There is nothing inherently wrong with Boeing throttle controls or other parts. The main problems with Boeing is management outsourcing production to Spirit Aerospace, and management wanting to strap a big engine to a low wing 737, that behaves like older 737, such that pilots don't need training for a new aircraft (i.e. fixing hardware changes with software abstraction). ---- Personally I don't care that much what happens to Boeing, I'm not American. But it's a fact that people are bashing on Boeing for the wrong reasons.
10:26 No, what's funny about Y2K is that the reason it didn't actually cause any widespread disasters was precisely *because* of the media attention it garnered, which caused institutions and companies across the US (and, presumably, the world) to get their asses in gear and fix it before January 1 2000. I know this because I was part of the efforts to fix the issue in a certain national bank's systems. They were still using mainframes from the 60s and 70s and their software was still largely written in COBOL, it was a massive effort consisting of contractors like me coming in and updating their shit.
It's kind of like when people point to the fact that you don't hear about the ozone layer anymore. They use that as an argument that scientists were wrong and so they are probably wrong about climate change too. When really what happened was everyone got together and found solutions to fix the issue so you don't hear about it anymore because it's not as much of an issue anymore. Y2K everyone got their shit together, found solutions, and fixed the issue so it ended up being a non issue
The media would have had no effect on this for the developers. Whether or not the media reported on it didn't change that all potentially broken software needed to be updated. Its 100% true that media fearmongering vastly overrepresented the situation.
@letcreate123 its definitely not the most talked about computer bug, especially because it didn't even happen. Crowdstrike is the most talked about bug.
I often think about how us successfully preventing/fixing things like Y2K, the hole in the ozone layer, acid rain, etc... led to complacency and denial with covid, climate change, etc...
12:30 My 80 year old grandpa was a victim of this. He said his car suddenly accelerated on its own causing him to crash into our local gas station. Nobody believed him and just figured he was an old man and got confused, but he insisted he did not hit the gas. He complained to Toyota but they also discounted him. Literally a month after he died, my grandma got a letter in the mail about the recall and his proof that he really didn't accelerate.
This bug is why I'll never buy a Toyota. Not _that_ it happened but _how_ it happened. Their code was terrible, and I can't trust my life to terrible code.
nha nha it really is a feature! 0:10 "can you really call yourselves "peaceful" if you're not capable of causing great harm? or are you just harmless."
@@StrikerEureka85 I’m not commenting as a correction to the video, I’m just stating it’s interesting that even Sid Meier himself doesn’t know where it began :)
@@kkjdaniel Apparently it was due to Gandhi notifying the player that he has access to nukes at some point in the game, and the absurdity of Gandhi 'threatening' to use nukes made people create memes about it, even though he actually very rarely used them. Over the years the subtlety got lost and the whole thing became just 'Nuclear Gandhi'
@@Liword132 Also, in the original Civ1, you couldn't have two civs with the same colours in the game. The Civ that shared the colour with Indians was the Mongols, so if the most aggressive civ was not in the game, other civs had more opportunity to shine. So in a game with Genghis Khan, you'd have to fight off his frequent invasions, but in games without him, you'd see the world peacefully develop, until some scientific-oriented civ-like the Indians-invents nukes.
Worst part about the Therac was that the Company KNEW about it. But they did not want to fix it to save money and proceeded to only placebo fix the issue, which cost the life of multiple people and injured many more. Imagine winning against cancer, going to your last Session and then you get to expirience one of the most gruesome deaths imaginable just because some guy deemed your life less important than a bit of saved cash.
@@RillianGrant At that time it was unknown that it was a certain string of inputs that caused it and that it was easily replicable, they probably took their chances of it not happening again.
For sociopaths it's not about the money, it's all about saving face. These people are already filthy rich to begin with. What they care about is what others think of them, but in a twisted kind of way.
Not just made a world a better place but a more secure one SO MANY people switched from sWhatsApp to more secure app like Telegram and also it was literally free advertising for the alternatives messaging apps like Discord etc..
@@DRSDavidSoft The locks didn't "fail". In fact, they worked perfectly. The badge readers that were supposed to _unlock_ the door couldn't reach the systems needed to verify access, so they left the doors locked. (it's also a fire code violation)
That wasn't a bug per se, that's just how BGP works. Remember the guy accidentally announcing UA-cam's ranges? Yeah.. still working as intended. It's easy to screw up a system built on trust me bro.
He was talking about all the money wasted worrying so much about it. Of course most companies fixed their shot because otherwise they'd have huge problems.
I think he meant that ordinary people whose livelihoods were non computer related expected that banking systems, etc. would crash and hence stockpiled unnecessary amounts of food and other supplies which was money essentially wasted by them (instead of storing in a bank).
It's so frustrating. It was no big deal exactly *because* of the big ruckus. Only that made people spend the time, money, and effort to fix all of this. Without the ruckus, the people who thought it wouldn't be a big deal (and who oh so smugly said I-told-you-so afterwards) wouldn't have given us the budgets to fix it.
@@ThomasBlank-np5uvit's a lose-lose situation, you either warn people causing a mass panic that causes people to prepare for the disaster but people die in the panic, or you warn people but no one listens and dies in a disaster believing they were safe
You’re correct. But there still a lot of people that took advantage of the panic and sold snake oil preparation solutions. Which is probably the wasted money he is referring to. He is definitely oversimplifying but nothing he said is actually wrong. LGR has a great video on the subject.
@@srsa2436 He specifically says the money that was spent to fix it, not the money random people wasted stockpiling. He just didn't research this video enough and a lot of the information is wrong.
Small error for the y2k bug, the reason nothing happened was that many companies, governments, etc updated their systems and hardware so the integer overflow wouldn’t happen, the bug did cause damage but at a smaller scale than what people were expecting
I experienced a Y2K bug on my digital watch that incorrectly calculated leap-day in 2000. I vividly remember the day, because I was flying from Italy to the US. I looked at my watch and it displayed March 1st. But it was actually February 29th. The leap-year rule is every 4 years, unless the year is evenly divisible by 100 AND not divisible by 400. I guess the programmers didn't know about the 400 part.
I think the reason most people assume it wasn't a big deal is because most people thought it would effect all computers, while it was only a big problem in mainframes and other institutional systems that are invisible to the average computer users.
Which is the whole point, it was blown out of proportions, and was kind of a test run before the real Bullshit of the Millenium, which was the corona pandemic.
About the last Boeing thing: They didn't even announced the planes had this system, because then every pilot would need mandatory retraining. So when the planes did that the pilots had no effing idea what was going on. Oh and it was there to fix a engineering mistake "in software"...
Worse -- it wasn't to fix an engineering mistake. They deliberately designed the aircraft to avoid triggering retraining requirements and tried to work around the hardware problems this caused in software.
@@einargs Worse, the engineers knew about the lack of redundancy of MCAS, but they were told it is too expensive to add additional sensors, which would have solve the software limit
In short - Boeing murdered almost 350 people for profit! No redudancy and neither the airlines or pilots even knew the system existed, even less what to do when it went haywire because Boeing did not want to spend money on pilot training so they just kept quiet about MCAS and then - disasster!
And the reason they didn't want to retrain? Because the planes wouldn't sell. Airlines would have to ground pilots until their training completed, which would cost them lots of money and slow down air traffic for a few weeks. Boeing is absolutely culpable, but the MCAS shitshow is a child of unchecked greed and carelessness in the modern aviation industry as a whole.
The Y2K bug was real and while the world didn't end, it did create global issues. It's most likely due to the widespread media attention that most bugs were fixed to avoid severe disasters. However despite years of preparations and spending $100s of million, not everything was patched in time and issues did occur. Those issues ranged from funny things like offering a 105 year old woman a spot in a daycare center, all the way to severe issue with Nuclear Power Plants giving false radiation readings. Check Wikipedia for a documented list of confirmed issues that happened on Jan 1. 2000. There is also a list of confirmed fixes that were patched in time, which prevented severe issues.
Two fun facts abouty the Morris-worm: 1. It never hit Norway, because the call came over from the US to literally pull the plug on the connection to the rest of the network 2. Robert Tappan Morris' dad, Robert Morris, was the chief cryptologist (expert in ciphers and codes) of the National Security Agency
The Y2K bug was legit. Myself and many I used to know in the space got paid HUGE bucks to fix dates in old code. A buddy at the time actually bought a Supra (the F&F body style) within the first month of working for a large bank.
Y2k - My brother works at a hospital where they can't analyse the blood because of dates of birth and it reappears every 10 years, at which point they put a hacky fix in and wait another 10 years. This happens because it uses 2 digit years. Their hacky fixes involve finding symbols that they can substitute for numbers because of poor input validation, like using 15 for 1915 and !5 for 2015.
I'm a software dev, and these kind of shortcuts are pushed on us regularly. Most devs will learn early on in their careers that pushing back against it is futile. The execs want short term results for short term profits. They don't care about what happens in 10 years when they no longer have anything to do with the company.
@@BudgixousPretty sure it is, fireship made a voice model of his own voice to speed up his ability to put out videos. This sort of thing shows up on his other videos regularly. Doesn't mean the whole video is AI gen, but pretty sure he's largely automated the voiceline part of his videos.
Actually, there's a lot of bugs that do. I don't know why, but aside from Heartbleed, Spectre and others too have a logo. Someone with a spare PS license probably made it...and i kinda like it. xD
I just went through two bugs in the video. Really nice to see that you explained the actual issue from the point of view of a developer, some background and the repercussions. Thanks.
Y2K bug is a great example of the curse IT/infosec profession (also in many other areas as well, including lots of stuff CGI artists work on and also tons of prevention focused areas etc) suffers from; when you do everything right, no one notices it...
The code shown at 5:00 in the AT&T Network Switch Cascade crash is backwards. Break statements in C do not break out of if statements, only loops and switches. The issue was that it exited the switch statement early without doing the intended work, and later code overwrote the data that should have been processed.
Yeah, it is just straight up wrong in the video. "When the destination switch received the second of the two closely timed messages while it was still busy with the first (buffer not empty, line 7), the program should have dropped out of the if clause (line 7), processed the incoming message, and set up the pointers to the database (line 11). Instead, because of the break statement in the else clause (line 10), the program dropped out of the case statement entirely and began doing optional parameter work which overwrote the data (line 13). Error correction software detected the overwrite and shut the switch down while it couls reset."
@XDarkGreyX didn't know webb runs JS, although it appears to only use it for high level stuff i doubt any aircraft uses something like JS (without having done any research ofc)
@5:14: No, break statements don't interact with if blocks. Just switches and loops. I googled the bug, and other sources show something different from your screen (Something that looks a lot like Apple's goto fail bug from a few years ago.)
yeah, couldn't replicate the bug with gcc... perhaps they used another compiler? and what does the C standard has to say about this? 🤔 don't have the C standard at hand rn.
this confused me too as a c programmer. I had never heard of break statements working this way, sp I wrote an example program to test this and couldn't reproduce the bug, the break just exits the whole switch statement.
16:48 Big issue happened: "where were all the programmers when we needed them?" Big issue avoided: "damn, everyone really overreacted, nothing ever really happened, huh. Why don't they work on actual issues we need them on?"
@@XDarkGreyX the joke was better this time around because that shit wasn't even a software bug, it was a lunatic that even used an Logitech Controller to guide a DIY submarine to Titanic.
0:20 it's an overflow error. Underflow is when a floating-point operation is smaller in magnitude than the computer can represent. An integer wrapping in _either_ direction is an overflow error.
Damn, he is right. Over/under really means most/least significant digits side. Thought one could argue underflow happens with integers too, it is just much more expected.
@@RickWeberEcon You can do a binge for simple clay working with primitive technology and learn the writing from one of Irving Finkel's lessons about Cuneiform.
The chase bank situation wasn't a bug. I am saying this with a 100% straight face, it was a feature. The idea was that you could deposit the check and immediately have access to the funds as a form of convenience instead of having to wait for the bank to properly process your check and make you late on rent. The "chase glitch" was literally just check fraud, there was never a glitch involved.
That's funny, you don't look like Leena Kurjenniska. In fact, you don't look like a woman at all. And "Fortinet" is a funny name for Synopsys Software Integrity Group. It's almost as if you pulled your claim out of your ass or something.
the only reason Y2K didn't cause widespread disaster was because we prepared for it a number of less critical systems that was intended to be replaced was also left running with the intention to see what would happen, and as predicted they went completely nuts... so had we not spent the money to prepare for it, it quite likely would have been disastrous
11:09 the stupidest part of this situation was that this was a private entity that messed up trying to do something that is bordering on illegal(manipulating short positions in a malicious manner) and when they messed up they proceeded to socialize their losses and demand a bailout.
"Speaking as one of the devs who actually worked on the original Civ, yes Gandhi tended to nuke you. It was not intentional, but resulted from the fact that Gandhi usually didn't built much of a military, and advanced rapidly in tech. So when you betray your alliance with him and attack, his only recourse was to nuke you."
Sounds sus quoting someone without the name, also I would not be surprised if Sid Meier forgot or lied about the bug. half the internet believes him it was hoax.
A truly wonderful compilation of various bugs and issues. With your fast pace content, a 16 minute video feels like a 40 minute documentary, and I'm here for it 😎
8:20 The NASA Mars Climate Orbiter crash wasn't caused by 2 NASA teams miscommunicating with each other, but by a sensor part feeding data in lbf/s, even though the manufacturer's own documentation specificed that it returned N/s. It's a case of bad documentation, not of programmer error.
During my Intro to C course as a Sophomore back in 96, our professor talked at length about the Therac-25 problems. That has stuck with me for my whole career
Triggered! I graduated in summer 1998, and the very first work I was doing to fixing the Y2K problem! Nothing went wrong in 2000, BECAUSE I WORKED MY ASS OFF along with many others fixing the damn problem. The lack of gratitude winds me up.
Fun fact, some Y2K fixes were only temporary fixes as they figured they would deal with it properly once they had time. Now some of these fixes have run out in various Australia public services systems and people have to fix Y2K bugs in 2024.
Worst ever bug I've ever caused was I disabled a line of code that filters out car fleets by a trucking company's branch. My reason for doing so was to test an isolated bug on that particular page. I thought I had reverted it before checking in the code to TFS because we had a live release that night, but when I logged in the next morning, I noticed my mistake As soon as my lead developer came in, I informed him about my mistake and apologised profusely. He chastised me and said that I was "just wasting space" at the company; which contributed to the company's decision to make my position redundant a few days later I learnt a lot from that ordeal, and I occasionally make the same mistake but always catch it before it goes live. Thankfully no-one noticed EDIT: Technically it was breaking data protection law, this was before GDPR, which prevents data from one branch being viewed from a different branch
Your lead developer should have caught the bug during code review. Sounds to me like you weren't the only one who made a mistake that day, and he should get off your back if he isn't going to take responsibility for doing the things a lead developer is supposed to do.
@@futuza It wasn't because of that particular mistake. I was already on a performance improvement plan because I struggled early on in my career as a software developer; my work ethic was terrible, I was immature and very inexperienced The company was in the middle of creating a shortlist of people to get rid of and I was already on the chopping block. I was upset but secured a new position about a week later There were no code reviewing processes and the testing wasn't always confirmed by the testing team. The tech stack was also a complete disaster
I like the subtle reference the Bongo Moth brings when it shows up. In fact it was due to a moth that the term for software errors of that nature are called bugs.
The Boeing MCAS system was not to prevent stalling, but to ise software to augment the maneuvering characteristics of the 737 MAX to match those of previous iterations of the 737. This was done solely as a money saving exercise, since it allowed pilots to do a simple difference training instead of a new type rating drastically speeding up the time and reducing the cost to the airlines to get things rolling. They essentially did this because that's what Airbus did with the A320, but the A320 didn't require any fancy software to keep the same handling characteristics. The 737 did, because fitting the larger engines to the wings meant shifting everything on the plane ever so slightly. The other problem with the system is that it only ever read from a single Angle of Attack Vein. If you paid for both available sensors (yes, it was an optional add-on when it shouldn't have been), the system would just flip-flop between them at every system start-up. It should have been reading both sensors for redundancy, and in hindsight it is quite obvious why. The last big problem with the design was that it had more control authority than the pilots. If it started sending bad commands to the horizontal stabilizer, the ONLY thing a pilot can do is to completely de-power the stabilizer and use the manual trim wheel.
The horrifying thing about the Therac-25 bug is that the bug arose from experienced operators when inputting the incorrect procedure. The machine was supposed to "lock up" and administer the procedure but operators who were fast enough to correct the mistake before the lock up actually caused higher doses to be administered. So slow, unfamiliar operators wouldn't cause the error to trigger but fast, experienced ones did. I remember studying this when taking my computer ethics course in uni. I don't remember the full details, but that's the gist of it.
I actually got my start in IT because of the Y2K bug. I took in 1996 what is now called a bootcamp, where I learned COBOL (yes, I did). Before I could get a programmer job, I got a job as a tester, and I kept building on that over the years.
Alternate theory: because of all the attention, the critical Y2K bugs got dealt with ahead of time. I did experience some Y2K-related bugs in the first days of 2000 but they were minor and quickly corrected. I also remember identifying an easily-dealt-with issue with Intel 286-based systems. But considering the 386 had been out since 1985, we didn't have many 286 systems left where I worked.
8:55 French : tous les paramètres propulsifs sont normaux, la trajectoire [...] Translated : all propulsion parameters are normal, the trajectory [...]
love this!!! im so facinated by computer bugs but most videos either go way over my head or gloss over the root cause; understanding why they happened is what makes them so interesting to me. you hit a perfect balance, great explanations that a non-techie like me can still mostly understand :-)
I went to click Play on a video I paused, but I hit Next instead, and found myself in a Fireship video 4mins after posting. It was no accident. It was destiny.
Or, on the flip side, we can see just how error-prone human programmers are and the goalpost we have for AI that makes 0 mistakes is likely unreachable, but AI that still make mistakes will still be used because the alternative is humans that make mistakes.
@@charlielarson1350 I think it is a little more complex than that. Right now, programmers need to check the quality of code generated by AI. They should understand the code, so that the integration can be done. If the AI can do this without the human supervision, then there is other aspects of programming. The solution, documentation, new ideas, issues mentioned in the video and so on. If AI can also do these, then all knowledge based jobs are in danger not just programming. But this seems to a big jump from current state of AI. So, I think in the future, AI will help humans increase efficiency rather than replacing them.
@@krishnarajt1743 so you're wanting agent AI. If you want an AI to help you write code, it can do that. If you want an AI to help you write documentation, you can do that. If you want an AI to help you generate ideas based on what's already there, it can do that. We just need the piece of the puzzle that orchestrates all of these individual tasks together, AKA long-term planning rather than current request-response schema. ChatGPT was released 2 years ago and since then we haven't gone more than a few months without a new model getting better performance, 5x cheaper, new functionality and it hasn't slowed down. o1 model is their first generation of chain of thought and they're claiming they've utilized RL which completely changes the game if true. You think 5 years from now we'll still be considering AI a smart auto-complete? We are potentially 1 next generation model away from college graduates being completely useless.
Code in an aircraft, especially a military one, most likely is Spark and not C++. In the case of the MAX, you could argue it was not a bug to use only one sensor, since it was according to boeing specification. It was bad design.
I appreciate that you note Nuclear Gandhi being caused by a bug is an urban legend (I believe in actuality it's just a funny consequence of the way all leader AI behaves with nukes, combined with the fact that Gandhi's science focus means he'll always get nukes before everyone else). An actual example of a game bug being a feature would be move-canceling to perform combos in fighting games. Players found they could glitch attack animations to start their next attack early and create a string of attacks the opponent couldn't escape from, which, of course, is now a standard game mechanic for that whole genre.
I do not think the chase thing was a glitch. Just trying to be a good bank and letting people have there money right away and let the check clear later.
This man deserves an Ig Nobel Prize. His videos first make people laugh, and then make them think... and then make them save human lives and billions of dollars.
// TODO: I'll comment it out later
At least put your initials in there slacker!
🤣
@@noanyobiseniss7462 Ill only do it when blame still shows my name
lol
Noice
"The helicopter has automatic throttle control based on external sensors"
Just gonna check the manufac-
And it's Boeing.
Every Airbus and military jet fighter, plus most modern business jets have autothrottles. Luckily they all work fine
And this is even before they were dei
@@LuLeBe Its not luck. It is engineering 😅
There is nothing inherently wrong with Boeing throttle controls or other parts.
The main problems with Boeing is management outsourcing production to Spirit Aerospace, and management wanting to strap a big engine to a low wing 737, that behaves like older 737, such that pilots don't need training for a new aircraft (i.e. fixing hardware changes with software abstraction).
----
Personally I don't care that much what happens to Boeing, I'm not American. But it's a fact that people are bashing on Boeing for the wrong reasons.
don't stay near windows too much, always check for red dots around you, and never leave the house alone
10:26 No, what's funny about Y2K is that the reason it didn't actually cause any widespread disasters was precisely *because* of the media attention it garnered, which caused institutions and companies across the US (and, presumably, the world) to get their asses in gear and fix it before January 1 2000. I know this because I was part of the efforts to fix the issue in a certain national bank's systems. They were still using mainframes from the 60s and 70s and their software was still largely written in COBOL, it was a massive effort consisting of contractors like me coming in and updating their shit.
Yeah, this is *the* most talked about tech bug in history, why do UA-camrs keep spreading misinformation about it??
It's kind of like when people point to the fact that you don't hear about the ozone layer anymore. They use that as an argument that scientists were wrong and so they are probably wrong about climate change too. When really what happened was everyone got together and found solutions to fix the issue so you don't hear about it anymore because it's not as much of an issue anymore.
Y2K everyone got their shit together, found solutions, and fixed the issue so it ended up being a non issue
The media would have had no effect on this for the developers. Whether or not the media reported on it didn't change that all potentially broken software needed to be updated. Its 100% true that media fearmongering vastly overrepresented the situation.
@letcreate123 its definitely not the most talked about computer bug, especially because it didn't even happen. Crowdstrike is the most talked about bug.
I often think about how us successfully preventing/fixing things like Y2K, the hole in the ozone layer, acid rain, etc... led to complacency and denial with covid, climate change, etc...
12:30 My 80 year old grandpa was a victim of this. He said his car suddenly accelerated on its own causing him to crash into our local gas station. Nobody believed him and just figured he was an old man and got confused, but he insisted he did not hit the gas. He complained to Toyota but they also discounted him. Literally a month after he died, my grandma got a letter in the mail about the recall and his proof that he really didn't accelerate.
Womp womp
I'm sorry for your loss
I remember some were dismissing it as BS, even I was skeptical till now. heh
This bug is why I'll never buy a Toyota. Not _that_ it happened but _how_ it happened. Their code was terrible, and I can't trust my life to terrible code.
nha nha it really is a feature! 0:10
"can you really call yourselves "peaceful" if you're not capable of causing great harm? or are you just harmless."
The myth of the Gandhi bug still gets perpetuated but Sid Meier himself confirmed in his book that it never happened and isn’t even sure how it began.
hence "urban legend"
I was just typing this then saw your comment. Its funny cause the whole video is on bugs
@@StrikerEureka85 I’m not commenting as a correction to the video, I’m just stating it’s interesting that even Sid Meier himself doesn’t know where it began :)
@@kkjdaniel Apparently it was due to Gandhi notifying the player that he has access to nukes at some point in the game, and the absurdity of Gandhi 'threatening' to use nukes made people create memes about it, even though he actually very rarely used them. Over the years the subtlety got lost and the whole thing became just 'Nuclear Gandhi'
@@Liword132 Also, in the original Civ1, you couldn't have two civs with the same colours in the game. The Civ that shared the colour with Indians was the Mongols, so if the most aggressive civ was not in the game, other civs had more opportunity to shine. So in a game with Genghis Khan, you'd have to fight off his frequent invasions, but in games without him, you'd see the world peacefully develop, until some scientific-oriented civ-like the Indians-invents nukes.
Worst part about the Therac was that the Company KNEW about it. But they did not want to fix it to save money and proceeded to only placebo fix the issue, which cost the life of multiple people and injured many more.
Imagine winning against cancer, going to your last Session and then you get to expirience one of the most gruesome deaths imaginable just because some guy deemed your life less important than a bit of saved cash.
What was their calculation for that? Dead patients tend to be extremely expensive.
@@RillianGrant At that time it was unknown that it was a certain string of inputs that caused it and that it was easily replicable, they probably took their chances of it not happening again.
Companies face no actual punishment for ending people's lives, so its just an acceptable loss and cost of doing business for them which is despicable.
For sociopaths it's not about the money, it's all about saving face. These people are already filthy rich to begin with. What they care about is what others think of them, but in a twisted kind of way.
@@Xeonerable Even worse, they actually go after the programmers, instead of their managers and CEOs...
Let's not forget Meta's BGP bug which took down Facebook & Instagram making the world a better place while it lasted.
Not just made a world a better place but a more secure one SO MANY people switched from sWhatsApp to more secure app like Telegram and also it was literally free advertising for the alternatives messaging apps like Discord etc..
And that it also caused the data center door locks to fail
@@DRSDavidSoft The locks didn't "fail". In fact, they worked perfectly. The badge readers that were supposed to _unlock_ the door couldn't reach the systems needed to verify access, so they left the doors locked. (it's also a fire code violation)
That wasn't a bug per se, that's just how BGP works. Remember the guy accidentally announcing UA-cam's ranges? Yeah.. still working as intended. It's easy to screw up a system built on trust me bro.
Not really a bug, just an employee messing it up. And it's pretty easy to do given the most important internet protocols rely on "trust me bro"
10:32 Y2K never caused and disasters BECAUSE the money was spent on fixing things
Classic example of the curse of a successful intervention.
"Why all the panic? Nothing happened!".
I was looking for this comment, we spent half a trillion to prevent it, next big thing will be PQC
@@lexus4tw Y2K38 could be interesting too... there's probably still a lot of 32 bit systems used in critical environments
It’s insane he actually put that in the script without researching it or thinking about it for a second.
He was talking about all the money wasted worrying so much about it. Of course most companies fixed their shot because otherwise they'd have huge problems.
We tested, found, and fixed dozens of unix Y2K bugs. That money was *not* wasted.
I think he meant that ordinary people whose livelihoods were non computer related expected that banking systems, etc. would crash and hence stockpiled unnecessary amounts of food and other supplies which was money essentially wasted by them (instead of storing in a bank).
It's so frustrating. It was no big deal exactly *because* of the big ruckus. Only that made people spend the time, money, and effort to fix all of this. Without the ruckus, the people who thought it wouldn't be a big deal (and who oh so smugly said I-told-you-so afterwards) wouldn't have given us the budgets to fix it.
@@ThomasBlank-np5uvit's a lose-lose situation, you either warn people causing a mass panic that causes people to prepare for the disaster but people die in the panic, or you warn people but no one listens and dies in a disaster believing they were safe
You’re correct. But there still a lot of people that took advantage of the panic and sold snake oil preparation solutions. Which is probably the wasted money he is referring to. He is definitely oversimplifying but nothing he said is actually wrong. LGR has a great video on the subject.
@@srsa2436 He specifically says the money that was spent to fix it, not the money random people wasted stockpiling. He just didn't research this video enough and a lot of the information is wrong.
Small error for the y2k bug, the reason nothing happened was that many companies, governments, etc updated their systems and hardware so the integer overflow wouldn’t happen, the bug did cause damage but at a smaller scale than what people were expecting
I experienced a Y2K bug on my digital watch that incorrectly calculated leap-day in 2000. I vividly remember the day, because I was flying from Italy to the US. I looked at my watch and it displayed March 1st. But it was actually February 29th. The leap-year rule is every 4 years, unless the year is evenly divisible by 100 AND not divisible by 400. I guess the programmers didn't know about the 400 part.
yeah, programmers won the day essentially. companies relying on software spent a looooot of resources on identifying y2k problems in 99. ^^
@@carrion1234 my college still uses software so old that it has f13 and above function keys
I think the reason most people assume it wasn't a big deal is because most people thought it would effect all computers, while it was only a big problem in mainframes and other institutional systems that are invisible to the average computer users.
Which is the whole point, it was blown out of proportions, and was kind of a test run before the real Bullshit of the Millenium, which was the corona pandemic.
About the last Boeing thing:
They didn't even announced the planes had this system, because then every pilot would need mandatory retraining.
So when the planes did that the pilots had no effing idea what was going on.
Oh and it was there to fix a engineering mistake "in software"...
Worse -- it wasn't to fix an engineering mistake. They deliberately designed the aircraft to avoid triggering retraining requirements and tried to work around the hardware problems this caused in software.
Worse still that there was no option to manually disable the MCAS if something did go wrong
@@einargs Worse, the engineers knew about the lack of redundancy of MCAS, but they were told it is too expensive to add additional sensors, which would have solve the software limit
In short - Boeing murdered almost 350 people for profit! No redudancy and neither the airlines or pilots even knew the system existed, even less what to do when it went haywire because Boeing did not want to spend money on pilot training so they just kept quiet about MCAS and then - disasster!
And the reason they didn't want to retrain? Because the planes wouldn't sell. Airlines would have to ground pilots until their training completed, which would cost them lots of money and slow down air traffic for a few weeks.
Boeing is absolutely culpable, but the MCAS shitshow is a child of unchecked greed and carelessness in the modern aviation industry as a whole.
The reason Y2K was not a catastrophe is due to thousands of people like me working 7 days per week remediating old code.
The Y2K bug was real and while the world didn't end, it did create global issues. It's most likely due to the widespread media attention that most bugs were fixed to avoid severe disasters. However despite years of preparations and spending $100s of million, not everything was patched in time and issues did occur. Those issues ranged from funny things like offering a 105 year old woman a spot in a daycare center, all the way to severe issue with Nuclear Power Plants giving false radiation readings. Check Wikipedia for a documented list of confirmed issues that happened on Jan 1. 2000. There is also a list of confirmed fixes that were patched in time, which prevented severe issues.
Two fun facts abouty the Morris-worm:
1. It never hit Norway, because the call came over from the US to literally pull the plug on the connection to the rest of the network
2. Robert Tappan Morris' dad, Robert Morris, was the chief cryptologist (expert in ciphers and codes) of the National Security Agency
Mom, I’m not a failure but a feature :p
someone pin this comment pls
@@Equalisys 🤣
Mee 2!
Bro has comment pfp
Todd Howard: "I'm not your mom but i know son. "
I feel the corrupted blood incident in WoW deserved a mention.
Was that a bug? It was a lot of fun!
Lol, I remember that sh*t lololololololololololololololol
The Y2K bug was legit. Myself and many I used to know in the space got paid HUGE bucks to fix dates in old code. A buddy at the time actually bought a Supra (the F&F body style) within the first month of working for a large bank.
Y2k - My brother works at a hospital where they can't analyse the blood because of dates of birth and it reappears every 10 years, at which point they put a hacky fix in and wait another 10 years.
This happens because it uses 2 digit years. Their hacky fixes involve finding symbols that they can substitute for numbers because of poor input validation, like using 15 for 1915 and !5 for 2015.
Yikes and they are aware of this.. What could go wrong
ah, cobol. lol
I'm a software dev, and these kind of shortcuts are pushed on us regularly. Most devs will learn early on in their careers that pushing back against it is futile. The execs want short term results for short term profits. They don't care about what happens in 10 years when they no longer have anything to do with the company.
Surely after 9 they could move to A. That gives them another 260 years.
8:07 I’ve never heard NASA pronounce Nassau before. Must be a feature.
Still beats people pronouncing Nassau "Na-su-aw".
This is because these are AI generated
@@avwie132 no it isn't...
That didn't sound like Nassau at all, much closer to "Na-saw"
@@BudgixousPretty sure it is, fireship made a voice model of his own voice to speed up his ability to put out videos. This sort of thing shows up on his other videos regularly. Doesn't mean the whole video is AI gen, but pretty sure he's largely automated the voiceline part of his videos.
Can't wait for Y2038.
I love that Heartbleed is so notorious of a bug that it has its own logo
Actually, there's a lot of bugs that do. I don't know why, but aside from Heartbleed, Spectre and others too have a logo.
Someone with a spare PS license probably made it...and i kinda like it. xD
I think Heartbleed (or something around that timeframe) was what started the trend of giving major vulnerabilities a logo and their own website.
@MichaelPerna1289
9 hours ago
Omg I designed that heartbleed logo when I worked at Fortinet
He created this logo!
Dude who designed the logo commented above. Seems everyone is watching Fireship. :)
@@TDCIYB77 Wow, I had no idea!
I just went through two bugs in the video. Really nice to see that you explained the actual issue from the point of view of a developer, some background and the repercussions. Thanks.
Y2K bug is a great example of the curse IT/infosec profession (also in many other areas as well, including lots of stuff CGI artists work on and also tons of prevention focused areas etc) suffers from; when you do everything right, no one notices it...
The code shown at 5:00 in the AT&T Network Switch Cascade crash is backwards. Break statements in C do not break out of if statements, only loops and switches. The issue was that it exited the switch statement early without doing the intended work, and later code overwrote the data that should have been processed.
Yeah, it is just straight up wrong in the video. "When the destination switch received the second of the two closely timed messages while it was still busy with the first (buffer not empty, line 7), the program should have dropped out of the if clause (line 7), processed the incoming message, and set up the pointers to the database (line 11). Instead, because of the break statement in the else clause (line 10), the program dropped out of the case statement entirely and began doing optional parameter work which overwrote the data (line 13). Error correction software detected the overwrite and shut the switch down while it couls reset."
*Me slightly panicking that I've written some deadly code because I've been wrong for years*
16:28 I like the idea of the plane's systems running on JS
would explain a multitude of things
My mind always jumps to Webb and I think there are more things running on JS than you think. Node is not the biggest crime.
@XDarkGreyX didn't know webb runs JS, although it appears to only use it for high level stuff
i doubt any aircraft uses something like JS (without having done any research ofc)
Bet it uses C witch I refuse to learn lol
@@joshua476 it's good, especially to learn how programs work in the lower level
@5:14: No, break statements don't interact with if blocks. Just switches and loops. I googled the bug, and other sources show something different from your screen (Something that looks a lot like Apple's goto fail bug from a few years ago.)
GCC and G++ agree with you
yeah, couldn't replicate the bug with gcc... perhaps they used another compiler? and what does the C standard has to say about this? 🤔 don't have the C standard at hand rn.
+1, was looking for this comment
this confused me too as a c programmer. I had never heard of break statements working this way, sp I wrote an example program to test this and couldn't reproduce the bug, the break just exits the whole switch statement.
i guess Fireship doesnt know C and only javascript n html
Damn sponsorblock works fast, 1 minute after upload
Can't watch youtube without it these days..
@@HiImKyle I wish it worked in incognito mode.
It works if you enable to work in incognito@@muhdiversity7409
@@muhdiversity7409 you can make it work (check the "allow in incognito mode" box in its options page)
Goated community
16:48 Big issue happened: "where were all the programmers when we needed them?" Big issue avoided: "damn, everyone really overreacted, nothing ever really happened, huh. Why don't they work on actual issues we need them on?"
14:13 Looks like the bugs couldn't escape the "THREAC-25" again
0:16 this Gandhi bug is a myth
Parallel universe 😂
Off to a great start when 15 seconds into the video they already are repeating a myth as fact.
There's something funny about the same car manufacturer having a breaking problem and an acceleration problem.
Writing faulty code that ends uo killing someone is a nightmare scenario.
0:35 "real men test in prod"
I WAS DRINKING MY COFFEE WHEN THIS APPEARED, THANKS!
remember to feed your monitor!
He has used the joke 3 times at least. You must be new here.
@@XDarkGreyX the joke was better this time around because that shit wasn't even a software bug, it was a lunatic that even used an Logitech Controller to guide a DIY submarine to Titanic.
0:20 it's an overflow error. Underflow is when a floating-point operation is smaller in magnitude than the computer can represent. An integer wrapping in _either_ direction is an overflow error.
Its actually not and the bug never existed
@@AdoreHorror Which was also mentioned in the video -_-
Damn, he is right.
Over/under really means most/least significant digits side.
Thought one could argue underflow happens with integers too, it is just much more expected.
@@AdoreHorror it is - look it up. And I was never arguing the bug existed.
It's also not an unsigned *integer* if it overflows to 255, it's a char or a byte
Of all the channels encouraging me to learn book binding and calligraphy, Fireship is my favorite!
If you want a real hard copy, go all the way back to kiln burned clay tablets.
@@OperationDarkside I’m probably due for a three week binge of Primative Technology. Maybe he’s got a video on knot language or something
@@RickWeberEcon You can do a binge for simple clay working with primitive technology and learn the writing from one of Irving Finkel's lessons about Cuneiform.
The chase bank situation wasn't a bug. I am saying this with a 100% straight face, it was a feature. The idea was that you could deposit the check and immediately have access to the funds as a form of convenience instead of having to wait for the bank to properly process your check and make you late on rent.
The "chase glitch" was literally just check fraud, there was never a glitch involved.
16 minutes of fireship video lets gooo
Omg I designed that heartbleed logo when I worked at Fortinet
good job mate
GJ! It was all over the internet for that one year when Heartbleed was identified
no you didn't
That's funny, you don't look like Leena Kurjenniska. In fact, you don't look like a woman at all.
And "Fortinet" is a funny name for Synopsys Software Integrity Group.
It's almost as if you pulled your claim out of your ass or something.
Fr?
15:42 Fireship is not suicidal
We all agree to this statement 100%
Correction, the 2003 blackout contributed to over 100 deaths (no traffic lights, chaos, etc.)
Its just a statistic. No one actually paid for the crime\mistake.
Wtf how do you die of chaos when its just dark
Hey man, you take that back. Zune was awesome. I'll die on that hill 😆
0:14 The Ghandi nuke thing in Civ is a myth. Some guy made a joke one time and it became an undying meme. It never happened.
Yo I just looked it up and can't believe what you are saying is true. My life has been a lie...
He did specify it's the Urban Legend
the only reason Y2K didn't cause widespread disaster was because we prepared for it
a number of less critical systems that was intended to be replaced was also left running with the intention to see what would happen, and as predicted they went completely nuts... so had we not spent the money to prepare for it, it quite likely would have been disastrous
3:35 that’s not *a glitch*. That’s one of the oldest forms of bank fraud. It’s check kiting
15:28 That wasn't an accident.
Also, murrica never apologized, never paid reparations.
Iran "recently" downed ukrainian aircraft, now they kill them with their drones and missiles.
@@SirusStarTV you are justifying a bad thing with a bad thing , it doesn't work like that
The lesser of 2 evils don't always need to be put as an example. Just vote in don't vote out.
Thats a question for philosophy yt channel
For the replier not for the OP of comment
@@SirusStarTV because they were under threat of bombardment by the USA
the Gandhi civilization bug turns out to be, funnily enough, a case of the Mandela effect
No, it's just misinformation.
@@MarcinKralka which is what Mandela effects are. People rember shit wrong all the time. You can't forgot that you forgot.
People keep saying that but was Gandhi still going ballistic out of nowhere in that game (on purpose)? Or that part was also a legend?
11:09 the stupidest part of this situation was that this was a private entity that messed up trying to do something that is bordering on illegal(manipulating short positions in a malicious manner) and when they messed up they proceeded to socialize their losses and demand a bailout.
Classic late stage capitalism move
@@hugomazeas4297 There's nothing capitalist about bailouts.
@@hugomazeas4297 Demanding a Bailout is classic late stage Socialism move
@@hugomazeas4297government bailouts are capitalism now?
"Speaking as one of the devs who actually worked on the original Civ, yes Gandhi tended to nuke you. It was not intentional, but resulted from the fact that Gandhi usually didn't built much of a military, and advanced rapidly in tech. So when you betray your alliance with him and attack, his only recourse was to nuke you."
Who are you quoting?
Sounds sus quoting someone without the name, also I would not be surprised if Sid Meier forgot or lied about the bug. half the internet believes him it was hoax.
@ it was from the other video on the myth. I’ve heard sid discuss the same logic in other interviews.
@@derodomtommy3716 it's a total myth
8:36 I like to point out that 1996 was 28 years ago. NOT "few".
LALALA CAN'T HEAR YOU.
he is talking about the orbiter fail which was in '99. So 1996 is just 3 years before that. don't post ret**ded comments.
A truly wonderful compilation of various bugs and issues. With your fast pace content, a 16 minute video feels like a 40 minute documentary, and I'm here for it 😎
Bro you could make an entire series just with the content of this video
8:20 The NASA Mars Climate Orbiter crash wasn't caused by 2 NASA teams miscommunicating with each other, but by a sensor part feeding data in lbf/s, even though the manufacturer's own documentation specificed that it returned N/s. It's a case of bad documentation, not of programmer error.
3:20 this wasn't a glitch, this wasn't a bug. It how the system works and is supposed to work. All they did was cheque fraud.
0:59 CRAPPY????? are you kidding me Zune was way better than ipods. Loved mine so much and the software was actually amazing
Yup that was my jam, the interface looked so much nicer than the iPodds imo
I loved my zune hd
zune RULED!
Lmao 🤣
Typical Zune autist
This video should be included in all introduction to programming courses. All of them!
During my Intro to C course as a Sophomore back in 96, our professor talked at length about the Therac-25 problems. That has stuck with me for my whole career
Thanks, I will show this video at my boss when I take entire prod down on a friday afternoon so he knows it's not that bad after all
Triggered!
I graduated in summer 1998, and the very first work I was doing to fixing the Y2K problem!
Nothing went wrong in 2000, BECAUSE I WORKED MY ASS OFF along with many others fixing the damn problem.
The lack of gratitude winds me up.
Gratitude?
I am sure that you were compensated
Fun fact, some Y2K fixes were only temporary fixes as they figured they would deal with it properly once they had time. Now some of these fixes have run out in various Australia public services systems and people have to fix Y2K bugs in 2024.
Worst ever bug I've ever caused was I disabled a line of code that filters out car fleets by a trucking company's branch. My reason for doing so was to test an isolated bug on that particular page. I thought I had reverted it before checking in the code to TFS because we had a live release that night, but when I logged in the next morning, I noticed my mistake
As soon as my lead developer came in, I informed him about my mistake and apologised profusely. He chastised me and said that I was "just wasting space" at the company; which contributed to the company's decision to make my position redundant a few days later
I learnt a lot from that ordeal, and I occasionally make the same mistake but always catch it before it goes live. Thankfully no-one noticed
EDIT: Technically it was breaking data protection law, this was before GDPR, which prevents data from one branch being viewed from a different branch
working people make mistakes. if you have to chastise people for making errors then you are a bad leader
Your lead developer should have caught the bug during code review. Sounds to me like you weren't the only one who made a mistake that day, and he should get off your back if he isn't going to take responsibility for doing the things a lead developer is supposed to do.
Ouch, but their loss for firing you for a mistake anyone could have made due to their incompetent code review and testing processes
@@futuza It wasn't because of that particular mistake. I was already on a performance improvement plan because I struggled early on in my career as a software developer; my work ethic was terrible, I was immature and very inexperienced
The company was in the middle of creating a shortlist of people to get rid of and I was already on the chopping block. I was upset but secured a new position about a week later
There were no code reviewing processes and the testing wasn't always confirmed by the testing team. The tech stack was also a complete disaster
Sid Myers said it himself in interviews that the Ghandi's bug is a myth
Hence "urban legend"
You should have included the "cannot print on tuesdays" bug
I like the subtle reference the Bongo Moth brings when it shows up. In fact it was due to a moth that the term for software errors of that nature are called bugs.
Fireship: one of the few channels you don’t need to watch at 1.25 speed 😊
5:38 i got your sponsor as an ad with the same voice lol
I think it's time to change my spaghetti C++ code to memory leaking unsafe untested Rust
a moth playing the bongos is now my favourite representation of a bug lmao
@16:50 Love this rounding error
The Boeing MCAS system was not to prevent stalling, but to ise software to augment the maneuvering characteristics of the 737 MAX to match those of previous iterations of the 737.
This was done solely as a money saving exercise, since it allowed pilots to do a simple difference training instead of a new type rating drastically speeding up the time and reducing the cost to the airlines to get things rolling.
They essentially did this because that's what Airbus did with the A320, but the A320 didn't require any fancy software to keep the same handling characteristics. The 737 did, because fitting the larger engines to the wings meant shifting everything on the plane ever so slightly.
The other problem with the system is that it only ever read from a single Angle of Attack Vein. If you paid for both available sensors (yes, it was an optional add-on when it shouldn't have been), the system would just flip-flop between them at every system start-up. It should have been reading both sensors for redundancy, and in hindsight it is quite obvious why.
The last big problem with the design was that it had more control authority than the pilots. If it started sending bad commands to the horizontal stabilizer, the ONLY thing a pilot can do is to completely de-power the stabilizer and use the manual trim wheel.
Holy crap. This puts the importance into perspective. Pretty sure I won’t be dreading fixing bugs, writing tests and handling exceptions from now on
9:00 its illarious how the guy says everything if fine so far as the rocket blows up xD
FYI braking bug scene is from the movie: Hangar 18 (1980).
Yep. MST3K did the movie back in the UHF days of the show (pre-cable). I like the movie (and MST3K's handling of it).
I really appreciate your content! Each one of your videos is, somehow, diabolically more enlightening than the prior.
The voice transition at 6:39 : it's not a bug, it's a feature
AI voice
Fireship is one of the few channels that I watch at normal speed
Fun Fact: The first computer bug was an actual bug-a moth found in a Mark II computer in 1947.
The horrifying thing about the Therac-25 bug is that the bug arose from experienced operators when inputting the incorrect procedure. The machine was supposed to "lock up" and administer the procedure but operators who were fast enough to correct the mistake before the lock up actually caused higher doses to be administered. So slow, unfamiliar operators wouldn't cause the error to trigger but fast, experienced ones did.
I remember studying this when taking my computer ethics course in uni. I don't remember the full details, but that's the gist of it.
if (isStalling && sensor1 == sensor2) { pushNoseDown() }🤣Awesome as always Fireship
I actually got my start in IT because of the Y2K bug. I took in 1996 what is now called a bootcamp, where I learned COBOL (yes, I did). Before I could get a programmer job, I got a job as a tester, and I kept building on that over the years.
Alternate theory: because of all the attention, the critical Y2K bugs got dealt with ahead of time.
I did experience some Y2K-related bugs in the first days of 2000 but they were minor and quickly corrected.
I also remember identifying an easily-dealt-with issue with Intel 286-based systems. But considering the 386 had been out since 1985, we didn't have many 286 systems left where I worked.
That's what actually happened. A real issue that got fixed thanks to the hard work of tens of thousands of people
lol, that bit of code(00:52) is why I love this channel.
ive been trying to cancel netflix for weeks
I got rid of Netflix by assigning my account to my ex. Done and dusted.
@@muhdiversity7409 based
is that a hard process?
I (as a backend developer) like your way of presenting a frontend bug after each backend bug 😀
8:55
French : tous les paramètres propulsifs sont normaux, la trajectoire [...]
Translated : all propulsion parameters are normal, the trajectory [...]
love this!!! im so facinated by computer bugs but most videos either go way over my head or gloss over the root cause; understanding why they happened is what makes them so interesting to me. you hit a perfect balance, great explanations that a non-techie like me can still mostly understand :-)
I went to click Play on a video I paused, but I hit Next instead, and found myself in a Fireship video 4mins after posting. It was no accident. It was destiny.
i appreciate the amount of research went into this , thanks Jeff!
The "Nuclear Gandhi" bug never actually existed in Civ 1. It was an internet myth that ended up ascending into an actual thing in later Civ games.
you must be very old
Hey! The Zune was awesome. OLED screen, SSD, and crystal clear audio back in the iPod days!
This shows how complex the industry is and gives me confidence that AI will not replace programmers.
Or, on the flip side, we can see just how error-prone human programmers are and the goalpost we have for AI that makes 0 mistakes is likely unreachable, but AI that still make mistakes will still be used because the alternative is humans that make mistakes.
@@charlielarson1350 I think it is a little more complex than that. Right now, programmers need to check the quality of code generated by AI. They should understand the code, so that the integration can be done. If the AI can do this without the human supervision, then there is other aspects of programming. The solution, documentation, new ideas, issues mentioned in the video and so on. If AI can also do these, then all knowledge based jobs are in danger not just programming. But this seems to a big jump from current state of AI. So, I think in the future, AI will help humans increase efficiency rather than replacing them.
@@krishnarajt1743 so you're wanting agent AI. If you want an AI to help you write code, it can do that. If you want an AI to help you write documentation, you can do that. If you want an AI to help you generate ideas based on what's already there, it can do that. We just need the piece of the puzzle that orchestrates all of these individual tasks together, AKA long-term planning rather than current request-response schema. ChatGPT was released 2 years ago and since then we haven't gone more than a few months without a new model getting better performance, 5x cheaper, new functionality and it hasn't slowed down. o1 model is their first generation of chain of thought and they're claiming they've utilized RL which completely changes the game if true. You think 5 years from now we'll still be considering AI a smart auto-complete? We are potentially 1 next generation model away from college graduates being completely useless.
The segues from one bug to the next are masterful 🙂
Code in an aircraft, especially a military one, most likely is Spark and not C++. In the case of the MAX, you could argue it was not a bug to use only one sensor, since it was according to boeing specification. It was bad design.
Spark is a flavor of Ada, isn’t it?
Digital gold vid right here, compiled news fast paced, really like this format 👍🏼🙏🏼
Watching this instead of bug fixing...
I appreciate that you note Nuclear Gandhi being caused by a bug is an urban legend (I believe in actuality it's just a funny consequence of the way all leader AI behaves with nukes, combined with the fact that Gandhi's science focus means he'll always get nukes before everyone else).
An actual example of a game bug being a feature would be move-canceling to perform combos in fighting games. Players found they could glitch attack animations to start their next attack early and create a string of attacks the opponent couldn't escape from, which, of course, is now a standard game mechanic for that whole genre.
I do not think the chase thing was a glitch. Just trying to be a good bank and letting people have there money right away and let the check clear later.
This man deserves an Ig Nobel Prize. His videos first make people laugh, and then make them think... and then make them save human lives and billions of dollars.
8:15 for people still defending the imperial system. Utter insanity
16:18 Fireship just mentioned my country. Wuhu
Only to be an example for someone's mistake
😂😊
What a good Halloween movie. A 16mn Fireship video 🥳🥳
AND lets let AI run 2 ton death machines without so much as passing the same driving test a human must pass.