thank you for WAITING 2 MONTHS for a video lol. i had to finish up school for 3 whole weeks then make this thing. so yeah. anyways sources and music are below as always sources: en.wikipedia.org/wiki/Unix_time - wikipedia article on unix time en.wikipedia.org/wiki/2,147,483,647 - 2 147 483 647 en.wikipedia.org/wiki/Power_of_two#4294967296 - 4 294 967 296 and other powers of 2. en.wikipedia.org/wiki/Orders_of_magnitude_(numbers)#109 - more power of 2 stuff en.wikipedia.org/wiki/Quadruple-precision_floating-point_format - 4 byte fpp format en.wikipedia.org/wiki/Octuple-precision_floating-point_format - 8 byte fpp format THE LAST DAY of a MINECRAFT WORLD?! - minecraft time breaks finalforeach.itch.io/cosmic-reach - cosmic reach download music: PELAGIC - density & time dirt rhodes - kevin macleod prologue - brendon moeller total drag - c418 cliffside hinson - c418 a slow dream - emily a. sprague beyond space - chill carrier MODBAP - density & time
Oh hey, time for Y2K Part 2 in 14 years, hope I'll still be around by then to see all the scammy end of the world stuff people will try to peddle to old people.
Fun fact: 2^1024 is a big thing in an idle game called antimatter dimensions. One you reach that much antimatter, the game forces you to prestige and gain infinity points. You then after some time break infinity and get way more antimatter
i've also played way too many incremental/idle games and apparently 2^1024 is quite common in alot of other games as well in fact its a very special number in general cause it's the max number javascript (the language most browser idle games use) can normally reach, and the most common libraries used to get around this are quite fittingly named "breakInfinity" & "breakEternity"
@@jazzygiabao5560 to add to this, the reason it's very common is becuase it's the max value of 64 bit float (also known as a double), which is the only number format in javascript by default (as noted from jazz, you can use libraries to add custom formats)
32 bit devices can use 64 bit numbers, just not in a single operation. So if the software is fixed (pretty much every open source libc has been for years), they'll be fine. If the software isn't fixed, they'll overflow to 1901, which may or may not be a fatal error depending on how the software needs to use timestamps.
Doing the math, that's a very reasonable estimate. I think the formula would be [(2^33)-1]/60 for the amount of time in seconds. Correct me if I'm wrong.
@@nootnootboopssnoots9707 the figure of ~4.5 years is based on a stryder7x video. the specifics are that the counter is an unsigned 32 bit(4 byte) integer that goes up by 1 per frame 30 times per second.
@@tristantheoofer2 it's unfathomably densely incorrect. By hundreds of billions of years. 64-bit unix time allows to represent time until year 292277026596, December 4th 14:29:40 (enough time to forget about the problem entirely since human civilization will probably collapse sooner, or at least have unrecognizably different computer systems from current ones at that point, if any). This corresponds to a maximum integer increase of 4294967298x, while your estimate assumes a 64-bit integer can only store ~4.3x larger of a number than a 32-bit one.
I absolutely despise learning how to code, but videos like this that cover computer science principals and the significance of certain numbers in computing will always fascinate me. Thanks for taking the time to explain everything as well as you did, it's always nice to learn something new.
For the float/double -> inf/nan issues, there are a really common example I can think of: idle/incremental games. A lot of the smaller ones that don’t use libraries like exptantanum and have broken final upgrades will let you reach 1e308 at the end of the game, breaking everything. This is so widespread that it has been memed on in other games, such as Antimatter Dimensions where reaching 1e308 is when you unlock the “break infinity” mechanic.
This also affected UA-cam when Gangnam Style reached over 2 billion views in 2014; the view count was a 32-bit integer back then, so once the view count reached THAT number, it became frozen. UA-cam had to update the view count to a 64-bit integer as a result.
i think briefly it even rolled into the negatives lol. but yea i know that happened and have for a while. i really think they shoulda just made it 64 bit to begin with yk?
i recently found your channel and so far it's been nothing but bangers i absolutely love this kind of content, will def be looking forward to anything else you make ^_^
In the now discontinued game, Angry Birds Epic, the events that occurred in the game used the Unix timestamp to calculate when a programmed event will occur. Sadly this now means that without modding it, it is unable to play said events, even though they are programmed in the game, however sometime around the 22nd century, long after the Unix timestamp has overflowed and has even rolled back into the positives, the events will start working once again just as nothing had happened up until they pass again.
12:43 the reason things still work at nan in ROBLOX is cuz when your character is at nan for whatever reason the server takes network ownership (aka the server takes control of the physics of your character) and puts your character back to where it was before nan and buggy lighting is just when the camera goes to nan
@tristantheoofer2 quite the fast reply lol, anyway I saw you play roblox? I had a fun video idea where perhaps you go over roblox glitches like emote clips, speed glitches and other miscellaneous things and explain how they work? I know quite a lot of glitches and know how most of them work but can't really find anyone expressing that knowledge with others properly. I would do something similar myself but I have absolutely NO experience in editing or video creating.
@fefe5237 oh actually hell yeah thats a really good idea. i might cover that in a video i make about tiered obbies at some point (will be called "the most unknown roblox obbies" or something) or i might make a seperate video on them entirely
@tristantheoofer2 sick dude! If you need to you can reach out and I could help you with some of the glitches if you do decide to take that idea up, there's a lot more glitches than you think lol
12:27 Funny thing theres a game that uses its own number system internally to bypass this its called antimatter dimensions(it literally makes fun of 10^308 being infinity) I think it did convert to using the bignum system after a while
@@HHHjb_ mobile has gotten the update (everything past eternity is now available on mobile) the update came out on april 1st 2024, with bug fixes coming out as well
yknow i never expected for someone to be at the same community as roblox and minecraft considering their communities are usually enemies, although i play both and it depends on my mood, its a weird sight to see
wait the communities tend to be enemies?? how??? ive played both games for as long as i could remember (age 4 for minecraft and 6 for roblox) and ive never heard anything about this
@@CuongTruong-cb3mk In the game Balatro, if you reach the 39th stage in Endless Mode (Ante 39), the scoring requirements get so high that they exceed the 32-bit floating point limit of 2^1024, causing them to become nan. This means that Endless Mode currently does have an end - and that’s Ante 39.
@@thanhtamnguyenthi7073You know, the most known use of the number 525 is the approximate number of votes (in millions) casted in the 1964 Illinois at-large House of Representitives election, in which 525 million votes were caster because the redistricting process had not resulted in a map yet.
8:48 I think that happens because games calculate high scores like this If current_score > high_score: high_score = current_score Since current score overflows and goes back to the negatives, it's never higher than the 32bit limit
Good video. I subscribed from the last limit video you made. This one also is super entertaining and intriguing. For someone so young, you make amazing content. Thanks.
8:03 Unix time is always seconds since 1970, no matter how long (how many bits) it is. It still "counts up by one", there's no reason for it to do otherwise. So, signed 64-bit time works up to year 292277026596 AD. You are probably thinking of the so-called "Unix nanos", which is Unix time, but nanoseconds since 1970 instead of seconds. That would mean it "counts up by one" per nanosecond, which yes, means that after a second it would have "counted up by 1 000 000 000", but it has many immediate states between 0 and that. Moreover, the length of the time value is not defined, and whether it's 32-bit or 64-bit depends on the OS, not the CPU. Due to _reasons_ the length the CPU uses is often used for values when programming, so that's where the misconception comes from. On Linux (which is the most important Unix OS today), 32-bit devices indeed long used 32-bit time. Support for 64-bit time on 32-bit devices was added to both the kernel and all leading libc libraries (IIRC). However, this change requires breaking compatibility with old 32-bit software that relies on 32-bit time - and the decision is up to whoever builds the OS (usually this is the distribution, like Ubuntu). Debian, which I use, has decided to not support 64-bit time on IA-32 CPUs (32-bit x86, what older "regular PCs" use) to not break compatibility (such CPUs probably will not be used at all anymore in 2038, but the support once again for _reasons_ is required for 32-bit software to work on x86_64 CPUs), but decided it is worth it on other 32-bit CPUs like 32-bit ARM (as those chips are still commonly used, and don't have as much binary-only software). The other Unix OS, macOS, will before 2038 manage to not only completely leave IA-32 hardware support, but also x86 hardware support entirely ;)
oh shit wow. and yeah mac os is really fucking wierd in that regard. actually on apple devices you can set your time up to 2068 but itll still track past that point
@@tristantheoofer2 Interesting... I got a bit more into it, and found that macOS internally uses an epoch separate from the Unix epoch, that being 2001-01-01 00:00:00 GMT. That means the last moment possible to represent there would be, assuming a 32-bit value, 2069-01-19 03:14:07 GMT. This is just a year after 2068, so seems like they just made an arbitrary decision here, so the clock can still run for a year after that. Funnily enough, looks like macOS returns the time since it's (2001) epoch not as a 32-bit integer, not as a 64-bit integer, but as a 64-bit float. I assume 2069 comes from the fact they use the value as a 32-bit integer in some places. macOS, as an Unix OS, still has to convert the time to a Unix (1970) epoch for software written "for Unix", and not "for macOS". If they convert it to a 32-bit value, some things would still freak out between 2038 and 2069, like the `date` text utility. I imagine they use a 64-bit value, but I wouldn't bet on it at this point... and I don't have a way to check it, as I don't have a Mac. This situation also reminded me of something I observed on Linux someday. You can only set the clock up to 2232-04-18 23:47:15 UTC, even with 64-bit time, but it will continue to work from that point. Turns out, the Linux kernel internally keeps time as the mentioned Unix nanos! That means we can indeed observe 2262-04-11 23:47:16 UTC as the last possible timestamp. But wait... 2232 is 30 years before 2262... Yeah, similarly to macOS, they decided to cut it out before the actual last moment possible to represent, but 30 years instead of just 1 year. They justify this by that, since Linux is big in the server world, some servers might stay up for a very long time, and 30 years seemed reasonable.
@GrzesiekJedenastka OH YEAH ok that makes perfect sense. infact its actually exactly 1 year and 12 hours off of the apple os unix limit time. its kinda cool
Dude keep up the work these type of videos are realy interesting and helps other people learn about these cool things fr ur the best if I could have I would have told about u to my friends but they dont know english so cant sadly :(
I will say this number is absolutely broken in every game pretty much. It’s so funny how everything basically breaks at this one number. Anyways completely unrelated nice jtoh skills. My best is a remorseless.
if you go to the game tower creator or tc for short you can use a zipline that just breaks if you stay in too long and then you just go flying and the freeze for a moment. but if you then jump you get flinged really high. 4:01
I were casually watching this video and then you talked about a Roblox video, that’s when I realized that you were the person that made the Roblox zone video that I commented on. I remember talking about playing walk on walls and falling in the void.
aw damn if you missed the max roblox games one then you should totally watch that. fair warning tho that i accidentally forgot to explain a few things but thats about it
5:21 It just occurred to me that Mario 64 uses a mix of integers and floats, while the integers just wrap around causing the world to loop infinitely the floats on the other hand are still floats, this means that Mario 64 also should have some sort of similar kind of bug at play when you get far enough.
@@tristantheoofer2 Interesting! I'd like to see some of the glitches that happen there, of course I know that the collision on the world will just loop infinitely without error, but I'm curious how the graphics will be rendered for Mario when he gets too far from spawn.
so, if i manage to age to 2147483647 years old, then, i would start aging backwards, until i get to zero years old? or would i turn into a freaking wall of 128 meters of height?
crab champions also has a 32 bit integer funny when it comes to its money, in which if it goes beyond the 32bit integer limit, anything you buy in the shop becomes free.
another two games that i find seriously interesting when it comes to the 32 bit limit is cod world at war and black ops 1, especially zombies mode. seeing how the game breaks and you have an insane time limit until you hit a reset has made for some of the most ridiculous speedruns ive ever seen
8:56 its because it cant go above that as it would cause memory errors as it cant write or even read from its area on memory because that memory doesn't exist as its the 32bit integer limit and going above it causes 1. An overflow 2. Underflow (depending on circumstances) 3. A crash (most likely) 4. Write some random shit to melory and corrupt random memory sectors (this will happen if the program doesn't have safemem or memprotection) TLDR: it cant use any memoty lazy there because shit flips the fuck
@@tristantheoofer2 fun fact the pendulum goes gradually faster on mario 64 and will eventually cause a crash for this exact reason (it takes like 24 years though)
As a programmer, I'm reeallly interested in everything related to computers, especially stuff relating to binary and transistors (if you recall I commented on your "How many possible Roblox games" post, the SROSUs and shit lol) so these videos are great in my opinion. I advise everybody to learn about stuff like this and how computers work; it really is amazing to know what's actually going on at the sight of a bug, etc. Furthermore I just wanted to say I appreciate the extreme amount of effort you put into these videos. For example, mentioning Perlin noise per block to explain how the Farlands coordinates correlate to the 32 bit integer limit. Most videos dedicated to the Farlands not go into that much detail. And that's just a small part of this video. I don't know how to phrase this other than it sounding like an insult (trust me, it's not), but this video took 2 months to make. Another UA-camr could make this in 2 weeks.. (bear with me) HOWEVER it wouldn't nearly be as well-researched, well-edited, and overall well-presented. Choosing quality-over-quantity isn't rewarded nearly as much as it should be on UA-cam; this is why I admire UA-camrs like you, Kurzgesagt, LEMMiNO, etc. Because you choose the approach that, instead of benefiting you, benefits your audience. Which is just a great quality. I'm only subscribed to 7 channels over 4 years, and you're one of them. Lastly, I also have ADHD, and I know how hard it is to make something like this. Even with the fact that I know video editing, I could not get close to the skill level needed to have the motivation to make this with ADHD. That's enough yapping!
@@tristantheoofer2 for reference, that's 2^24 I'm guessing that you might be referring to 24 bit RGB (as in, each colour gets 8 bits (leaving 255 values for each colour)) this means that there are 2^24 (16777216) possible colours in that colour space
the reason of the character glitching in roblox is not because of integer overflow, most 3d games use floating points to store the position of every object, the character breaks because the numbers get less and less accurate as they get bigger
I was at first confused as to why game consoles have more memory than computers but now I see why “to prevent these kinds of bugs and improve the overall experience”
eh i think its just cus on consoles you download a LOT more games than on pc typically, so they need more memory (eg 1 terabyte) to compensate for that
and with the bits it’s the same thing the PS2 and Nintendo 64 are 1st to be 64 bit but computers stayed 32 bit only until late 2006 and the Phones and devices went to 64 bit in 2015 and technology is only now becoming 64 bit only
12:08 Ooooooo That's why Balatro score is capped at that, it makes a ton of sense. I wish there was a possible way to increase it, but maybe the only way is a 128 bit system.
fun fact for work at pizza place in roblox: if you sit on a sit-able objects like chairs etc, and then equip the gravity coil and hold jump, you'll be propelled up the world with such a super speed that it doesn't take even a minute for you to notice the precision of the rendering becoming less precise and more "pixalated"
Hey Tristian, I have a question, what is the music used at 8:39, I've been looking forever and i can't find it, it sounds like a minecraft song in reverse
For sdome reason every video has been recommended to from your chaneel and on every video i click i reallise in the middle of the video that its that far lands guy
thank you for WAITING 2 MONTHS for a video lol. i had to finish up school for 3 whole weeks then make this thing. so yeah. anyways sources and music are below as always
sources:
en.wikipedia.org/wiki/Unix_time - wikipedia article on unix time
en.wikipedia.org/wiki/2,147,483,647 - 2 147 483 647
en.wikipedia.org/wiki/Power_of_two#4294967296 -
4 294 967 296 and other powers of 2.
en.wikipedia.org/wiki/Orders_of_magnitude_(numbers)#109 - more power of 2 stuff
en.wikipedia.org/wiki/Quadruple-precision_floating-point_format - 4 byte fpp format
en.wikipedia.org/wiki/Octuple-precision_floating-point_format - 8 byte fpp format
THE LAST DAY of a MINECRAFT WORLD?! - minecraft time breaks
finalforeach.itch.io/cosmic-reach - cosmic reach download
music:
PELAGIC - density & time
dirt rhodes - kevin macleod
prologue - brendon moeller
total drag - c418
cliffside hinson - c418
a slow dream - emily a. sprague
beyond space - chill carrier
MODBAP - density & time
Hey, but what about the 64 bit integer limit
Almost nothing can process above that
I didnt even notice i already subscribed to you, I dont know but i love your videos I find them interesting in a strong sense
Also at 1:23 you could have taken an invisible shortcut i know this cause I'm a JToH veteran
@@rainhat1 fr
@@rainhat1 fire lmao
hearing "2038" and "in 14 years" in the same sentence threw me for a LOOP. how on earth are we further from 2010 than we are from 2038...
wait holy sh 2010 is 14 years ago? damn I was born 2011 and I feel something for this even tho I didn't exist in 2010
@@HeavenlyNovaemy birthday was today and I was like "wait i'm 14 what. "
@@realcinnamongoldi became 14 5 months ago and i was not confused
Cause 2010 was 14 and a half years ago
@@realcinnamongold happy birthday
If you look at the year then we are at the middle of 2010 and 2038, but if you also take account of the current day then yeah you're right
kids nowadays will unfortunately never get to understand the horrors of 65535
32767 and 65536 moment
And 256
As a fallout player, I do. Trust me… we know…
As a minecraft bedrock player, 16777216
255 moment
Games when an integer goes over 2,147,483,647: you are now negative!
Endless stairwell: you haven't seen ANYTHING.
me when my dart monkey has negative xp
That makes absolutely sense.
Because we count sign as a number.
Coughs* warframe* coughs*
BigInt chillin in the corner
mmmmm i love standing still for 524288 and become invincible to crystal spikes
is this a gd moment orrrrrr what
@@tristantheoofer2 Celeste I believe. Look up Spinner Stunning for more on that.
wait, is this the Celeste bug about waiting for so long the spikes can’t properly load?
@@tristantheoofer2celeste thing
Oh hey, time for Y2K Part 2 in 14 years, hope I'll still be around by then to see all the scammy end of the world stuff people will try to peddle to old people.
same. ill be 30 :P
by 2038 I'd be 27 (if april ofc, but 32-bit unix time ends in jan of 2038 so I'd be 26)
@@HeavenlyNovae13 + 14 = 27
I'll be 32 in Jan 2038
Y2K2
You could also argue that 37 is more broken, because of its use outside of computer science, in probability.
Now, THAT is more interesting!
wait seriously? how?
@@tristantheoofer2a roulette wheel (in europe) has 37 numbers on it and is commonly used as an example of your chances of winning.
@@tristantheoofer2 Watch the veritasium video about it
@larrywhitney ok
Fun fact: 2^1024 is a big thing in an idle game called antimatter dimensions. One you reach that much antimatter, the game forces you to prestige and gain infinity points. You then after some time break infinity and get way more antimatter
i've also played way too many incremental/idle games and apparently 2^1024 is quite common in alot of other games as well
in fact its a very special number in general cause it's the max number javascript (the language most browser idle games use) can normally reach, and the most common libraries used to get around this are quite fittingly named "breakInfinity" & "breakEternity"
@@jazzygiabao5560 to add to this, the reason it's very common is becuase it's the max value of 64 bit float (also known as a double), which is the only number format in javascript by default (as noted from jazz, you can use libraries to add custom formats)
many incremental games especially on roblox for whatever reason can go up to 2^^^^1024
@@noobiii what's 2^^^^1024????
@@Revoltition a lot....................
7:49 So if that date passes, there is no way to revert back to its normal state, and the 32-bit devices will be in the history book, right?
32bit devices will roll back to 1901 and that could break some applications
So let me get this straight all 32-bit devices on that day will be back to 1901/1971 devices?
@@Girl-SuS-30001901/1970... tho the 32 unix timestamp was likely a standard in 1971
Unless its a signed value which cant overflow @@Mizu2023
32 bit devices can use 64 bit numbers, just not in a single operation.
So if the software is fixed (pretty much every open source libc has been for years), they'll be fine. If the software isn't fixed, they'll overflow to 1901, which may or may not be a fatal error depending on how the software needs to use timestamps.
Never thought a video about a single number would be this well documented and put together, thank you for making this
np my dude ^^
there's multiple numbers
baking a cake for ~4.5 years is perfectly acceptable in paper mario 64.
Doing the math, that's a very reasonable estimate. I think the formula would be [(2^33)-1]/60 for the amount of time in seconds. Correct me if I'm wrong.
@@nootnootboopssnoots9707 the figure of ~4.5 years is based on a stryder7x video. the specifics are that the counter is an unsigned 32 bit(4 byte) integer that goes up by 1 per frame 30 times per second.
8:20 dude just predicted with an unfathomably dense presition the exact date and time when our devices are going to stop working
yes that is correct
@@tristantheoofer2 it's unfathomably densely incorrect. By hundreds of billions of years. 64-bit unix time allows to represent time until year 292277026596, December 4th 14:29:40 (enough time to forget about the problem entirely since human civilization will probably collapse sooner, or at least have unrecognizably different computer systems from current ones at that point, if any).
This corresponds to a maximum integer increase of 4294967298x, while your estimate assumes a 64-bit integer can only store ~4.3x larger of a number than a 32-bit one.
would be correct if we were talking about like.. 34.1-bit integers (??)
he didn't predict it? also it would be easy to find out
Not really. Unix Time doesn't account for leap seconds and leap seconds goes both positive and negative. The dates mentioned won't be very precise.
I absolutely despise learning how to code, but videos like this that cover computer science principals and the significance of certain numbers in computing will always fascinate me. Thanks for taking the time to explain everything as well as you did, it's always nice to learn something new.
np dude ^^
coding is fun tho
Oh hell yea, another video about stupidly large numbers breaking stuff! Always a treat to watch! Oh yeah, here's 15 ringgits for ya! Keep it up!
no way holy shit fuck yeah thank you for the donation my dude ^^
hey, that's my favorite number
Same
Are you broken too?
me too
no way :o
@@tristantheoofer2 Firwst
Me watching this video knowing damn well I don’t know jack shit about math
Actually its more of programming and computer science than math.
@maxwell6881 computer science is literally math
@@tristantheoofer2 According to the dewey decimal system, they are different subjects, cs at 000, and math at 510.
@maxwell6881 ok but cs relies a lot on math
@@tristantheoofer2 fair
For the float/double -> inf/nan issues, there are a really common example I can think of: idle/incremental games. A lot of the smaller ones that don’t use libraries like exptantanum and have broken final upgrades will let you reach 1e308 at the end of the game, breaking everything. This is so widespread that it has been memed on in other games, such as Antimatter Dimensions where reaching 1e308 is when you unlock the “break infinity” mechanic.
"reaching 1e308 is when you unlock break infinity"
idk if this dude ever played antimatter dimensions
2,147,483,647 is my favorite number due to how fun it is to mess around with due to how hilariously broken things can get.
HOLY SHIT DUDE TYSM FOR THE SUPER THANKS ^^
also same lol 2 147 483 647 is my favorite number as well probably
This also affected UA-cam when Gangnam Style reached over 2 billion views in 2014; the view count was a 32-bit integer back then, so once the view count reached THAT number, it became frozen. UA-cam had to update the view count to a 64-bit integer as a result.
i think briefly it even rolled into the negatives lol. but yea i know that happened and have for a while. i really think they shoulda just made it 64 bit to begin with yk?
@@tristantheoofer2 the creators of UA-cam thought a video would never reach 2bi views, it's too much
i recently found your channel and so far it's been nothing but bangers
i absolutely love this kind of content, will def be looking forward to anything else you make ^_^
ty ^^
In the now discontinued game, Angry Birds Epic, the events that occurred in the game used the Unix timestamp to calculate when a programmed event will occur. Sadly this now means that without modding it, it is unable to play said events, even though they are programmed in the game, however sometime around the 22nd century, long after the Unix timestamp has overflowed and has even rolled back into the positives, the events will start working once again just as nothing had happened up until they pass again.
12:43 the reason things still work at nan in ROBLOX is cuz when your character is at nan for whatever reason the server takes network ownership (aka the server takes control of the physics of your character) and puts your character back to where it was before nan and buggy lighting is just when the camera goes to nan
8:22 we're probably gonna have 128 bit computers by then
we already can make 128 bit cpus its just isnt useful
u mean 1024 bit
@@CuongTruong-cb3mk maybe 512 bits or 256 bits is more correct
I love breaking limits and stupendously large numbers, and your content is the definition of it. Great work as always 👍
really cool video explaining floating point precision, integer limits and giving examples of it!! make more dude!
lmao dw i will. at this point my channel
is becoming about all that and a lot of cs topics lol
@tristantheoofer2 quite the fast reply lol, anyway I saw you play roblox? I had a fun video idea where perhaps you go over roblox glitches like emote clips, speed glitches and other miscellaneous things and explain how they work? I know quite a lot of glitches and know how most of them work but can't really find anyone expressing that knowledge with others properly. I would do something similar myself but I have absolutely NO experience in editing or video creating.
@fefe5237 oh actually hell yeah thats a really good idea. i might cover that in a video i make about tiered obbies at some point (will be called "the most unknown roblox obbies" or something) or i might make a seperate video on them entirely
@tristantheoofer2 sick dude! If you need to you can reach out and I could help you with some of the glitches if you do decide to take that idea up, there's a lot more glitches than you think lol
Man, I freaking love these videos. They always hit different, and the quality is off the charts. YOUR STILL UNDERRATED THOUGH!!
The real ones will remember the og broken number, 256.
ah hell yeah lol
also known as 0
Finally, i can see the 32-bit integer limit clearly for more that 0.25 seconds
We got video entirely made for a single number before GTA 6 💀
lmao did you really thibk gta6 was gonna be out before that point???
Have you... never heard of Numberphile?
this is basically just watching someone info dump why is it so entertaining ToT
being real arent all my videos? lol
ok now look I know you shouldn't just go around calling people autistic but look how easy they make it ToT
@sgtmajorkiwi ngl this is actually kinda accurate
This is some of the most high quality content relating to large numbers I've seen. Keep up the great work brother.
12:27 Funny thing theres a game that uses its own number system internally to bypass this its called antimatter dimensions(it literally makes fun of 10^308 being infinity) I think it did convert to using the bignum system after a while
literally exactly what I was thinking of when he said that.
@@PiercyPlayz I best the game 5 times before the update(I havnt checked if mobile has it yet but if it does bigger numbers be apon us)
@@HHHjb_ mobile has gotten the update (everything past eternity is now available on mobile)
the update came out on april 1st 2024, with bug fixes coming out as well
@@aceae4210 Yooo, reinstalling on mobile rn
very high quality video, i just discovered your channel and by judging from the start, i liked it already. keep it up ;)
yknow i never expected for someone to be at the same community as roblox and minecraft considering their communities are usually enemies, although i play both and it depends on my mood, its a weird sight to see
wait the communities tend to be enemies?? how??? ive played both games for as long as i could remember (age 4 for minecraft and 6 for roblox) and ive never heard anything about this
@@tristantheoofer2 same
you sure u arent talking about battlefield vs cod?
I love that you used jtoh footage in the background
12:16 the most well known example of the 2^1024 limit at the moment is probably the Balatro Ante 39 Nan-e-inf glitch
and what does this mean
@@CuongTruong-cb3mk In the game Balatro, if you reach the 39th stage in Endless Mode (Ante 39), the scoring requirements get so high that they exceed the 32-bit floating point limit of 2^1024, causing them to become nan. This means that Endless Mode currently does have an end - and that’s Ante 39.
"the most well known example of the 2^1024 limit is" proceeds to show like the least popular use
@@thanhtamnguyenthi7073You know, the most known use of the number 525 is the approximate number of votes (in millions) casted in the 1964 Illinois at-large House of Representitives election, in which 525 million votes were caster because the redistricting process had not resulted in a map yet.
Slight misconception: in 9:37, note blocks tuned was shown twice.
oh shit sorry lol
@@tristantheoofer2 is ok lul
8:48 I think that happens because games calculate high scores like this
If current_score > high_score:
high_score = current_score
Since current score overflows and goes back to the negatives, it's never higher than the 32bit limit
but in osu it DOES overflow. theres no way the score would change if it tracked it like that
billion, 147 million, 483 thousand, 647. That's how many views I wish you had lol. Thank you for making this amazing content :)
genius
lmao tysm for the super thanks ^^ (also nice way of combining it with the comment lol)
Ah yes, TOAST background footage
jtoh is love jtoh is life
being real its probs the best background footage that there is lets be honest lmao
@@tristantheoofer2Jtoh is an amazing game to be fair
@@Lilpodo agreed
Ah yes, video.
No no, its a photo
@@TheDucky16 this is the most broken number that exists.mp4
nah i think its a txt file
@@turkey_sandwhich nah, its a exe file!
@@TheDucky16Nah, gotta be a .wav
Ayyyyy, fellow enby
Great video! I'm not the best at math and know very little about coding, and yet I recognize this number anywhere xP
hell yeah ty dude ^^
this video was conveniently watched a day before i had to explain someone what that number meant
Good video. I subscribed from the last limit video you made. This one also is super entertaining and intriguing. For someone so young, you make amazing content. Thanks.
ty ^^ ive been enthralled in this stuff since i was like 10 and wanted to know everything about it lol so yeah
8:03 Unix time is always seconds since 1970, no matter how long (how many bits) it is. It still "counts up by one", there's no reason for it to do otherwise. So, signed 64-bit time works up to year 292277026596 AD.
You are probably thinking of the so-called "Unix nanos", which is Unix time, but nanoseconds since 1970 instead of seconds. That would mean it "counts up by one" per nanosecond, which yes, means that after a second it would have "counted up by 1 000 000 000", but it has many immediate states between 0 and that.
Moreover, the length of the time value is not defined, and whether it's 32-bit or 64-bit depends on the OS, not the CPU. Due to _reasons_ the length the CPU uses is often used for values when programming, so that's where the misconception comes from.
On Linux (which is the most important Unix OS today), 32-bit devices indeed long used 32-bit time. Support for 64-bit time on 32-bit devices was added to both the kernel and all leading libc libraries (IIRC). However, this change requires breaking compatibility with old 32-bit software that relies on 32-bit time - and the decision is up to whoever builds the OS (usually this is the distribution, like Ubuntu).
Debian, which I use, has decided to not support 64-bit time on IA-32 CPUs (32-bit x86, what older "regular PCs" use) to not break compatibility (such CPUs probably will not be used at all anymore in 2038, but the support once again for _reasons_ is required for 32-bit software to work on x86_64 CPUs), but decided it is worth it on other 32-bit CPUs like 32-bit ARM (as those chips are still commonly used, and don't have as much binary-only software).
The other Unix OS, macOS, will before 2038 manage to not only completely leave IA-32 hardware support, but also x86 hardware support entirely ;)
oh shit wow. and yeah mac os is really fucking wierd in that regard. actually on apple devices you can set your time up to 2068 but itll still track past that point
@@tristantheoofer2 Interesting... I got a bit more into it, and found that macOS internally uses an epoch separate from the Unix epoch, that being 2001-01-01 00:00:00 GMT. That means the last moment possible to represent there would be, assuming a 32-bit value, 2069-01-19 03:14:07 GMT. This is just a year after 2068, so seems like they just made an arbitrary decision here, so the clock can still run for a year after that.
Funnily enough, looks like macOS returns the time since it's (2001) epoch not as a 32-bit integer, not as a 64-bit integer, but as a 64-bit float. I assume 2069 comes from the fact they use the value as a 32-bit integer in some places.
macOS, as an Unix OS, still has to convert the time to a Unix (1970) epoch for software written "for Unix", and not "for macOS". If they convert it to a 32-bit value, some things would still freak out between 2038 and 2069, like the `date` text utility. I imagine they use a 64-bit value, but I wouldn't bet on it at this point... and I don't have a way to check it, as I don't have a Mac.
This situation also reminded me of something I observed on Linux someday. You can only set the clock up to 2232-04-18 23:47:15 UTC, even with 64-bit time, but it will continue to work from that point. Turns out, the Linux kernel internally keeps time as the mentioned Unix nanos! That means we can indeed observe 2262-04-11 23:47:16 UTC as the last possible timestamp. But wait... 2232 is 30 years before 2262... Yeah, similarly to macOS, they decided to cut it out before the actual last moment possible to represent, but 30 years instead of just 1 year. They justify this by that, since Linux is big in the server world, some servers might stay up for a very long time, and 30 years seemed reasonable.
@GrzesiekJedenastka OH YEAH ok that makes perfect sense. infact its actually exactly 1 year and 12 hours off of the apple os unix limit time. its kinda cool
I knew about all of this previously, but thanks for making this video so I can show this to people that don't know!
Dude keep up the work these type of videos are realy interesting and helps other people learn about these cool things fr ur the best if I could have I would have told about u to my friends but they dont know english so cant sadly :(
aw thats sad abt your friends dude. rip. also ty ^^
i love numbers
Noticed those xor gates at the end of that roller coaster track at the end there, cool
oh yeah i was briefly into redstone logic so i made those
I will say this number is absolutely broken in every game pretty much. It’s so funny how everything basically breaks at this one number. Anyways completely unrelated nice jtoh skills. My best is a remorseless.
cant be a day of youtube without a well edited video about the most random topic ever !!
if you go to the game tower creator or tc for short you can use a zipline that just breaks if you stay in too long and then you just go flying and the freeze for a moment. but if you then jump you get flinged really high. 4:01
integer overflow jank, my beloved
I were casually watching this video and then you talked about a Roblox video, that’s when I realized that you were the person that made the Roblox zone video that I commented on. I remember talking about playing walk on walls and falling in the void.
oh fire. did you enjoy this video btw? bc it blew up so fast that in 6 hours it has 3.5k views almost
@@tristantheoofer2 yeah I loved the video, just like the floating point and Minecraft void ones (I might’ve missed the max Roblox games one though)
aw damn if you missed the max roblox games one then you should totally watch that. fair warning tho that i accidentally forgot to explain a few things but thats about it
5:21 It just occurred to me that Mario 64 uses a mix of integers and floats, while the integers just wrap around causing the world to loop infinitely the floats on the other hand are still floats, this means that Mario 64 also should have some sort of similar kind of bug at play when you get far enough.
yep he does actually and with sblj's and stuff you can yeet him all the way to infinity and crash the game
@@tristantheoofer2 Interesting! I'd like to see some of the glitches that happen there, of course I know that the collision on the world will just loop infinitely without error, but I'm curious how the graphics will be rendered for Mario when he gets too far from spawn.
theres actually a video by pannenkoek2012 showing this off afaik
@@tristantheoofer2 I can't find it, can you send a link?
How did bro fall at 1:06 that's a skill issue
lmao idk man im trash :P
@@tristantheoofer2 it was a skill issue, you're not trash >:[
oh hey wtif
Hey guys it’s me final impact from windows ultimate world
Wtif
bro i looked into ur channel and immediately subbed cuz wtf u post every type of content i find interesting i’m ur biggest fan now 🙏
so, if i manage to age to 2147483647 years old, then, i would start aging backwards, until i get to zero years old? or would i turn into a freaking wall of 128 meters of height?
you would cease to exist
@@tristantheoofer2 why?
bc 2 147 483 648 years ago you didnt exist
@@tristantheoofer2 sorry, for some reason i dont understand😭, my brain is really small apparently
@@kikospagnoYT lmfaoooo
every time i hear about "2038" this is the only thing that i think of. imo, this is gonna be a huge thing in the whole world LOL
My favorite funny numbers guy uploaded again
ty lmaoooo
computers: see a bunch of numbers together
also computers: *didn't seem number enough to me*
No way, random cosmic reach reference. I love that game. Been playing it since day 1.
lol fire. i got it just for the video when someone i know discovered that glitch
10:13 hardest ring 6 tower:
real!!!
crab champions also has a 32 bit integer funny when it comes to its money, in which if it goes beyond the 32bit integer limit, anything you buy in the shop becomes free.
4:50 I think there probably are better examples than Cosmic Reach since its so similar with Minecraft
true but other games i tried finding stuff for i either couldnt or there was literally nothing about it
another two games that i find seriously interesting when it comes to the 32 bit limit is cod world at war and black ops 1, especially zombies mode. seeing how the game breaks and you have an insane time limit until you hit a reset has made for some of the most ridiculous speedruns ive ever seen
8:56 its because it cant go above that as it would cause memory errors as it cant write or even read from its area on memory because that memory doesn't exist as its the 32bit integer limit and going above it causes
1. An overflow
2. Underflow (depending on circumstances)
3. A crash (most likely)
4. Write some random shit to melory and corrupt random memory sectors (this will happen if the program doesn't have safemem or memprotection)
TLDR: it cant use any memoty lazy there because shit flips the fuck
true. hell in super mario 64 it DOES crash lol
@@tristantheoofer2 fun fact the pendulum goes gradually faster on mario 64 and will eventually cause a crash for this exact reason (it takes like 24 years though)
@@DJ-Dashy yeah ik
As a programmer, I'm reeallly interested in everything related to computers, especially stuff relating to binary and transistors (if you recall I commented on your "How many possible Roblox games" post, the SROSUs and shit lol) so these videos are great in my opinion. I advise everybody to learn about stuff like this and how computers work; it really is amazing to know what's actually going on at the sight of a bug, etc.
Furthermore I just wanted to say I appreciate the extreme amount of effort you put into these videos. For example, mentioning Perlin noise per block to explain how the Farlands coordinates correlate to the 32 bit integer limit. Most videos dedicated to the Farlands not go into that much detail. And that's just a small part of this video. I don't know how to phrase this other than it sounding like an insult (trust me, it's not), but this video took 2 months to make. Another UA-camr could make this in 2 weeks.. (bear with me) HOWEVER it wouldn't nearly be as well-researched, well-edited, and overall well-presented.
Choosing quality-over-quantity isn't rewarded nearly as much as it should be on UA-cam; this is why I admire UA-camrs like you, Kurzgesagt, LEMMiNO, etc. Because you choose the approach that, instead of benefiting you, benefits your audience. Which is just a great quality. I'm only subscribed to 7 channels over 4 years, and you're one of them.
Lastly, I also have ADHD, and I know how hard it is to make something like this. Even with the fact that I know video editing, I could not get close to the skill level needed to have the motivation to make this with ADHD. That's enough yapping!
And nice toie gameplay lol
Nah, 1 minute in and I already learning more than what school taught me.
are you trying to major in computer sciences or coding ?
@@MindOfFish Yeah XD
This feels like I shouldn't understand most of this, but for some reason I do anyway.
9:14 EVERYTHING OF 2^31.
literally saw a post on reddit made by you about this video. good video!
9:54 they probably have no texture for the "-" sign
ok being real i didnt think of that and you might be right
or it just goes to the nearest "point" there (a point being an available number)
The highest number in games is Technical -1 (after that comes 0 and after that 1 and everything starts from the beginning)
7:40 you know that windows calculator has a time conversion calculator built in, right?
uh no??
It can convert between commonly used units of time. Unfortunately, the calculator does not have a Unix time converter.
I discovered that kind of number in HCR1 when I collected many coins and at that point the coins couldn't keep up until it reaches the maximum limit
the 32 bit integer (2147483647) is basically the Stan Lee of video games, constantly showing up everywhere.
same with 16777216 in some instances but yeah lol
@@tristantheoofer2 for reference, that's 2^24
I'm guessing that you might be referring to 24 bit RGB (as in, each colour gets 8 bits (leaving 255 values for each colour))
this means that there are 2^24 (16777216) possible colours in that colour space
@aceae4210 yes. and at 16777216 units away, if you use floats the precision gets so bad its 2 units
0:30 JToH Jumpscare
lmao real
@@tristantheoofer2 do u play jtoh? i also play it i could play wit u lol
@tunashift yeah i do sometimes lolpl
@@tristantheoofer2 could we play sometime together i could help you play it ig
In Australia in 2038, Easter Sunday and Anzac Day will fall on the same day. Will lead to a super long weekend.
12:06 fun fact 179.769 uncentillion is the max cash in AdCap
ik. actually its 10 uncentillion but they capped it there bc if you got to 179.769 uncentillion it rolled to infinity
@@tristantheoofer2 infinite money glitch
Damn, I've waited this kind of video so long!
Thank you for making it! Now i kinda understand why after 16.777.216 I can't get odd numbers anymore
np dude ^^
the reason of the character glitching in roblox is not because of integer overflow, most 3d games use floating points to store the position of every object, the character breaks because the numbers get less and less accurate as they get bigger
ohhh that makes sense
yeah i got that, but i thought it was worth a mention anyway and games out that far are stupidly broken regardless basically no matter what
I liked the JToH gameplay in the background.
lmao ty. thatll become a recurring theme likely
man fails toast gameplay and then resets
@aeiouuioea yes
oh wait theres also short footage of toie
tocn footage too holy fuck
I was at first confused as to why game consoles have more memory than computers but now I see why “to prevent these kinds of bugs and improve the overall experience”
eh i think its just cus on consoles you download a LOT more games than on pc typically, so they need more memory (eg 1 terabyte) to compensate for that
and with the bits it’s the same thing the PS2 and Nintendo 64 are 1st to be 64 bit but computers stayed 32 bit only until late 2006 and the Phones and devices went to 64 bit in 2015 and technology is only now becoming 64 bit only
0:47 insert tpot intro
why
@@tristantheoofer2 why not
@@tristantheoofer2you summoned a community
Duh power of 2... 👹👹
@@Spaceyguy_animates wait does tpot stand for "the power of two"
12:08 Ooooooo
That's why Balatro score is capped at that, it makes a ton of sense.
I wish there was a possible way to increase it, but maybe the only way is a 128 bit system.
3:32 decimal overload
i think that is because of their coding lenguage which propably just spits an error when you try to add something pass the 31 bit limit
1:55 I laughed when I saw "1.71103"
yeah i like making each part a reference to something
Very good content. Nice quality. Keep it going dude
1:47 what in the farlands is this!
4096 blocks tall one
fun fact for work at pizza place in roblox: if you sit on a sit-able objects like chairs etc, and then equip the gravity coil and hold jump, you'll be propelled up the world with such a super speed that it doesn't take even a minute for you to notice the precision of the rendering becoming less precise and more "pixalated"
oh thats pretty cool actually nice lmao
1:11 numbers bases
i do remember a number that makes the sky box pure black in roblox
but it is so hard to say I can’t even say it
isn't that the 64-bit int limit or something (aka 9223372036854775807, yea I have that memorized)
@@HeavenlyNovaeit would definitely be that uh
big number ovo
Hey Tristian, I have a question, what is the music used at 8:39, I've been looking forever and i can't find it, it sounds like a minecraft song in reverse
oh i think i know the one you mean. its called "a slow dream" and its by emily a sprague
@@tristantheoofer2 Yo thank you so much bro!
@white_suit_nickironswordfa208 np np
I got really surprised when you mentioned Cosmic Reach, I've been really loving that game's devlogs lol
lol yeah some dude on my discord server found it and i wanted to include it somewhere
For sdome reason every video has been recommended to from your chaneel and on every video i click i reallise in the middle of the video that its that far lands guy
real
gotta love the integer limit
Short-32767, Integer-2147483647, Long-9.223*10^18, Float-about 3.4*10^38, Double-1.796*10^308
6:30 yeah i know the Y2K38 problem
Fun fact: in geometry dash the attempts dont actualy stop at the 32 bit limit, instead it goes down until it hits -2147483647
amazing
I can never remember the number itself, but when I see it, I KNOW what it is.
That's how much I've seen the highest 32-bit positive integer.