it means that they will eventually get caught. Trust me, I've done it before in the past and I'm not happy with those moments when it comes to other things in life, not video games. Because if you're lying about video games.And cheating in such you have a sad life. If I'm going to the cheat in anything, I'm going to be open with it because I'm not part of a speed running community.I don't do that stuff.I video game for fun.So, if I'm cheating and somebody says that I am, I'll be like.Yes, you're right.I am and that's what I meant to do because i'm not competing for anything.@idontwantahandlethough
That "reacting before it happened" claim is complete BS; he said, softly "oh my god" because he was dreading what might be about to happen, I'm pretty confident I don't need to see greater context to be right about that.
That accusation earned a raised eyebrow the moment I heard it. It's a tile-based game, so you have the entire time it takes to move between tiles to regret hitting a button. As any gamer can tell you, frames feel a lot slower when you screw something up.
On top of that, I stream games frequently, and I often know when I've just hit a wrong button or executed the wrong action. Even if it hasn't happened yet, it's like I can feel it in my brain and it's too late to stop. Difficult to explain but sometimes you know you've messed up before before the results are seen.
Brood is a great example of the Dunning-Kruger effect in action. Learned a few things (poorly) about audio analysis and thought he was a sleuth, yet failed on something so simple as inrush current on startup.
Audio is its own monster. There is a reason audio engineers exist. This guy tries to accuse someone (wildly) over "splices," but for what purpose? He didn't understand enough about audio to make his conclusions, and his methodology got exposed to be faulty and idiotic. The cherry on top was using AI to try coercing it to match his "findings." XD
This actually irks me a lot. I hope this wasn't the case, but it almost feels like this person wanted to make the next big "CHEATER EXPOSED" video and tossed around wild accusations at a speedrunner *WHO HAD HIM AS A VIP*.
Funny that I just watched another video about people asking chatGPT to find bugs in public code projects like curl and then submitting those bugs for bug bounties. This is causing a deluge of bogus reports, either reporting on bugs that were fixed years ago or just inventing bugs that don't exist, and it's a huge waste of time for the project developers to respond to these. AI is great for some stuff but it's also a good way for charlatans to generate nonsense for their own benefit.
Brood should have been fully transparent about his lack of experience and his full sources, and not been so sure of himself that he had caught a cheater. Like it's not just that his knowledge was questionable, he came at Blink like "Bro, I gotcha, you're done now mate just wait".
Yeah. The issue was not with how much he taught himself but that he should have realized he was nowhere near knowing enough to accuse someone after what amounts to a dozen tutorials you find online.
ChatGPT for anything you need an expert for is insane. Like I maybe use it if I can't find what I need in a couple of pages on google for things like drop sources in an MMO. No way in hell am I every going to trust it for anything an actual expert exists for. Especially when I have no f-ing clue if it's hallucinating or not.
Imagine going through all this to accuse someone of splicing a >5 hour run because they got "consistently" lucky in one of the first splits of again, a >5 hour run. It's almost like the first chokepoint of a run that long is going to be by far the most attempts at it and therefore a lot of opportunities to get decent luck.
@@ztriker4406 Worst part about Dream is he had the skill as he was actually great at the game he did not need to cheat and he would have gotten the record eventually. It's really sad to see a genuinely skilled person fall to cheating, but it's more common than you think with games that are very RNG heavy especially if the devs updated it to be more RNG heavy
I'm loving the fact that the legit audio engineers in the speedrunning community can actually hear the walls and the equipment in the room and determined it to be legit just from that. It's like having Batman himself explain in a multipage technical report how he did that one thing in The Dark Knight.
the other cool thing about mains hum is that there are slight fluctuations in the grid's frequency (e.g. if it's 60hz it might be between something like 59.99 and 60.01), and that can also be used to somewhat accurately date a clip if the region's grid publishes mains frequency data. NESO in the UK does this at 1 second resolution
The moral of the story is: large language models are not a viable replacement for proper research and education. You cannot skip the hard work required to learn a skill. Stories like this need to spread so that we can shake the public perception that LLMs are magic answer machines with perfect knowledge.
The moment I heard "Taught himself" I knew that this "Brood" character had made up his mind already and no amount of proof otherwise was going to work. And then I find out he used _Chat GPT_ You could _not_ get more blatant in your reaching. I'm honestly disgusted (but not surprised) that he lied to himself so thoroughly . I hope he gets kicked out of that community, because it's clear to me he doesn't deserve to be _in_ it.
I'm sorry, asking GPT to tell you how to prove your point and then refusing to acknowledge that you're wrong is absolute tool behavior. The first half can be a whoopsie, but the second half? Ban him from the community. Absolute tool.
There is absolutely nothing wrong with making a claim about potential speedrunning cheating, but you better be very sure that you have a baseline of evidence before making it public. If you can't be sure. You should reach out to people individually and ask them for help.
I'm still new to the speedrunning scene, so please forgive me for ny ignorance, but I just wanna know... why? Why did Brood go out of his way to try and blame Blink for something as serious as this? Why didn't Brood be an adult and talk about this kinda stuff in private, instead of putting it on blast? Why was Brood using ChatGPT for serious claims? Maybe it's just me, but I'm confused as to what Brood's endgame was from all this
@@toumabyakuya Definitely a fan of Karl. Interestingly, Brood has the same problem as the people who cheat - he wanted to take a shortcut to get fame. In this case, he wanted to expose someone (anyone), but doesn't have the skill to do it or the patience to find a person who is actually cheating
Brood is a complete d-bag.... He was just out for blood for some personal reason it seems completely apparent with how hard he went after it, and how he released info even when being told not to.
This has been one of my favorite videos of yours... Super interesting going over the investigation yourself and seeing how you would do the investigation, as well as how it just goes to show that if you go into a study with the conclusion already in mind, you can find ways to find what you're expecting.
This is why accusations shouldn't be made by people who don't know what they're talking about. Armchair accusers hurt people's reputations and waste everybody's time. Brood should get a temporary sanction.
Phenomenal video. Hits you with the audio analysis lore dump then hits you with a haymaker of a twist in that it was ChatGPT trying to screw the speedrunner the entire time.
brood using chatgpt in the first place is laughable. but he goes a step further to ignore what it says to get confirmation bias out of it. and bonus points for STILL thinking blink cheated his run despite being proven wrong. this man isn't a clown, he's whole damn circus. what a joke. shame on the mods too for not punishing brood for this joke of a false accusation made with such weak and easily disproven evidence. bad actors like this needs to be and an example of.
Trusting chatgpt for anything is a bad idea but for complicated technical tasks? Pff. It can't even get basic historical dates and stuff correct for pubquiz questions
"I wonder how he could come to such a conclusion." The answer actually had me facepalming -.- Basing anything on ChatGPT, which is known to hallucinate 'facts' is just so ridiculous
20:20 "Mainly, how did Brood come to the conclusion that there was evidence of splicing?" Me: "Because he's an idiot." *roughly 40 seconds later* > ChatGPT ... oh, yup. Called it.
The way he prompts it too like "how to detect mains hum (or whatever it is called) in Adobe audition" he's trusting a predictive text bot to teach him stuff and he's not even asking the questions well, it's painful
-Moron trusts an "AI" to tell him how to find a hacker. -Causes 300+ hours of trouble for a 100% legit streamer. NEVER trust anything "AI". It's broken shartposting machines and nothing more.
The fake reaction complaint is odd to me. Speedrunners know when their inputs are wrong; they don't need to wait to see what happens to know their movement order wasn't right, so they could easily react just as or slightly before the results are shown on screen. Or maybe I'm missing something about the specific game?
to answer on this statement. you are 100% correct. i went the wrong way loading onto the stairs and knew since this was the second time in the same weekend i made the same mistake and ended up falling in that hole the previous run as well. i didn't know that you can avoid that hole (i wanted to take the shortest path to get back on track) since i normally never make that mistake, so i didnt know the backup
Poopoo on the accuser. I'd have some sympathy if, after all this, he let it go and apologize. Nope, just doubling down. He's got all the proof that Blink didn't cheat.
Copper tech here, that burst distortion is what happens when a current is passing through copper that has partially oxidised. Its very common in old electronics.
Its common, someone gets suspicious tries to understand how to actually find out and either gets bad or even wrong information and reaches a faulty conclusion. It was handled in a bad way and we all should be careful with throwing allegations around.
Agreed. He found ONE "proof" when stuff like this requires multiple "proofs", or at least something that reinforced his one "proof". Also if he had more patience and checked his stuff with actual experts(like abyssoft with multiple persons and experts) this wouldn't happen. I could see myself doing the first mistake, but in no way in hell all of the consectuve ones.
The people saying his innocence is proven by his transparency would be some of the same ones persecuting him if they saw the accusation video first. If the caustic treatment made the runner quit sharing before this MASSIVE effort went into proof, the same transparency is proof logic would say he was guilty when instead people had just broken his heart. Cheers those who did the hard work! Cheers to the runner for rising above another challenge!
Whenever I see people get accused, the mods have mountians of evidence and are very cautious about the accusation as they know it will possibly ruin someones speedrunning life. This guy just cared about catching someone and getting popular himself some steet cred.
Some people actually think ChatGPT is an all-knowing super search engine that can accurately explain any topic rather than a text predictor with a massive dataset and a penchant to people please. Those who believe the former rarely stop to consider that, maybe, if you could just ask the all-knowing genius box to solve a certain problem, someone else would have tried it already.
Except chatGPT never accused anyone of anything. No this person got a crash course from GPT and thought he was all of a sudden an expert. I don't believe it was GPT that was in his Livestream chat pretty sure that was blink just based off of initial.observation.
The ChatGPT reveal was positioned as the big twist that tore the entire narrative of the accuser down, so including it in the title would have required a restructuring of the video around another point (probably the reveal of innocence).
the fact people ACTUALLY use ChatGPT for something so serious tells me humanity is literally doomed. You're telling me you literally used chatGPT instead of learning yourself... and you exposed someone purely based on computer generated responses..... Bruh.
@@ZPu95cookbooks are written by people who actually cook, it's more like getting taught how to cook by someone who's never stepped foot in a kitchen, sure they might be able to guess how to cook an egg but at the end of the day they are guessing at every turn
I think you mean anyone can post recipes on the internet, to publish an actual cookbook you do need to know what you're doing (although someone did let Ned Fulmer publish one so it's not an inflatable system lol)
This is what Brooder had to say about the video on his Discord server: > they didn't address my main concern (interference) and they didn't even look at the originally recorded footage to check for zero padding or other forensic markers and he misrepresented the chatGPT thing So now not only is Blink still a cheater, but Abyssoft and everyone involved didn’t do a thorough enough investigation. He’s genuinely pathetic.
This definitely feels like one of those things where he's working backwards, trying to find proof of his accusations, rather than analyzing the evidence to determine whether or not he's cheating
Can Brooder get awarded some type of Dunning-Kruger award? He basically speedran being a clueless regard by using chatGPT, so it should be on a leaderboard somewhere.
Imagine just wanting to play games really fast, but you end up being the focus of a full-scale investigation just ‘cause some guy couldn’t believe that you were good at video games.
I was NOT expecting Chat GPT being dragged into this... I can't believe people actually use it to learn things, given how unreliable it is, and not just to cheese out college homework.
False accusations should be treated almost as severely as faked runs. The individual doing this should never be involved in analysis of any speedrun ever again.
Brooder cannot just accuse Blink without VERIFIEDS evidences. Immagine a mechanic (with no qualifications) telling you to replace the whole engine, just because he talked with an AI. Cheating must be discouraged in competitive playing, but trying to humiliate someone that till now has play fair, is awful. He should have worked on his allegations in private with a sound expert, and knowing Blink perificals devices. And after verify his finding with a mod, made the post public (especially not during a stream)
Maybe cause he provided some evidences about his theory, even if they weren't verified. At least now we know that someone without background experience shouldn't be listened.
@@Mario6732Because Brood didn’t back down and in fact doubled down even when given an opportunity to form a different opinion. Real damage was done here.
The main takeaway I got is that Blink is an absolute legend and Brood is a loser. I don't need ChatGPT to tell me that much Shoutouts to everyone that helped defend him, he was accused unfairly and I'm very glad the community came to his aid
As soon as it was discovered he learned via ChatGPT, the whole allegation should have been thrown out (unless the experts were already well on their way to completing it, or done; I'm tired, I mighta missed that bit). Based on what was brought up in this video, Blink handled things very professionally. Hats off to him for that, as well as for SMASHING that 5h window.
Go to expressvpn.com/ABYSSOFT and find out how you can get 3 months of ExpressVPN free!
nah
No thank you
THUMBS DOWN!!!
nah, it's actually a terrible vpn. Look into your sponsors.
your content is genuinely pretty good, which is a rarity on youtube, but sadly you have to stoop really low when it comes to money
Absolute gigachad. In the middle of a cheating investigation, gives everything asked for, continues to do runs with increased proof standards, crushes WR with a hand cam.
Yeah I almost cried when I found out he finally got his sub-5 among all of this absolute nonsense
It's like he got so annoyed at Brood that it doubled his focus lmao
@@codetakuSpite can make one do crazy things.
Honestly he should have just given nothing and got it anyway. The witch hunts are going way too far, and kneeling to them doesn't help anyone.
@@sadscientisthououinkyouma1867 giving nothing implies guilt, and there's no witch hunt, just one dude with chat gpt. Most others assumed innocence but wanted to do their due diligence.
@@AceWolf456 "giving nothing implies guilt" in what world? Prove to me you haven't murdered anybody, the format of which I demand uninterrupted stream of your entire lifespan such that we can verify for a fact you are not lying.
If you refuse to meet my simple demands you are implying guilt (for not submitting).
EDIT: also "most others assumed innocence but wanted to do their due diligence" is just a contradiction. You can't assume someone is innocent while also investigating them, because the suggestion they are innocent would mean they are not worth investigating. If he's innocent why investigate? If there's a chance he was not innocent and that justifies the investigation, then the assumption isn't that he is innocent but rather the assumption is he could potentially be guilty.
Blink was treated completely unfairly. I figured this video would end with Brood sheepishly admitting he was wrong and apologizing. But no, even after wasting everyone’s time with this nonsense, he still insists it’s cheated. Must have found a new ChatGPT model to learn from. Have some humility and give it up
I'm with ya. I can understand being convinced of something and then finding out you were wrong and apologizing. Not the end of the world! It happens!! But when you can't even admit you were wrong and say "yah looks like I messed up, sorry bros!", there's something seriously wrong with you. Just so, so, SO disrespectful of everyone's time and energy.
Yeah, this guy was a desperate, bad faith actor the entire time. Really wanted to be some cheat-busting legend, or really hated Blink for some reason, or both. Makes sense for someone who would use AI to try and learn something as in-depth as audio analysis. I hope he's banned from the community, but I doubt it.
Some people in chat can be motivated by peak hater energy. They see someone about to PB or get WR, and decide their time to troll has come.
@@SmokeRulz I hope he ends up on Liveleak. there's way too many people online that needs to be in a liveleak compilation...
@@idontwantahandlethough for that alone. this mf should be banned from accusing people.
they aren't doing it for the good of the sport, but for clout. and unhonorably
Imagine actually using chatGPT to learn something you have 0 experience in and then use this to accuse someone of cheating, absolute Dunning-Kruger
Its been a problem in my university. It can be really useful when you already experienced in the field and can shift through the bullshit it gives you, but the amount of people submitting work entirerly based on it is way too high
I've gotten into political arguments with people online (poor choice, I know lol) and they used chatGPT as their "rebuttal". It was pretty clear they had no idea what they were talking about before that, but it _certainly_ was obvious afterwards!
People are wild. What's even the point of life if you're not going to bother learning anything?! Learning new things is like the best part of life!!
once again the dangers of AI are apparent. In here, the stakes were relatively low (one person stood to lose their credibility in a niche community), but imagine if this perception that AI is a reliable source of information is used in law (we've seen this happen, actually - chatGPT simply invented a court case), or medicine, or other high-stakes environments where wrong information can lead to someone losing their freedom, or even their life.
AI is NOT reliable. Stop using it.
@@NameHere2243 exactly, as long as you already kinda know the answer you expect from the LLM and are ready to "shift through the bullshit" (I really like that wording), it can be ok. But LLMs are very convincing in the false informations they output and will mislead users that lack basic knowledge on the topics they are asking them about. It is sadly also a problem in my university, with students failing to understand they LLMs are not magic, and just not questionning anything chatGPT tells them :/
Then imagine using your chatgpt expertise to tell people that actually do this for a living that they wrong and you just presented your argument wrong.
Dude never had an argument
Deciding you are right and then looking for supporting evidence is NOT how this is done, especially not by using a technology that is well known for its ability to be painfully wrong, when you have a hypothesis, the first thing you should be doing is trying to prove it wrong.
That's a great way of putting it. Things like this are WAY too common online because it's always conclusions first, evidence later.
Wow. Imagine seeing you here oh great Whisky hoarder.
And the fact if you aren't proficient in such usage of technology, your results can be off.
Chatgpt wasnt even wrong. Chatgpt gave the accurate conclusion that brood is wrong. Then, brood demands chatgpt to support him.
Guilty until an LLM proves you innocent, then ignore the LLM and proceed to accuse the runner publicly anyway.
Have every expert in the community prove you wrong, stand by your false accusation.
????
Profit
Maybe a hot take, but if I was a mod of this community, based on what's presented here, Brood would be getting a ban for at least a year for this BS. Tormenting a community member, wasting multiple people's time trying to research his allegations, blatantly ignoring mod's instructions to not publicly post the allegations before the mods do; all the while not being able to confidently stand behind what he's accusing without a chatbot there holding his hand. Absolutely baffling behavior.
Completely agree. If the allegation had any basis in reality, an apology would maybe be enough, but given how much time of everyone’s he’s wasted with his stupidity and the fact that he’s still not admitting he was wrong, there needs to be some serious reprimand
i’d give him lifetime ban, this is completely idiotic and wasted so many hours of innocent people’s time
A bit of a hot take. Assuming lessons were learned and no malice, I think they should just move past it. Witch-hunting the guy for revenge isn't going to make anyone but uninvolved parties feel better by satisfying their bloodlust. Maybe Brood could make a public apology, though
@@displayer6023 assuming they were learned. Ig we'll see if he finally does change his mind now
@@displayer6023 I think giving them a time-out from the community might be for the best for both sides. It gives Brood some time off to think about things and also makes sure they aren't around for any potential witch-hunts. By the time they get back it would all have calmed down.
Like, even for Blink themselves, imagine how frustrating it is to be accused of something you didn't do, and have to go through a full on investigation where doubts are cast upon you, and there's even supposed evidence, even though you know you never cheated. Honestly I wouldn't be surprised if this killed a bit of their passion for speedrunning this game. This sort of stuff doesn't go away with just "oopsie i guess i was wrong, anyways"
The fact that Brood published their findings before mods could have a proper look is probably the worst part. This could all be avoided if they just listened to the people who actually know what they're talking about. I don't want to speculate why they did that, maybe just stupidity (Hanlon's Razor and all that) but the fact that this was very avoidable is still there.
The biggest sin of all this isn't the accusation, nor the ChatGPT usage... it's the attempt at public humiliation by going into his stream and calling him out.
In my opinion there’s nothing wrong with using chatGPT to become an amateur at something but please realize - you’re an amateur as much as someone who just did random tutorials.
@@mogalixir Less so, actually. Watching tutorials actually makes you use what's called a "brain". Trusting a predictive text generator ... doesn't.
Yeah that part really bugged me too. There’s nothing wrong with holding a run up for closer inspection, that keeps things fair for everyone. But the public call out with poor information before his “proof” was even out was really crappy.
I'd still say it was the ChatGPT usage, but yeah, what a damn troglodyte this guy was. He was so hungry for that big GOTCHA that he made himself look like a tool even before the GPT reveal.
the biggest sin is the Dunning Krueger delusion underlying his confidence to accuse in the first place
Being accused of something when innocent is a huge deal, mentally and emotionally. I'm glad Blink did well after it all.
yeah it's a really awful feeling :/
they say IRL, the punishment is the process.
If you got accused of murder and had to go through a whole trial, just because you were proved innocent doesnt mean its all ok
And even when someone's proven innocent, it doesn't always mean things will go back to how they were. Court of public opinion means some people will think "Well you were accused, so you must have done SOMETHING sketchy", or people will spread misinformation about how you were proven guilty, and people who know absolutely nothing about the situation will either be driven away, or jump on the mob bandwagon blindly.
I got accused of stream sniping after dropping in during a match twice; once at start to say glhf and again later because a player gave up during match and I wanted to see if I could tell they were a group.
I replied to something (at this point just standing around cause I was going to let them all go) and suddenly streamer and chat started accusing me of sniping even though I had done nothing but message twice and my gameplay showed no hints of ss. It really hurt my feelings so I just decided to end the game instead of letting them go; I had left the stream again by that point. But because of how Twitch handles the Viewer list (not updating for sometimes 4 whole minutes even on refresh) they accused me of still being there. Even went on about me needing to "mute stream" to not hear callouts when 1; delay is so bad any info on your location in game is void, and 2; they had one callout that entire match and again delay.
I didn't find out till later watching vod they had called me a clown and a whiner and kept accusing me of sniping, how disappointing it is when it's one of their own (vtuber), and banned me. It really really hurt and ate me up inside for WEEKS. I even had my own VOD as proof I didn't snipe since it was my first month streaming. I didn't even want them to admit they were wrong publicly I just wanted an apology for assuming and being so rude.
It took forever to convince myself not to let it eat me up so much and that it wasn't worth all my energy. The accusations stick for a LONG time when you know you're innocent and it hurts because people won't just listen or take 2 seconds to look even just a minute longer at the proof.
@@Someguyhere111 Don't turn this one situation into this big thing lol. I'm pretty sure that once people find out that Chat EBT was used, they wouldn't take them seriously
This is an obvious attention seeker who publicly accuses someone of cheating on baseless grounds using arguments that he was absolutely unqualified to make. What's more offensive is the amount of effort that everyone had to take in order to disprove this nonsense, not to mention the mental toll it takes on you being accused. Disgusting behaviour.
"a lie travels around the world twice before the truth has time to take a dump and find his keys" (or something like that)
Now take what you just said and apply it to medicine and vaccines. So much time and money has to be waisted proving people who should never have had a platform in the first place wrong.
He was ignorant and over-confident, but I see no reason to think he didn't believe what he said, or that he wasn't trying to do good.
Prime example of the Bullshit Asymmetry
@@fumoffu_l >waisted
Bro wanted to become the next legend who busted a fake speedrunner.
Temu Karl Jobst
"Hello, you absolute fairytale"
@@infinidominion lol
He became a legend all right, just not in the way he thought.
He was literally brooding.
26:58
>"If there is spliced footage, I think you should be permanently banned. If I somehow made an error and you are exonerated, you should be able to do whatever you like"
huh???? And the guy who falsely accused a man using techniques he had no idea how to use and relied on ChatGPT for stays????? Please tell me that the guy got banned for falsely accusing him with flimsy and garbage evidence.
Oh my GOD I didn't even catch that part. He sounds so self-righteous. I would have left the community out of shame after that... but clearly he's built of a different type of IQ points...
This alone should be reason to permanently ban Brood from the community. He went all-in with his demands, turnaround is only fair.
@ArnieMcStranglehold "Personally I hope I made an error. I just don't think I did."
What an asshat
I really wish this was pointed out in the video because it really shifts how Brood was approaching a possible cheating run. He doesn't even seem to think it was possible he was wrong in that same paragraph. That and the fact he doesn't seem to accept that the experts don't think there is any evidence of cheating and instead says he's getting better at his technique. I wouldn't be surprised if he accuses another innocent person. Definitely should get banned to some extent to prevent further occurrences, especially since he didn't listen to requests to not make the accusation public before the mods got a chance to look over the evidence.
Worst part is that after 3 audio engineers reported there was no splice he still believed there was
Jaw dropped at the chatGPT reveal. Literally trying to use a shortcut to expose someone else for supposedly using shortcuts, the irony is astounding.
I legit shouted "are you fucking kidding me?!" at my screen at the reveal. JFC causing so much drama because you are stupid enough to trust AI.
It definitely gives it a "This is a witchhunt" feel.
MamaMax has left the chat (if you know then you know 😂)
Using AI is a banworthy attempt and when the reveal happened I laughed so hard.
I actually found what we saw of his logs with ChatGPT fascinating. Perhaps I'm reading too much into it, but it felt like a lot of his "questions" to ChatGPT were pointed towards finding the results he wanted, and not what was actually there.
He had already decided Blink was guilty, and wanted to know how to make the "evidence" say that, instead of checking if the evidence actually showed he was guilty.
I was shocked that Brood wasn't banned at the end of the video. It's a big deal to accuse someone of something without knowing how to do it, let alone defying a moderator's request that it not be posted publicly.
I support a fact checkers existence to keep runs legitimate. It's good brood is around. He wasn't hostile and was professional about his approach. As someone with that investigative mind myself I want to be wrong so tell me why I am. If I am its a lesson for me that shouldn't lead to a reputation altering incident.
Innocent before being proven guilty.
@@nightwintertooth9502 Fact checkers and validators are important, but I have to ask: are you high?
Brood was SO far from professional, and made a mockery of the people who ACTUALLY do stuff like this by thinking he could supplant years of education and knowledge by using ChatGPT. He needs to be laughed out of the community, not extoled for being a "professional", because he wasn't.
Maybe you just randomly clicked around the video, but the ACTUAL professionals are the kind of fact checkers we need, not a con artist like Brood who doesn't give a shit about the community and just wants to try and get internet famous for catching a cheater. There were plenty of ways to do this right, and Brood did none of them.
Begone, shill.
@@nightwintertooth9502 professional about his approach?????????? dude, his only source was sloppy chatgpt self-learning, wtf are you talking about???
@@nightwintertooth9502blink himself has talked about how extremely stressful it was to be falsely accused, even with flimsy evidence. Brood WAY jumped the gun and ignored mod directions to let them review his evidence privately. He could have had his evidence reviewed and achieved the same result without wasting a LOT of peoples' time and putting blink through all this.
@@nightwintertooth9502 You are talking absolute shit - Brood had no idea what he was talking about. Just trying to be a smartass and create drama.
Brood really had the balls to accuse someone live on stream after learning how to analyze audio from ChatGPT. Dude should be ashamed.
What is the point of making a comment like this?
@@PapaBlue1 social validation for hating ai
Not the brightest move a man can do.
@@PapaBlue1 What is the point of making a comment like that?
@xenn4985 oh no, not my heccing AI!!! It must be defended at all costs!!!!
For me the biggest indication that he did not cheat is the absolute cooperation and participation in the investigation. Anytime someone cheats and then people ask them questions about their setup or for help analyzing their gameplay, it's always "oh my setup is in another state, my brother took the console, so now I don't have it, and I'm using a different cart and cables and..." etc. Instead the dude described every part of it in detail and even loaded analysis tools onto a flashcart and provided every sample that was asked for. A cheater would never. He's as legit as it gets.
Honestly I can understand if a regular person does not feel the need to buy a flashcart so they can load it with analysis tools in the event that they are accused of cheating as it's not really a scenario a normal person would think of only a paranoid person. Like for some systems such as a GameCube you have to physically mod it in order for a flashcart to even work and most people don't have the know how or even technical skills to do that as improper modding can damage your console so a lot of people would even be hesitant to even attempt it
blink's cooperation with the investigation is a huge indication of his innocence imo. no cheater would record and share so much device information if they were actually splicing. can't imagine how it felt to be called out so publicly like this and then learn his accuser used chatgpt as his main source of learning about audio engineering, that's so wild
TWISTED WONDERLAND FAN SPOTTED???
I agree, but only a little bit. An actual cheater might still be very capable at gaslighting people and blowing smoke up their asses by leveraging their popularity. My first thought immediately goes out to Dream, who (other than getting defensive) still offered various files and other bits of information to the people investigating his cheating. Same goes for actual criminals being interrogated by law enforcement: often the right to remain silent is skipped because they are super-eager to prove their own innocence in all the possible ways.
Just because someone is helpful doesn't mean they aren't guilty.
@@Aviertjei dont think 'helpful' is the right word to describe a liar's attempt to weasel out of trouble, but i agree with your point.
I think the word 'cooperative' makes more sense here.
@@Aviertje i get your point but even dream had his uncooperative moments, iirc he said something like "I can't give you my mods folder because we routinely delete everything in it for our minecraft videos" which *could* be possible but it just sounds like something he made up so he didn't have to show his mods folder. And in the end, turned out it was because he was indeed using a mod. Sooner or later you'll hit a sore spot which they can't give you without revealing themselves.
Meanwhile with blink, there was not a single point in the video where he refused to provide something. He recorded like three baselines, gave device information, ran several tests... Idk it just feels too open for him to be suspicious. Though of course the real proof is in the data gathered, not his character.
LLMs can be useful if they point you to a resource. They have an enormous knowledge base and can find things for which you don't know the right words for.
However, relying SOLELY on an LLM is very stupid, as LLMs are still nowhere near as good as humans in reasoning ability (they have a heavy bias toward what has the most text devoted to it as being the most important). LLMs are a great search tool, and can help find particular passages in long documents that talk about a concept - without being hidebound to specific words like a word-search is. But LLMs can't outthink humans, and in certain ways won't ever due to their focus being skewed through a language lens.
Brood should be punished for not only terrorising another speedrunner, but for also breaking the rules when it comes to publicizing his findings.
It is ok to be skeptical, but not when you cause so much distress only to be wrong.
In honor of this man trying to bust a speedrunner with chatgpt, im going to prove abyssoft fakes all of his videos using reverse psychoanalysis that my grandfather learned selling nfts to a paralegal
I laughed hard with this one 😹
Hope he sold em out of an ice truck.
paralegal or paraplegic?
Yes
In defense of Abyssoft, I'm going to rebute your arguments using information that was revealed to me in a dream
something EVERYONE is missing.
When confronted on the issue of getting good knight luck, he says he only watched 2 videos.
All you had to do, is look at the last 10-20 knights on completed runs, and see how frequently they are lucky. All this sound engineer stuff is assuming their was inordinate luck in the first place.
Ofc there will be a correlation, a good early is more likely to be completed, but you would expect a splicer to cheat their best runs, otherwise they wouldnt need to cheat.
Zeals logic was...sound here.
Sounds like brood never learned the idea of survivorship bias
@@Exodyr This exactly. Of course an early game boss will have better luck than others in the run, you're more likely to reset if the early game goes poorly, since you've committed less time to this run
There’s a reason why, during the Dream cheating debacle, his accusers didn’t just take the drop rates from one or two runs with suspicious results.
They collected data from thousands of runs, even runs which ended prematurely, to prove that there was no conceivable way that random odds would produce such results.
this was highkey pissing me off before I even got to the chatGPT thing ngl
@@jellywillreturn oh it was by no means thounsands of runs :) it was all the runs on a few streams, but mpre than enough to mathematically proof it. Stand up maths has a great video on the dream cheating scandal and its super interesting
I’m not even shocked at the use of ChatGPT. My students turn in AI slop all the time and usually can’t answer basic follow-up questions, because they are using it as a replacement for knowledge rather than as a tool for enhancing it.
Shame on you Brood.
And before that they would just copy/paste wikipedia. I remember one presentation used the phrase "specific gravity" when talking about a battery. The teacher asked them what specific gravity is, and they had no answer.
@@shawn576 just follow the info in the hyperlink professor.
Claude is the better AI tool anyway, GPT is utter slop
You need to start reporting them for intellectual dishonesty.
Reminds me of the inception horn I heard when Math classes finally started to say I had to show my work...
Bro needs consequences for this. Wasting a tremendous amount of time of far more qualified people, and undoubtedly caused unnecessary stress on someone.
He would just make a new account though these assholes always do this
Prison at minimum, Deny Defend Depose at maximum.
imo, asking a speedrunner in a long speedrun thats trying to shave a couple seconds off to break a milestone to "open a menu" is a dick move. Hes a few seconds shy of sub-5hour and you want him to waste a few seconds.
I get that cheaters are pretty common in speedrunning...for some reason...but man i'd be annoyed in Blink's scenario at that request.
Yep. Dude was playing moderator, and clearly doesn't know how multiple things work. xD
> ... for some reason...
In the words of the (absolute legend) Karl Jobst, "Runners don't cheat to get a faster time. Runners cheat to get a time, faster."
Cheaters are common in anything that is competitive, from 'roids to riding buses during marathons to aimhacks on-stage during fps tourneys, splicing runs and having 3 people submit runs under 1 name, if it's a competition people will find a way to try and get an illegitimate edge.
Definitely a dick move. "Oh, you're in the middle of a highly complex, high stress performance? Stop and pay attention to me instead, or else you're cheating!"
I bet cheating isn't that common after all tho. I see a lot of speedrunning but not a lot of cheating. Maybe it's kept hushed up, or? ....
Bro deadass screenshotted waveforms and asked a chatbot if it looks bad. Buddy, your chatbot cant even beat a captcha about bikes.
Hey now, some of those bikes can be really hard to figure out!
"Hey chatgpt, tell me if this person is cheating and also give me insanely bad information about audio editing"
Calling this a "scandal" is generous. This is like someone going into a 50 page document about how a Mario speedrun is cheated because he totally touched the piranha plant and I asked chatgpt if touching piranha plants was allowed in Mario and it told me no.
Kudos to the runner for putting up with the inane BS of some schizo rando though.
Dude wasn't even using chatGPT correctly, and he thinks he's the next big audio engineer busting speed runners cheating.
I hope to refresh the page and have people realize that this is indeed THE Ken Ashcorp
@@Godseyyy Same bud. I love Ken's songs!
I did not expect to see Ken Ashcorp weighing in on a speedrun cheating scandal lol
@@Godseyyy wait what, you're right 😭
"Guilty until prove innocent" should not exist. The accusing side needs to provide more solid evidence of foul play, not the accused side who needs to go out of their way to prove they are innocent.
A lot of people have prejudice and bias, sadly.
This is why I prefer to use 'Innocent unless proven guilty', because 'until' in 'innocent until proven guilty' infers that guilt is there, it just needs to be proven.
That's just the speedrunning "community" for you, though such a low-trust "community" can hardly be called as such.
@@sadscientisthououinkyouma1867 This kind of thing is prevalent in just about every online space due in no small part to the wide-spread use of social media as a primary social interaction outlet. Headlines get clicks, accusations are functionally the same as a verdict, and misinformation is heralded as gospel if your social circles demand it to be so.
The whole of online society isn't interested in truth or evidence -- they want sacrifices.
I don't know man... He edited the audio so much when doing his research it feels like compromised the evidence to get the match he was looking for
Welp, I made the comment before the experts were introduced, I guess I wasn't far off.
Also: CHAT GPT? COME ON 💀💀💀
@@OmegaSauruseasy mistake to make without knowing what’s being done. I’m more familiar with light so it’s like you have bucket of paint and want to know what percentage of the color yellow is (565-590 nm) in it. You could shine light on it, and pass the reflected light through a prism and then simply don’t measure the other colors. The removal of all other sounds to find the 50-60hz hum is similar to not measuring the other colors out of the prism
I was thinking the same thing when I saw his audio stuff, just by looking at it I could already tell that nobody could tell what's going on.
@@HexCopper
The fact that brood CLEARLY didn't know what he was doing should be enough to be suspicious of the evidence this stuff should be handled by people who KNOW what they are doing so actual cheaters get caught and innocent people like blink aren't harassed by random people
That's the very definition of confirmation bias. He had a big pile of data he couldn't interpret, and he just stirred and poked at it until it became something that could be vaguely interpreted as evidence to his claims.
It's really wonderful to see a story of false accusations where the community (including the accused) comes together to make a show of proof that exonerates the accused. Excellent investigation and reporting. This is some class A effort on your part, Abbysoft. These kinds of investigative skills are professional grade.
He used... ChatGPT to get technical information.
Oh, brother.
In spite of the AI moniker, the glorified predictive text generator *is not actually intelligent.*
And as you point out at the end, I can only imagine how Blink felt given it seems they were innocent and being accused of cheating when they didn't.
GPT 4o used to be smarter, and then it got replaced by a new model named chatgpt-4o-latest-11-20 on the 20th of November, which was a smaller and therefore dumber model. If he used gpt-4o-05-13 or o1-2024-12-17 it probably would've been better.
TLDR: It might've worked with o1 or o1 pro.
@@talkalexis I strongly doubt it would have been that much better. AI Hallucination is an ever-present possibility.
@@talkalexis was that the version that spontaneously switched languages? or the one that couldnt do math? lmao
@@talkalexis no, alexis, you mistunderstand: NEITHER of them are smart. It did not "used to be smarter". NO AI is "smart" because they do not think. They don't know anything, they can't actually make connections between data points like humans can. The general public has VASTLY overestimated the "intelligence" of AI because of the shady way that AI companies have lied to consumers about their product. They shouldn't even be calling it "AI", in my opinion (I mean it technically is a type of AI, it's just not what people think of when they hear the word "artificial intelligence").
(it seems like you should probably know that if you know version #s, but the rest of your comment makes it sound like you don't understand that ChatGPT isn't "smart" in any real or measurable sense of the word.)
I was really surprised that the video didn't end with Brood getting banned. This is a huge thing to accuse someone of with no knowledge of how to actually do it, not to mention going against a mod's wishes to not post it publicly.
People should be required to disclose if they've used AI for something so we can throw out their opinions sooner.
I mean, it depends both on what they used AI for and if they checked what the AI gave them to make sure it was all corret.
Or better yet, if they really know what they're doing.
@@toumabyakuyaonly time I’ve found it useful was I was trying to navigate apples awful documentation I asked it how to get volume name from a disc drive; it made up an API that didn’t exist but it was in the correct library as the actual one you need so I looked up documentation for that and found the real one
That said, if I was trying to do this a few years ago, when google actually worked, this would’ve found it on the first search “macOS get volume name of drive” or whatever
@@toumabyakuya It really doesn't. Requiring the use of Generative AI in any part of your "professional" process is all the proof anyone needs to know you are not actually a professional.
@@fumoffu_l Why? Isn't AI a tool like any other? Or are you arguing now that we should judge a construction worker based entirely on the brand of drill he uses? Are you saying that we should judge a teacher based on the brand of textbooks they use to teach their students? Because, if you aren't, then why should we judge somebody sorely because they used AI? Meanwhile the end product is of high quality, why should it matter?
So basically.
A person with very little knowledge or experience in audio analysis, asks ChatGPT an AI chatbot for instructions on how to perform an audio analysis. And this person, with very little knowledge is somehow able to filter out 'bad' responses, and is somehow able to use the AI as a learning tool to perform an audio analysis. But instead of this being some fun, or a hobby on the side for fun. He uses the results of his 'audio analysis' to accuse another person of cheating, i.e antisocial behavior.
Was he not tipped off, when he himself recognized that the AI gave 'bad' answers? Could it possibly mean that the chatbot was fundamentally incapable of teaching audio analysis to a person to a degree sufficient enough to be 'court admissible'. There is a very good reason why witnessing as an expert in a court has a HUGE barrier to entry.
"There is a very good reason why witnessing as an expert in a court has a HUGE barrier to entry."
and honestly, still not as high as it should be in many places (like the U.S.). but that's a conversation for another time!
amateur armchair audio analysis
Yep pretty much. Honestly I think the experts kind of wasted their time studying the audio because all we really had to do was ask Brood about how he learned to analyze audio in the first place. The exact moment he said "I used Chat GPT" discredited him immediately and was a death sentence to his entire ""proof"" of cheating.
And how would he know which answers are "bad" when his knowledge of the subject is functionally zero, anyway? Were they "bad" responses because they didn't lead to the conclusion he already started with?
He essentially used ChatGPT to give him the confirmation he was looking for. In other words, he was just demanding his confirmation bias be satisfied and that's why Brood still thinks the runs are cheated despite multiple experts saying they're not.
Conclusion: ChatGPT is NOT an expert and should not be relied upon as such.
Brood really thought he had a gotcha there, really just seems like a case of dunning-kreuger
the thing is, it's not like ChatGPT was always that far off, but the dude doesn't even understand audio enough to know how to even prompt it?
he's asking such general questions anyways, like ChatGPT doesn't have ANY of the info needed, such as what input devices are being used, is there a noise gate, are shielded cables being used, etc. If you gave this info to an expert without any other prompts, they would probably give similar responses (outside of using too much gain to detect the discrepancies, but there are several layers of issues with how this was asked before it got to that point)
Point being, somebody well-versed in the topic could ask a more nuanced question to the AI and get an answer that might at least be a jumping off point to figure out the problem, but using ChatGPT alone without having even the most rudimentary knowledge of a subject is ridiculous.
and also he ignored what ChatGPT said anyways! his method of prompting got false positives and he just continued forward and ignored it instead of concluding that his methodology was flawed. ChatGPT literally gave him the answer he needed but he had a conclusion that he needed to follow.
@@SuperAmazingJared Thing is, at this stage chat bots are good at following instructions but that's all it can do really. Sometimes it seems that these bots are smarter than they should, but it's really only the result of having a good enough system prompt and/or knowledge base and/or can look up the information online. Which means if the bot is asked to do something that the bot haven't know how yet, then the bot can't do it correctly. The bot wouldn't become an expert until they were told how to become one.
Also there's the concept of maximum token which is basically the "memory" of the bot. When a chat goes too long some of the chat histories just have to be ignored in order to make room for the more recent topic. And somehow most chat bot services just don't provide the information about how much memory capacity is already taken up and how exactly does this memory got cleaned up if it's full, and often times the bot can just malfunction without any warning, and a user without much knowledge might not be able to tell that the bot has already forgot how to do its job and can only pretend to still know things.
@@SuperAmazingJared I think so too. In my experience, ChatGPT says what you want to hear but not necessarily what is correct. When you try to "correct" GPT, it will just go along with it and rephrase the conclusion accordingly but it cannot understand context. People need to remember ChatGPT isn't a search engine, it is a text predictor and it's just outputting text not understanding what you or itself is even saying.
And likewise I think Brood had already judged Blink of being guilty and was looking for evidence for that, and rather than assessing any evidence gathered and reevaluating his opinion, he just dismissed evidence that didn't go with his foregone conclusion or tried to reframe existing evidence to better suit his narrative.
It's not chat gpt really, it would be the same just google searching anwsers and looking at multiple sources. It's more when you're trying to learn from the ground up something new and have no knowledge to contest possibly wrong or info in wrong context is where you start to snowball a problem.
This feels like the original claim was not made in good faith. Like the intention was "I want to take this guy *down*" rather than "I want to keep the competition clean and fair."
That's how a lot of speedrunner stuff is, it's about drama with these people, not enjoying a game.
this guy plainly misunderstands how chatgpt works. it doesn't "understand" anything. period.
@@SnackFoodCentral that is 50% true for the free version which brood used but you cant extrapolate that to chatgpt itself
o3 was announced and got 25% on a maths benchmark that terrence tao (really good mathematician) could only do 1 question of
@@talkalexis IT. DOES. NOT. UNDERSTAND. ANYTHING.
It might "look" like it is able to do stuff.
In practice, it has pretty much all of the knowledge on the internet, way more than a human for sure, but it has 0 intelligence or wisdom.
It literally predicts the next word that should follow.
Like the options above your keyboard on Android that try to predict the next word, but a bit more sophisticated.
It just does that over and over again.
It can find patterns in words and phrases that humans might not pick up on, and extrapolate from those patterns, sometimes producing really convincing results that look intelligent. They might even be correct too, but not necessarily, because it isn't connected to any fact checker and is just a word prediction algorithm.
@Xnoob545 it is predicting the next word, thats true. its just so really really good at doing that, that it mimics actual understanding to the point that it can be "smarter" than the smartest mathematicians, which has already been achieved via that benchmark i mentioned
kinda pointless use case/selling point for o3 tbh, its still a bot of course its going to process math blazing fast. Its like usain bolt trying to outpace an f1 on a racetrack, never gonna happen
@@middyjohn thats not how llms work though
consider how some llms fail at "is 9.11 or 9.9 bigger?" and how other state of the art llms only got 2% on those questions, llms arent calculators
the fucking double take i did at my phone at the chatgpt reveal. the trust people apparently have in a glorified chatbot with no true intelligence is astonishing. dude "talked" to chatgpt for pages and pages while its "memory" was going back probably a few hundred words at max and it never had a single clue about the context because its a glorified chatbot. fucking wow. glad the speedrunner was fully exonerated from such bs. what an unneeded mess.
Well, the memory of the bot is usually much more than that though, from a few thousand words to like 40K words. Definitely enough for solving a single question, and the issue is just that there's no indication if the memory has lapsed and a new chat need to be created to continue on the topic.
@FlameRat_YehLon today i learned then! i had futzed around with some llms in the past that had pretty limited memory retention, so i had assumed chatgpt was likely the same. thank you for the correction!
The accuser used ChatGPT and fed it data to analyse? Accusation should be automatically disregarded
Just that he used a LLM for analysing visual data should be a red flag.
Attacking Blink in the stream when told NOT to, should have had him instantly blocked by the mods.
this whole thing seems so pointless. a guy deluded himself into thinking he noticed something, prompted chatgpt to give false positives and fuel confirmation bias, and then made a smug accusation despite having zero actual knowledge of what he was talking about. waste of everyone's time.
imagine involving multiple actual experts to go over stuff and then learn that you had to do it because of some moron who conjured some bullshit with chatgpt. what a waste of everyone's time.
and then at the end of it all, that moron STILL BELIEVES CHATGPT OVER THE EXPERTS
Only to then go through multiple days, weeks, however long they took, doing the equivalent of genuine scientific analysis, to actually present findings, to have the amateur home-made armchair expert that caused this problem in the first place, basically say, "Yeah I'm gonna stick with the skills the internet and this predictive text generator have taught me, good luck losers!"
Just to say it, we had no idea he was relying on ChatGPT until much later
@dwangoAC By the way, really appreciate the work you guys do.
This is exactly how we ended up with anti vaxxers. Think about all of the time and money wasted fighting false information from these kinds of people. Think about how much real work could have been done with that time and money instead. Billions down the drain to appease these non contributors.
So glad to see Blink ended up crushing the sub 5 after all that bs
Happy ending. Brood fails to stop Blink getting WR.
Brood should be banned from accusing anyone of cheating.
Brood should be banned, period. From literally everything.
No, banning people for accusations just makes the community worse. Because, you are getting rid of those willing to put in the work to find cheaters. Preemptively you and anyone else are proven mindless sheep if you reply negatively.
@@lanceknightmare He's shown he's willing to lie and coax data if it means he gets the clout for getting someone banned.
@lanceknightmare The ban here isn't due to accusations, but rather due to a lack of proper procedure when evaluating the merit of said accusations, as well as unwillingness to recognize that the accusations have been disproven by people with actual qualifications to analyze the situation.
If he just admitted he was in the wrong and apologized, it would be fine, imho. The fact he is pushing back after being proven wrong is the main problem.
@@marcosmauricio7805 that too.
Blink helped me big time when he gave me the strat for easy auto-encounters on the Mideel coastline for my PS1 FF7 pure 100% run.
I saw his name and insta-clicked this vid thinking “ain’t no way in hell.”
Blinks a legend, this Brood dude apparently, is not.
It is crazy how many people here accused this innocent man of cheating. Imagine not watching the video.
Or seeing the thumbnail of the video....
They be speed running their comments, don't got time to check for accuracy. :P
Blame the thumbnail for that.
@@punkloser1465the thumbnail says "he's innocent"
the thumbnail that says "but he's innocent"?
A disappointing case of ego. I was really surprised one of the key points was 'responding too soon'.
Imagine playing a game so many times, it becomes second nature to be apprehensive about an area that can cause issues for your run. Sorry to hear Blink had to deal with something so poorly thought out and executed.
Yeah it was clearly a "oh i remember there's traps in here" and thenblink falls into a trap
Done it so many times it's not nearly enough expecally since its a reaction
Fr, “fake reactions” like….no, this man has just been playing and resetting for hours and is just getting increasingly frustrated.
This is something I brought up early in the investigation. In addition, what stream viewers are seeing is delayed by the upscaling and capture chain but he’s seeing it without that on a CRT.
As the old saying goes. Haters gonna hate.
Like if the traps positions don't move it's more than possible for them to have moved in a direction only to remember the trap in the position a bit too late and react to the trap they knew was there from likely countless runs.
people really think ChatGPT are omnipotent
No. Just able to create tutorials in this case. It’s Brood’s antisocial behavior that made him think doing tutorials was enough to accuse someone.
@@mogalixir Except as evidenced by this very video it can't even give you good tutorials... It's blind faith in a glorified autocomplete that was the real issue here. LLM's are garbage that cannot currently be trusted for tutorials, for information, for anything.
@@kezia8027the only thing you can really use ChatGPT for is just to have fun playing around with it and like nothing serious at all, and even then it’s limited becuase it’ll always go “as a large language model i cannot…” which is no fun rather than just bullshitting an answer which actually would be nicer in this case lol
No, only Brood.
No one in their right sane mind would use it for an investigation into splicing.
Wouldn't the correct term be "omniscient"?
Brood wants to be a Karl Jobst but became a MamaMax instead. Holy ChatGPT.
Nah, Karl isn't motivated by jealousy. Never has been.
@@abloogywoogywoo his not calling Karl that, he is saying that Brood had delusions of grandeur
@llwynoggwagle498 not that either, he just wanted to spite Blink and deny him the fruits of his own labor at grinding for better times.
I agree. The worst part is that these people seem to think Karl and Abyssoft only took off when they started talking about cheaters. They both have had successful YT careers long before Silly Bitchell became such a big thing. It's hard to believe they didn't think this was their fast track to fame. And that assumption is just wrong. Karl became huge after completing the Dam Agent speedrun untied WR, not by accusing people.
I haven’t finished the video yet but in my opinion the fact that Blink is so willing to go through all of this is making me think he’s innocent
I asked him a *lot* of tricky questions and he patiently supplied answers, video, and more. I had to work very hard to control bias. As in, my own, as I immediately suspected he was innocent based on how open he was.
I've just finished the video and honestly i'm kinda pissed at the accuser and feel so bad for blink....
He's cooperated the entire time it seems without any sorta pushback too to clear his name, also one of the final messages looks like the accuser still thinks he's a cheater which is WILD
This isn't a social deduction game, it doesn't matter how someone acts. The only thing that matters is what an analysis of the gameplay itself shows or doesn't show. I'm not saying anything one way or the other
I think he's innocent too, but you'd be surprised: statistically speaking, it's actually GUILTY people who are more likely to go along with the investigation, despite the reality that it very well may get them caught. On the contrary, people who are innocent [and improperly accused] are more likely to get (correctly) angry and lash out, _refusing_ to cooperate. That's just not what happened this time, obviously.
Human psychology is weird! :D
Where I know very little about what's going on and I mean next to nothing outside of this video. I agree with you because it's most likely the guilty person thinks they're never gonna get caught, so they're willing to go through with anything. They can even if it means that they will eventually get caught. Trust me, I've done it before in the past and I'm not happy with those moments when it comes to other things in life, not video games. Because if you're lying about video games.And cheating in such you have a sad life. If I'm going to the cheat in anything, I'm going to be open with it because I'm not part of a speed running community.I don't do that stuff.I video game for fun.So, if I'm cheating and somebody says that I am, I'll be like.Yes, you're right.I am and that's what I meant to do because i'm not competing for anything.@idontwantahandlethough
That "reacting before it happened" claim is complete BS; he said, softly "oh my god" because he was dreading what might be about to happen, I'm pretty confident I don't need to see greater context to be right about that.
That accusation earned a raised eyebrow the moment I heard it. It's a tile-based game, so you have the entire time it takes to move between tiles to regret hitting a button. As any gamer can tell you, frames feel a lot slower when you screw something up.
Yeah and his reaction did not even seem delayed anyway so it's just a moot point no matter how you look at it.
On top of that, I stream games frequently, and I often know when I've just hit a wrong button or executed the wrong action. Even if it hasn't happened yet, it's like I can feel it in my brain and it's too late to stop. Difficult to explain but sometimes you know you've messed up before before the results are seen.
Brood is a great example of the Dunning-Kruger effect in action. Learned a few things (poorly) about audio analysis and thought he was a sleuth, yet failed on something so simple as inrush current on startup.
To be fair the issues with his analysis aren't all that simple which is why he should have run his findings past an actual expert.
Audio is its own monster. There is a reason audio engineers exist. This guy tries to accuse someone (wildly) over "splices," but for what purpose?
He didn't understand enough about audio to make his conclusions, and his methodology got exposed to be faulty and idiotic. The cherry on top was using AI to try coercing it to match his "findings." XD
@@asterpw it is exactly why it is simple
This actually irks me a lot. I hope this wasn't the case, but it almost feels like this person wanted to make the next big "CHEATER EXPOSED" video and tossed around wild accusations at a speedrunner *WHO HAD HIM AS A VIP*.
Funny that I just watched another video about people asking chatGPT to find bugs in public code projects like curl and then submitting those bugs for bug bounties. This is causing a deluge of bogus reports, either reporting on bugs that were fixed years ago or just inventing bugs that don't exist, and it's a huge waste of time for the project developers to respond to these. AI is great for some stuff but it's also a good way for charlatans to generate nonsense for their own benefit.
Brood should have been fully transparent about his lack of experience and his full sources, and not been so sure of himself that he had caught a cheater. Like it's not just that his knowledge was questionable, he came at Blink like "Bro, I gotcha, you're done now mate just wait".
yeah it's kinda pathetic behaviour. Like that of a bully on the playground
Yeah. The issue was not with how much he taught himself but that he should have realized he was nowhere near knowing enough to accuse someone after what amounts to a dozen tutorials you find online.
He was jealous. Jealousy can make you do stupid hateful petty things.
ChatGPT for exposing cheating in speedruns is absolutely insane.
ChatGPT for anything you need an expert for is insane. Like I maybe use it if I can't find what I need in a couple of pages on google for things like drop sources in an MMO. No way in hell am I every going to trust it for anything an actual expert exists for. Especially when I have no f-ing clue if it's hallucinating or not.
If Brood clutched those straws any harder they'd collapse into black holes.
Imagine going through all this to accuse someone of splicing a >5 hour run because they got "consistently" lucky in one of the first splits of again, a >5 hour run. It's almost like the first chokepoint of a run that long is going to be by far the most attempts at it and therefore a lot of opportunities to get decent luck.
Unlike that Minecraft guy (forgot the name) who was constantly lucky this one is believable
@@ztriker4406 Worst part about Dream is he had the skill as he was actually great at the game he did not need to cheat and he would have gotten the record eventually. It's really sad to see a genuinely skilled person fall to cheating, but it's more common than you think with games that are very RNG heavy especially if the devs updated it to be more RNG heavy
I'm loving the fact that the legit audio engineers in the speedrunning community can actually hear the walls and the equipment in the room and determined it to be legit just from that. It's like having Batman himself explain in a multipage technical report how he did that one thing in The Dark Knight.
the other cool thing about mains hum is that there are slight fluctuations in the grid's frequency (e.g. if it's 60hz it might be between something like 59.99 and 60.01), and that can also be used to somewhat accurately date a clip if the region's grid publishes mains frequency data. NESO in the UK does this at 1 second resolution
Brood should consult chatGPT about why you shouldn't use LLMs as a substitute for actual knowledge/skills
I guess that Brood guy isn’t VIP anymore lol
The moral of the story is: large language models are not a viable replacement for proper research and education. You cannot skip the hard work required to learn a skill. Stories like this need to spread so that we can shake the public perception that LLMs are magic answer machines with perfect knowledge.
... ChatGPT 🤦
The moment I heard "Taught himself" I knew that this "Brood" character had made up his mind already and no amount of proof otherwise was going to work.
And then I find out he used _Chat GPT_
You could _not_ get more blatant in your reaching. I'm honestly disgusted (but not surprised) that he lied to himself so thoroughly .
I hope he gets kicked out of that community, because it's clear to me he doesn't deserve to be _in_ it.
not enough ppl are giving you credit for getting a bunch of qualified outside help to fact-check. good stuff. and banger vid 👍
I'm sorry, asking GPT to tell you how to prove your point and then refusing to acknowledge that you're wrong is absolute tool behavior. The first half can be a whoopsie, but the second half? Ban him from the community. Absolute tool.
There is absolutely nothing wrong with making a claim about potential speedrunning cheating, but you better be very sure that you have a baseline of evidence before making it public. If you can't be sure. You should reach out to people individually and ask them for help.
Unfotunately, there are many such cases!!!
ChatGPT users continue to prove that using the service actively tanks your intelligence to new lows.
well said
I'm still new to the speedrunning scene, so please forgive me for ny ignorance, but I just wanna know... why?
Why did Brood go out of his way to try and blame Blink for something as serious as this? Why didn't Brood be an adult and talk about this kinda stuff in private, instead of putting it on blast? Why was Brood using ChatGPT for serious claims?
Maybe it's just me, but I'm confused as to what Brood's endgame was from all this
Clout is a hell of a drug
Brood must be a fan of Karl Jobst or similar people and wanted to appear in their next video after catching a cheater in the act.
@@toumabyakuya Definitely a fan of Karl. Interestingly, Brood has the same problem as the people who cheat - he wanted to take a shortcut to get fame. In this case, he wanted to expose someone (anyone), but doesn't have the skill to do it or the patience to find a person who is actually cheating
Its likely a mix of jealousy, as he couldn’t be the one with a great run, and as people point out, wanting to be the one to catch a cheater
He's trying to get the drug called Clout.
Brood is a complete d-bag.... He was just out for blood for some personal reason it seems completely apparent with how hard he went after it, and how he released info even when being told not to.
It's because he's a competing runner in the same game. Pretty obvious he was trying to get iblink's WR time removed from the leaderboard.
This has been one of my favorite videos of yours... Super interesting going over the investigation yourself and seeing how you would do the investigation, as well as how it just goes to show that if you go into a study with the conclusion already in mind, you can find ways to find what you're expecting.
This is why accusations shouldn't be made by people who don't know what they're talking about. Armchair accusers hurt people's reputations and waste everybody's time. Brood should get a temporary sanction.
"how does this look in the sound thing?" has me wheezing 😭😭
Yeah I want a video about that. You know, the sound thing.
@@SushiElemental I'd watch the f out of that xD
Phenomenal video. Hits you with the audio analysis lore dump then hits you with a haymaker of a twist in that it was ChatGPT trying to screw the speedrunner the entire time.
brood using chatgpt in the first place is laughable. but he goes a step further to ignore what it says to get confirmation bias out of it. and bonus points for STILL thinking blink cheated his run despite being proven wrong. this man isn't a clown, he's whole damn circus. what a joke.
shame on the mods too for not punishing brood for this joke of a false accusation made with such weak and easily disproven evidence. bad actors like this needs to be and an example of.
Trusting chatgpt for anything is a bad idea but for complicated technical tasks? Pff. It can't even get basic historical dates and stuff correct for pubquiz questions
Not all tools are created equal.
9:45 this far it sounds like Brood has a personal vendetta against Blink.
Correct. Anything to stop him getting sub 5 hrs. Which Blink got anyway.
Brood can take his L and ask ChatGPT where to shove it.
"I wonder how he could come to such a conclusion."
The answer actually had me facepalming -.-
Basing anything on ChatGPT, which is known to hallucinate 'facts' is just so ridiculous
20:20 "Mainly, how did Brood come to the conclusion that there was evidence of splicing?"
Me: "Because he's an idiot."
*roughly 40 seconds later*
> ChatGPT
... oh, yup. Called it.
The way he prompts it too like "how to detect mains hum (or whatever it is called) in Adobe audition" he's trusting a predictive text bot to teach him stuff and he's not even asking the questions well, it's painful
-Moron trusts an "AI" to tell him how to find a hacker.
-Causes 300+ hours of trouble for a 100% legit streamer.
NEVER trust anything "AI". It's broken shartposting machines and nothing more.
The fake reaction complaint is odd to me. Speedrunners know when their inputs are wrong; they don't need to wait to see what happens to know their movement order wasn't right, so they could easily react just as or slightly before the results are shown on screen. Or maybe I'm missing something about the specific game?
to answer on this statement. you are 100% correct. i went the wrong way loading onto the stairs and knew since this was the second time in the same weekend i made the same mistake and ended up falling in that hole the previous run as well. i didn't know that you can avoid that hole (i wanted to take the shortest path to get back on track) since i normally never make that mistake, so i didnt know the backup
I thought people has learned after the lawyer used chatgpt - if it matters, you shouldn't use chatgpt
Poopoo on the accuser. I'd have some sympathy if, after all this, he let it go and apologize. Nope, just doubling down. He's got all the proof that Blink didn't cheat.
It's always easier to call someone a cheater than to be a real speedrunner.
Copper tech here, that burst distortion is what happens when a current is passing through copper that has partially oxidised. Its very common in old electronics.
Its common, someone gets suspicious tries to understand how to actually find out and either gets bad or even wrong information and reaches a faulty conclusion. It was handled in a bad way and we all should be careful with throwing allegations around.
Agreed. He found ONE "proof" when stuff like this requires multiple "proofs", or at least something that reinforced his one "proof". Also if he had more patience and checked his stuff with actual experts(like abyssoft with multiple persons and experts) this wouldn't happen.
I could see myself doing the first mistake, but in no way in hell all of the consectuve ones.
this is why you should wait for the proof check to come back before tossing around hasty accusations.
The people saying his innocence is proven by his transparency would be some of the same ones persecuting him if they saw the accusation video first.
If the caustic treatment made the runner quit sharing before this MASSIVE effort went into proof, the same transparency is proof logic would say he was guilty when instead people had just broken his heart.
Cheers those who did the hard work!
Cheers to the runner for rising above another challenge!
Can confirm
just hit a new PB in the Clicking A New Abyssoft Upload speedrun category
Probably used an autoclicker, cheater
someone needs to check the main hums with this one
Whenever I see people get accused, the mods have mountians of evidence and are very cautious about the accusation as they know it will possibly ruin someones speedrunning life. This guy just cared about catching someone and getting popular himself some steet cred.
Some people actually think ChatGPT is an all-knowing super search engine that can accurately explain any topic rather than a text predictor with a massive dataset and a penchant to people please. Those who believe the former rarely stop to consider that, maybe, if you could just ask the all-knowing genius box to solve a certain problem, someone else would have tried it already.
That first sentence does not line up with what i have seen/experienced
"ChatGPT falsely accuses speedrunner" would have been a better clickbait
Except chatGPT never accused anyone of anything. No this person got a crash course from GPT and thought he was all of a sudden an expert. I don't believe it was GPT that was in his Livestream chat pretty sure that was blink just based off of initial.observation.
Brood not blink my bad
@@davidwight5974Thats why it’s called “click BAIT”. It’s meant to be false
@@ToonyTails bet it does say click bait doesn't it XD welp time to take another reading class I guess
The ChatGPT reveal was positioned as the big twist that tore the entire narrative of the accuser down, so including it in the title would have required a restructuring of the video around another point (probably the reveal of innocence).
the fact people ACTUALLY use ChatGPT for something so serious tells me humanity is literally doomed. You're telling me you literally used chatGPT instead of learning yourself... and you exposed someone purely based on computer generated responses..... Bruh.
He used chatGPT to learn - the issue is he was still an amateur. What he did amounted to a dozen tutorials.
How dare people read a cookbook to learn how to cook! >:(
@@ZPu95cookbooks are written by people who actually cook, it's more like getting taught how to cook by someone who's never stepped foot in a kitchen, sure they might be able to guess how to cook an egg but at the end of the day they are guessing at every turn
@@Lacmene8 Anyone can write a cookbook so this is completely false.
I think you mean anyone can post recipes on the internet, to publish an actual cookbook you do need to know what you're doing (although someone did let Ned Fulmer publish one so it's not an inflatable system lol)
This is what Brooder had to say about the video on his Discord server:
> they didn't address my main concern (interference) and they didn't even look at the originally recorded footage to check for zero padding or other forensic markers and he misrepresented the chatGPT thing
So now not only is Blink still a cheater, but Abyssoft and everyone involved didn’t do a thorough enough investigation. He’s genuinely pathetic.
Also his friends / fans on the server are still treating Blink like he’s a cheater.
Leak the invite code
@@jellywillreturn if he did that the code would be deleted as you shouldn't promote harassment
This definitely feels like one of those things where he's working backwards, trying to find proof of his accusations, rather than analyzing the evidence to determine whether or not he's cheating
Can Brooder get awarded some type of Dunning-Kruger award? He basically speedran being a clueless regard by using chatGPT, so it should be on a leaderboard somewhere.
I think this is the first time I've seen "regard" used as a substitution of the r-slur, neat 👍
@@SynoPTL It's a stupid euphemism, like a lot of things from WallStreetBets.
Imagine just wanting to play games really fast, but you end up being the focus of a full-scale investigation just ‘cause some guy couldn’t believe that you were good at video games.
Brood def deserves a 6 month to a 1 year ban for everyone troubles
Lifetime. He went out of his way to try and ruin another persons credibility and it was likely out of jealousy at best.
I was NOT expecting Chat GPT being dragged into this...
I can't believe people actually use it to learn things, given how unreliable it is, and not just to cheese out college homework.
"He didn't want to make an accusation without evidence" proceeds to make accusations without evidence.
fabricated the evidence thanks to chatgpt wanting to fit the narrative of the human asking
False accusations should be treated almost as severely as faked runs. The individual doing this should never be involved in analysis of any speedrun ever again.
Brooder cannot just accuse Blink without VERIFIEDS evidences.
Immagine a mechanic (with no qualifications) telling you to replace the whole engine, just because he talked with an AI.
Cheating must be discouraged in competitive playing, but trying to humiliate someone that till now has play fair, is awful.
He should have worked on his allegations in private with a sound expert, and knowing Blink perificals devices. And after verify his finding with a mod, made the post public (especially not during a stream)
Maybe a question should be why the investigation team took his claim so seriously and did not verify brood's credentials first.
Maybe cause he provided some evidences about his theory, even if they weren't verified. At least now we know that someone without background experience shouldn't be listened.
@@Mario6732Because Brood didn’t back down and in fact doubled down even when given an opportunity to form a different opinion. Real damage was done here.
The main takeaway I got is that Blink is an absolute legend and Brood is a loser. I don't need ChatGPT to tell me that much
Shoutouts to everyone that helped defend him, he was accused unfairly and I'm very glad the community came to his aid
I think the use of ChatGPT makes it a tool-assisted speed accusation.
As soon as it was discovered he learned via ChatGPT, the whole allegation should have been thrown out (unless the experts were already well on their way to completing it, or done; I'm tired, I mighta missed that bit).
Based on what was brought up in this video, Blink handled things very professionally. Hats off to him for that, as well as for SMASHING that 5h window.