This Is the End of the Silicon Chip, Here’s What’s Next

[@guybonberry5252]

I gave up studying computers ages ago because teachers could never explain how code and transitors worked. Your short video would have solved my dilemma and briefly enabled be to understand the science behind computers. Thank you for educating an ol' man.

[@honkhonk8009]

transistors work is easy to understand. Theyre used to make Logic gates, such as If input 1 is on, then turn output 1 off. and if you put these logic gates together, you could make a small calculator. The actual calculator part i dont understand, but i understand how transistors make up logic gates, since you can search up the schematics for certain not gates

[@kamtroy2527]

If no one could explain it to you than you should have just read a book on it.

[@vivekchauhan4053]

@@kamtroy2527 exactly!!

[@Cryptonat]

I'm pretty sure they only made this video for advertising. They didn't really share much information. I feel like my time was wasted.

[@typingcat]

The only useful data was light computing and they did not even tell much about it. Even I knew about it more than this video explained.

[@ciraxa]

I didn´t watch it and came straight to the comments.

[@sure5291]

exactly, at the end you learned nothing or you gained nothing.

[@ma2i485]

how exciting

[@davemaverick8438]

as he stated in the end that one of the best uses of technology, as his example for amusement, is watching cat videos, while thanking nerds, you cant be more basic than that, so detailed info is secondary here

[@Nico1a5]

Progress got stuck for years until intel got a bit of competition

[@johnkaboly5313]

Well

[@doctor99267]

John Kaboly Earth is flat flat.

[@johnkaboly5313]

doctor99268 stop making fun of my stutter

[@ElTokeMaestro]

Compatition is what makes progress

[@grrr1351]

dAMD right!

[@stevensteven4863]

I am from the future being specific 2021 iPhone is only changing the case :-)

[@padanmuke8109]

😂

[@thefirstsin]

I'm from the year 2021 6/3/21 and 2nm is 4 years away and next year is Intel and amd's 3nm chip release.

[@zen608]

@@thefirstsin I'm from the future and Apple is spending billions of dollars trying to break the 1nm barrier O_O I'm very excited

[@marcog832]

All of you are now from the past

[@beastntenn]

Top tier comment

[@Master_Therion]

I asked this very question to my computer scientist friend, José. "After silicon chips get too small will we need to use a different element, if so which one?" He replied, "Si."

[@Dyslexic-Artist-Theory-on-Time]

That is the one we already use!

[@johnkaboly5313]

Earth is flat flat

[@johnkaboly5313]

Hades Subordinate Function to spread the truth

[@iRxyanDestinygtaandmoreL]

Master Therion im done, now i will be disappointed if i dont see your comment top on EVERY SINGLE SCIENCE VIDE EVER!

[@doctor99267]

Master Therion i don't get it.

[@brosch91]

It will be interesting to see what we do next when we hit the supposed chip shrinkage limit. Maybe then the computing industry can finally focus on optimization instead of chip shrinking. I'm sure current (and even older hardware) could be a lot more capable if optimization was a high priority. I feel our current and maybe even older hardware rarely realizes their full potential because of the lack of proper coding optimizations.

[@MetroidChild]

We've already abandoned typical transistors back around 2010, the switch was to FinFET which has a larger gate area for the relevant physical size, developments on GAAFET will probably extend the limit closer to 3nm. But it's worth mentioning that there are multiple types of transistors around

[@MetroidChild]

The way the video puts it the practical limits of silicon transistors is 5nm, but that only really goes for FinFETs, normal planar MOSFETs really had their last practical run around 20nm (maybe 10nm if we count planar FD-SOI). The same goes the other way too, advances in new transistor shapes (horizontal/vertical GAA-FETs and the like) will help to push beyond the limits of 5nm, probably to around 4-2nm.

[@mantisnomo5984]

Aren't you guys conveniently ignoring the 3rd dimension?

[@MetroidChild]

Depends on what you mean, FinFETs were called "3d transistors" by Intel back when they were introduced due to them judding upwards (and moreso true with modern taller/thinner FinFETs). The whole point of vertical GAA-FETs (or who knows maybe vertical TFETs) is to utilize the third dimension for better efficiency, but this can in some cases lead to _less_ dense transistors which is problematic for some areas. If you mean stacking transistors/dies on top of each other; that doesn't really change the way transistors are made, which is what the video again talks about (or doesn't depending on who you ask).

[@kercchan3307]

heat becomes a major issue with silicon chipsets the smaller the transistors get, luckily material science is slowly working on cheaper better materials

[@Dr_Mario2007]

Gallium Antimonide, Gallium Germanide, Gallium Nitride and Gallium Arsenide transistors are a few options for 3nm and smaller nodes, to list a few.

[@adinameissner2271]

Well, there are also some great potential upsides to the end of moore's law: like our gadgets won't need to be replaced every 6 months and programmers will have to actually start doing a good job since the bloatware hamster wheel will have to stop.

[@denisethasder8193]

We always find a way to improve technology. It’s just a matter of time and resources. Although I can’t say I’ll be paying $10,000 for a new breed of computer

[@RoScFan]

Denise Thasder some things are impossible. Dont forget we cant surpass the speed of light

[@midnightwatchman1]

or the quantum effects become more dominant

[@nltiro3387]

Denise Thasder if it let's me run pubg at medium graphics I will

[@davidbeppler3032]

RoScFan we can not surpass the speed of light....yet.

[@elenaa_28]

it's not possible...at all

[@R3MIXMODZ]

I went on a tour to a huge factory where I lived and they said their one of the biggest manufacturers for phone IC chips. They said throughout all of their careers, they have managed to have 3 IC components in the iPhone build and some with others. Keep in mind they have millions of dollars worth of components only to result with 3 IC Chips. Thats just amazing.

[@mantisnomo5984]

What's amazing is that their careers did not exist before the iPhone. When I worked at Apple, the "competition" between vendors to get their products included in the design of an Apple product was fierce. I lost my job because my boss and my boss's boss took bribes from a vendor to use their inferior part, but I refused to go along. Perhaps integrity is overrated, but I still believe criminals should be punished.

[@williamzhang963]

People made the same prediction for the 10nm node less than a decade ago, saying that would be the end to silicons, now we are 7nm. I read that Samsung is planning 3nm Gate-All-Around FETs in 2021, so maybe 5nm isn't the end.

[@seymoronion8371]

You don't need to worry about the claimed Ryzen CPU (7mn) 5-year burnout, because you can just by a new Intel CPU three years from now. Never you mind that Intel's new CPUs will also be 7nm...

[@doriandodo99]

2:28 CARBON NANOTUBE TRANSISTORS !! - NCIXKeys

[@warmon6]

oh R.I.P NCIX but that saying will live on. lol

[@qboycorvi]

Dorian Tomašinec Searched the comments if anyone picked up on this. Was not disappointed.

[@TheFlacker99]

Julia has a show now with tech news.

[@doriandodo99]

I know im subbed "snippy snippets"

[@Cineenvenordquist]

Cite, my dudes. Show an Advanced Nano Materials we never saw before.

[@somename1324]

Quick FYI, the manufacturing process is 14nm (and shrinking down to 7 soon) and 12 is available. The transistors are actually much larger but we can laser etch the fins down to a width of 7ish nm or 14nm for the chips referenced in this video (First gen AMD 500 GPU's, first gen AMD Ryzen chips, etc.)

[@greypaladin4560]

One neat option I have read about is to incorporate the quantum effects of the sub 5nm structures into the design--either using them directly, or having a correction system built in to the chips. But, that still probably won't get us very far for all the complexity it would introduce. I'm betting after 2021-ish machine learning and optimization will be applied to improve software and chip design for another few years while the industry figures out just where it is going to go.

[@tomnoyb5079]

A single ladder rung of DNA has 18 unique "states" vs silicon's two. DNA is 2nm vs silicon's 5nm theoretical limit. That's 22 times improvement over silicon. Not recommending DNA computing, rather that there are more compact (and thus faster with lower power) structures than silicon.

[@robinsuj]

I believe that DNA storage is a thing that's being seriously considered.

[@UltimatePerfection]

pabloeli29 Yeah, there was serious progress on that front. Prohibitively expensive though.

[@NotSafer]

I understand where you're coming from, but what you have to consider is that the typical computer transistor changes it's state purely with electricity billions of times per second. DNA is an interesting mechanism to store information, but it doesn't have the capability to change and process information, with that amount of electrons running through, it would rapidly degenerate. Each DNA strand merely holds a certain piece of information, so the comparison is not at all valid...

[@tomnoyb5079]

Ricardo - Surely there'd be many more technology challenges than that? Merely laid out the theoretical limit. A factor of 22 is seven-years on the Moore's-law curve. Moore's law originally postulated halving gate-lengths every eighteen-months. In other words, if Moore's law were to extend past 5 nm, DNA could only add seven-years to the curve. Is it possible to develop DNA to the point of mass production within the two-years left on Moore's law? One presumes not. Even if DNA computing were the correct technology direction, it couldn't save Moore's law. Moore's law has been the economic growth engine for the world. When it stops, what happens to economic growth? Most of us have never seen a world without that growth. We rely on it. Now it's over.

[@NotSafer]

Yes, there is certainly a lot more limitations, but I'm just laying out the basic limitations as for why DNA is not a feasible solution, it is simply because DNA does not compute, DNA stores data, it is an interesting factor when talking about storage technology, HDs, NAND flash, etc. But this video is in another subject that DNA plays no role in other than the mere length comparison. And of course, we are getting to the limit of Moore's prediction, we will probably have to move away from transistors to the next big thing in the next few decades.

[@retrobrw919]

Hey Trace. You're talking about FETs, or Field Effect Transistors with the gate/drain/source at the start of this video. Not regular transistors, or what we would call Bi-polar junction transistors in electronics, which have a collector, an emitter, and a base. Although, FET's are what most chips are made with now.

[@Toastmaster_5000]

Yes, and? Processors use FETs (some specifically use FinFETs) and have been for years. Trace isn't wrong.

[@shre6619]

Thanks,i was confused that my book had transistor made of emitter,base,collector ones. I didnt knew about these FET

[@ABaumstumpf]

they are still transistors - and those are affected just the same.

[@denniskarnes8680]

Yoshi Knuckles Exactly FET's and MOSFET's have been around for many years but this guy prob never heard of them.

[@billmyers6096]

Yoshi Knuckles, Yup, you nailed it! Different beasts altogether. And, Trace IS wrong, Peter Schmidt. Trace was historically talking about BJT transistors and never switched gears to a FET. It wouldn't have taken any effort for him to clarify and say "FET" when addressing the Source, Drain, and Gate explanation. Trace is ignorant, else he would have said, "Emitter, Collector, and a Base." This isn't the first time Trace has been wrong, and it won't be the last. It happens to all of us - myself included. However, after offering up a smoke sacrifice to the Electron gods about 40 years ago when I thought I'd just be able to replace a FET with a BJT in an amplifier circuit, I've never confused the two since. Ever!

[@GioGziro95]

0:34: „...and if the gate allows the electrons to pass...“ Seems like you're saying that the gate actually keeps electrons from flowing from the source to the drain which isn't technically correct. What keeps electrons from flowing through the transistor is the electric field between the source and the drain, and the gate helps the electrons to overcome it by attracting them. I'm oversimplifying the process but that's basically it.

[@ashwingiri1510]

I am from Future and can safely say that the new iPhone 15 pro max has 3 nm chip.

[@JoseRamirez-yh2ll]

Just when things are going good. Physics always ends up getting in the way. We always hit that wall! *Damn you physics!!!!!*

[@oari1150]

Physics. Making everyone his bitch since 14 billion B.C.

[@Azknowledgethirsty]

O'Ari 13.8BCE

[@tylerslagel5485]

Yes, and then we always wrecking ball our way through it. Don't worry.

[@40g33k]

Fish hits wall. Dam... He said

[@NocturnalRS]

If you're the driver avoiding eye contact in your commute to work is probably a good thing, otherwise that reference seems random lol.

[@burt591]

What if you keep the same transistor size but just put more of them? The CPU will be bigger, but what's the problem with that? Why do they need to make them smaller to make it faster? I think that's a viable thing to do at least up to 3 times the size form the ones we have now. What am I missing?

[@phillipcarpenter1638]

*burt591* The larger the cpu, the slower and hotter it runs.

[@burt591]

Yeah but still I think 3 times the size from the ones we have is feasible, just needs 3x bigger cooling system. It would be nice to know exactly how the size affects it.

[@dasemmiyogurt6288]

burt591 For Desktop/Tower Pc's it would work but whats with smartphones and laptops

[@burt591]

Good point

[@cortster12]

burt591 It would also cost three times as much, which then you'll just have a super computer that costs a ton. Because when you make them smaller, they use less energy per transitor, thus you get more bang for your buck.

[@Enonymouse_]

Until we acheive or exceed 1-2NM in size there is no point in trying to go beyond silicon based semi-conductors. We've been readily exceeding Moore's law since it's inception.

[@mykies2297]

A transistor has 3 leads base, emitter and collector. A Mosfet has gat, source and drain.

[@jeffo9396]

A MOSFET is still a transistor, though.

[@darkling65]

Temperature also effects the vibration of matter. I'd imagine that if you super-cool the same chip you will get less jumping of electrons.

[@pankajkaurav3155]

In 2021 scientist about to get 0 nm size of transistor Science is really amazing

[@GamerForLife514]

I want Seeker to talk about PC and Console Gaming, and how with more Powerful GPU's, CPU's and APU's will be in the future and Gaming Industry in 10 and 20 years will be huge.

[@nicolefawkes419]

I think a better CPU analogy would be to compare it to an engine, not a brain. Instead of pistons you have electricity. Its job is to crunch numbers... they run, they don't think. The computer is not smart, even with AI or ML - your software does that, the computer simply runs the software.

[@aenigmaticus_ca]

...For now.

[@RonanTetsu]

Nicole Fawkes It's not just software. it takes particular hardware infrastructure. DSPs, the GPU, and CPU are usually combined with special infrastructure to do something. There's some form of thinking, software or not.

[@mantisnomo5984]

The computer doesn't "crunch numbers." It is a symbol manipulator.

[@johnw1385]

You are a machine... Don't belive it get a map of what portion of your brain controls what functions... And get a power with a long drill bit... Write us after drilling your motor corex and let us know how non machine and magical brain is doing. Good luck

[@harleyme3163]

its not a brain... all it does is spit out 1's and 0's this is way its possible to make a cpu that use's light. think about a flashlight turning on off.. leds and a light sensor is = to a single transistor so that why he meant it might be larger then todays cpus

[@MartiniPinball]

2020 and we are already working on sub 3nm chiplets

[@ronniepirtlejr2606]

Why not just make the chips twice the size or quadruple the size and have one giant CPU but keep the gates the same size? More gates bigger CPU.

[@matthewsmith2385]

Ronnie Pirtle Jr facts

[@monkeyrobotsinc.9875]

uhhh they already do that. its called multicore.

[@RoboticSkillz]

Bigger CPUs need more voltage to run. Thus also getting hotter and needing better cooling solutions. That's really all that shrinking nodes is about. Being able to have more transistors in a small package, while keeping power consumption and temperatures low. What I mean is, you can't just quadruple the "CPU size". It'd have alot of problems, but the biggest one by far would be the power consumption.

[@benjaminmcintosh857]

Manufacturing a piece of silicon without defects get VERY expensive once you start making it big (exponentially so). And there are losses associated with using multiple silicon chips for the same processor. Also bigger chip - harder to cool; just look at Threadripper or Xeon Phi

[@niter43]

I don't have a good grip in field/might be wrong, but IMO y'all answering are missing main point. CPUs run in discrete steps by command of clock signal (imagine a metronome idk), so each transistor has to go into it's final state for given step in constant timeframe, before next clock/beat. Transistors relay on electrons, which can't be faster than speed of light. So, at 5ghz electron can only go for c*(1/5x10^9)=6 centimetres before it's late and next clock starts. Now add overhead of that transistors aren't connected in straight lines, that they aren't actually traveling at full speed of light and add some headroom/tolerances. Somewhat-somewhat in the same order of current CPU cores sizes. So, to summarize: big cores - low clock speeds. And probably there is not that much to add (and plenty of designs were tried?) to make core bigger, so when lithography shinks CPU manufacturers don't add more transistors to cores, but go for higher clock speeds. Adding cores is possibility (because they don't need to be perfectly synchronised on each clock?), but only effective for specific tasks / before certain amount of cores.

[@dandandan18]

It's 2024, and we're beginning to manufacture 2nm transistors

[@hoangkimviet8545]

I love the shirt of Trace. Benjamin Franklin and the kite :0

[@Dyslexic-Artist-Theory-on-Time]

Franklin was great!!!

[@johnkaboly5313]

Earth is flat flat

[@DonCDXX]

Benny F was one of the awesome ones.

[@jesusmejia1334]

He was an OG.

[@eriktran9682]

John Kaboly Ok, stop spreading bullshit dude.

[@sludgefactory241]

Well, they are pushing ahead with 3nm processes, so that no go under 5nm kinda got thrown out the window. Tsmc is having problems with the yields but it'll get there

[@TheJan1207]

An hour ago I was watching some videos by Joe Scott about this topic and a couple of minutes later you uploaded this video. I think the matrix is broken again.

[@irfanulkarim4992]

Baumstumpfkopf earth is flat

[@magicweaponr072]

Irfanul Karim no u

[@irfanulkarim4992]

Magic Weapon R0 Flat Flat Flat. Earth is flat and it sun is 6000 miles above. You believe in GPS does that even work? It's all fake. How can there be satellites when earth is flat. Btw Mars and Sun and Moon are round . Earth is flat and disc shaped.

[@johannesleirgul953]

"How can there be satellites when the earth is flat." How can earth be flat if there are satellites?

[@BRAMSTONER]

Irfanul Karim ok lets say earth is flat now what governments r still fucking ppl with no vasalin ice cube was flat

[@EgadsNo]

Don't forget clock speed. How well it handles heat is super important too.

[@okdoomer620]

Uhm... Processor speeds stopped increasing more than 10 years ago, this is not news. But computers are still getting faster though. It's just that you don't get the free speed up you used to get by simply increasing the clock speed. Today you have two or more cores in CPUs, modern GPUs even have thousands of cores, and more and more tasks are solved on GPUs now. You just need to be more clever about it, so software gets written with a parallel or multicore processor architecture in mind. I'm not saying that we won't have other than silicon based processors in the future, but you can get very far just with more cores and parallel programming... And if you don't find anything that's a lot (!) faster, you're going to have parallelism anyways, and in that case it would need to be cheaper and as reliable as silicon... Not going to happen soon...

[@wildhostage]

Whoopsie, seems 5nm wasnt the limit

[@ganopterygon]

we could get rid of windows 10 and increase performance 100x, just saying...

[@clayestes4640]

Aurelio Rockdriguez Windows 10 performance is way better than any of the previous windows...so unless they develop something else I don’t know where you’re gonna go lol

[@timeriderx]

It's not the operating system but the power of the ram, CPU and other factors that help Winblows do it's thing! Layer upon layer of crap programing needs so much power to work!!

[@Gunnareth]

ima keep it real with u chief i am NOT going to use linux for gaming

[@ciraxa]

Steam now supports more than 2.5k games on Linux, just saying... @@Gunnareth Should be usable for gaming in 5-10 years if there isn´t another open source OS by then.

[@PenisMcWhirtar]

Win98 FTW!!!

[@yenaarts]

One thing they have wrong in this video is the size of the transistors. 14-20 nm refers to the resolution that’s used when transferring the template to the silicon wafer when the processor is manufactured, not the gate itself.

[@LordDice1]

Killed the shawshank reference😁

[@OomNeil]

Made me laugh so hard!!

[@erikk77]

He nailed it!

[@aidangillett5396]

We stopped going for clock rate a long time ago - we were doing 3ghz+ in the early 2000's - and started pursuing more data in each clock cycle, and more cores to multiply the amount of work that could be done. We don't NEED smaller transistors we can either use physically much larger cpu's or many more of them. I kind of like the idea of having a motherboard with 32 cpu slots in it and you simply add as many cpu's as you want performance. Of course software development needs to change to be highly multi threaded, this would probably benefit most from a hypervisor that dynamically handles this and spreads the load so it appears to the software as 1 core/thread but is actually running on dozens at once

[@dragoola69x]

Truthfully I'm surprised we haven't used the engineering and technology that goes into to transistors on Silicon Wafers to make small and or Nano batteries that can be layers upon layers to increase energy density tenfold because what is a transistor BUT a p-type anode and n-type cathode battery essentially a bry voltaic pile on a microscopic even Nano scale but what do I know I'm just a dishwasher.

[@zeeeeeeeeeev6493]

what

[@adrianlowery7175]

There's a major reason why some tech doesn't get developed: market interests. If rich people aren't going to make a big enough profit from a tech evolution, that tech evolution won't happen.

[@bilboXbartok]

What you are talking about: "Source - Drain - Gate" is not a regular transistor it is what is called a FET or Field Effect Transistor, which is a quite different mechanism from a regular transistor! This might seem to be a minor point, except that these two devices work in quite different ways!

[@justrhyme123]

the content is so weak. you could just put " ....next is transistors using light or molybdenum sul...." in the title of the video instead of having me watch a 4 min video where the answer to the title is just 30 secs or less without any detail. rest of it is verrry basic info and ads

[@bigtime9597]

Crazy because TSMC (major chip manufacturer for AMD, NVidia and Intel) are already beginning production on 7nm architecture.

[@alamond3318]

"Small is actually good" not so sure about that...

[@irfanulkarim4992]

Stefan The Guy earth is flat.

[@Toastmaster_5000]

[@bfnvalley]

Growers not showers.

[@TheRosemontag]

That’s what she said.

[@dellconagher8121]

It's the tech that matters

[@icarusno]

Just being nitpicky about the physics, but it isn't the current that represents a 1 or 0, but the voltage level. Voltage is what triggers the gate in a transistor to open or close, which is what allow computations to be completed. Yes, current does flow during switching, but a stable voltage level (a stable bit) is what is computed.

[@rohandebbarma24]

This is the first video I've seen from your channel. I must say I got addicted to such an extend I can't describe. In it to learn more and more. Afterall knowledge is power! ☺️ Looking forward to more about Electronics. That's my favourite topic.

[@kineticstar]

Power consumption of a light based cpu and size will need to be made more viable before anything like that can be used. Switching on lazers and light emitting diodes at speeds required will need massive requirements for startup draw and thermal dissipation. Phase discrepancy of the wave guide will also be a headache. Fiber over longer spans is flexible but on distance like a cpu it will be difficult to work with and develop.

[@cameronbeyer7687]

CARBON NANO TUBES

[@irfanulkarim4992]

Cameron Beyer earth is flat

[@stardude2006]

Cameron Beyer Graphene Nano Circuits

[@mrboleus8240]

R&D should focus more on optimisation. Imagine we reached hardware limit of 99% with latest 7nm, while with software optimization we're at roughly 1%. On the other side, we can wait and pray for new, non silicon chip, or even quantum chip taking commercial market.

[@summertyme5748]

Ant man can fix anything that has to do with “quantum”.

[@___xyz___]

0:05 vacuum tube transistors are still superior to anything made today in terms of operating frequency. we just haven't found a way to mass produce vacuum channel transistor based processors yet.

[@dudewat212]

"Quantum Mechanic Valley" doesn't have as nice a ring to it.

[@venkathariharan1515]

Lol, good one!

[@BackFischKuchen]

A closed gate does not not represent a "zero". For a "one" a gate connected to the high voltage is open and for a "zero" the gate to the ground/low voltage is open.

[@craigcorson3036]

Does the voltage at the source affect the probability of tunnelling taking place? Seems like it would, but it's a quantum thing, so who knows?

[@Fred_Costa]

It very much does. (trust me, i'm on my second Quantum Mechanics course)

[@davidbeppler3032]

nope. our understanding of physics breaks at the macro and micro levels. "More research is needed."

[@craigcorson3036]

David Beppler Our understanding of physics at the macro level is quite complete. Admittedly, there remains much to be discovered about the quantum world. And, more research will ALWAYS be needed.

[@Toastmaster_5000]

I'm no physicist, but it definitely *seems* as though higher voltage does encourage tunneling. Higher voltage from overclocking is known to cause transistors to leak. Generally speaking, the smaller the transistor, the less voltage you want to use. But if there's too little voltage, you can't reliably activate the gates. Voltage gets harder to refine as it gets smaller. This is where the cost effectiveness comes into play.

[@craigcorson3036]

Gate voltage is not the same as source voltage. The problem as explained is that source electrons tunnel right through the gate if it's too thin, whether the gate is activated or not. It seems to me (though I don't really know) that dropping the source voltage would solve the problem. Gate voltage could remain whatever it needs to be. Obviously I'm just guessing.

[@jimpikoulis6726]

From what I understand by 2021-22 transistors within a Silicon Chip will stop shrinking. The drawback of S.C is heat generation as the ever increasing number of transistors makes it difficult to keep S.C within temperature limits. There are other limitations but for over 60 years S.C have served us well considering that the chip itself has kept pace with Moore's law.

[@forget2bhuman993]

this video was just the moores law video redone.... didnt actually say something is next... just 'oh we needa think of something new' yeah..... ok..... the videos substance was that? HAHAHAHA

[@undefeatedmc]

FOR GOD SAKES LIGHT COMPUTING! IT'S SO AESTHETICALLY PLEASING! Imagine gaming computer builds with mostly transparent illuminated parts ♥️

[@jsull81]

What about graphene transistors?

[@stardude2006]

J Snow 😊

[@itsmetheherpes1750]

quantum physics .

[@mantisnomo5984]

Probably.

[@kyleswoodworks2868]

it'sMe TheHerpes, quantum computers.

[@itsmetheherpes1750]

PLANET, i meant something else

[@MichaelClark-uw7ex]

The problem isn't just size but also distance between devices on the chips. Electricity travels at a finite speed so the farther it has to travel the longer it takes, modern CPUs are mostly 2 dimensional and some signals have to travel from one side of the chip all the way to the other which takes time. That leads to the logical conclusion that not only would making the devices smaller and therefore closer together improve speed but making them 3 dimensional would also shorten the distances even more and increase speed even more.

[@WhyteLis21]

Money should never be the limit for our technology advances. 😁

[@stardude2006]

WhyteLis21 Agreed

[@WhyteLis21]

stardude2006 😊👍

[@traso56]

who is going to pay for it then?? i would like advances too but nothing in this world is free

[@WhyteLis21]

traso as far as the world is concern, it is free. Only us human needs money or a currency of sort.

[@traso56]

then go ahead and tell the companies to make improvements for free because as far as everyone is concerned only humans can develop our techonoly

[@PrimiusLovin]

Personal random prediction with no serious investigation done on the subject: most of the performance we got from silicon chips was due to the shrinking of nodes that has been happening since the 1970's. We won't be able to shrink this technology much further, it's physically impossible. We'll need a completely new technology by 2030 if we are to keep the performance improvements (Moore's law) we're accustomed to in computing power...

[@mikechambers9129]

"Eat your heart out, Andy Dufresne!" ... Still laughing! Good one!

[@rahullamagna900]

How come TSMC and Samsung competing to build worlds first 3nm transistor if 5nm is the lower limit predicted?

[@socialmedia5534]

The limit is 2nm

[@honey4xi]

When Silicon chip reaches to the end, Quantum chip will be next. Will Quantum chip work in consumers' computers or only work in the industrial computers like IBM? *Will Intel and AMD make Quantum chip into computers for consumers?*

[@hihtitmamnan]

don't spread false info. quantum computers are nowhere close to x64 computers and their destination is totally different. also silicon is a material, quantum is mechanics of the microscopic world. jesus christ.

[@honey4xi]

@@hihtitmamnan Thank you for telling me the differences.

[@robertlong6311]

Larger processor clusters is an easy way to solve this. Just make bigger chips. When the main processor chips get to be about 5 pounds this will probably max out and and we will have a better tech to switch over to. Hopefully, before they get to the 5 pound mark of course.

[@dryaldibread2327]

This Is The End Of Seekers, Here's Is What's Next xD

[@perseverance8]

Even if & when they run into Fab limitations there are a number of design tactics that can be employed such as stacking working elements which is essentially being done with "3D NAND" flash storage.

[@romanvonungern-sternberg1322]

Why dOeS your thumbnails look so gooood 🤤🤤🤤

[@abulhassanmohammad]

Source Gate Drain is only applicable to -FET based transistors, which constitute the overwhelming majority of transistors used today but they're by no means the only ones. We use different notations for other types. Nice video nevertheless :)

[@adityas820]

Hi (Sorry for my bad english)

[@adityas820]

alex mercer ......thnx....I knew that I am missing something....

[@vvv2k12]

was* missing

[@adityas820]

vvv2k12 ....I've already said sorry 😒😒

[@pc_screen5478]

aditya solanke I've already said *I'm* sorry

[@adityas820]

AH SEU VOU ...............I am saying again .....SORRY....or I AM SORRY......

[@muramasa7537]

So if other elements are used to make new types of chips , will the coding system be same or will they change making us learn new methods or programming languages ?

[@CreamedCurry]

Graphene?

[@stardude2006]

Azure Co Yessssss !!!!!!!!

[@ChadKovac]

The electrons aren't "teleporting" through the gate, they're simply passing through it without intereference due to the space between the atoms.

[@판사님저는오늘만삽니]

damn... i wish i was born 100years later...

[@i_smoke_ghosts]

판사님 삽니다 저는오늘만 you must be from the future coz yor english is so futuristic ! Coz i mean wat are those → 오 오 ?!

[@digitalgaming3701]

판사님 저는오늘만 삽니다 or u can time travel 100 years later.. pretty cool though

[@davidbeppler3032]

Koreans are the perfect people. Ask them. ;)

[@damagecontrol7]

판사님 저는오늘만 삽니다 just get cryogenically frozen and wait it out 🤔🤔

[@shiroeloghorizon9771]

But what if u were forgotten and acidentally wake 500years later where everyone drink gatorade coz it have "electrolytes"

[@purplepanda8753]

Can we all agree that saying “we could use carbon nanotubes” sounds awesome?

[@Dyslexic-Artist-Theory-on-Time]

Zeros and ones of electronics is like waves and particles is quantum mechanics!

[@npm1811]

An artist theory on the physics of 'Time' as a physical process. Quantum Atom Theory Huh?

[@kindlin]

What's with your name bro. Tho, it summarizes you post pretty well, I'll admit.

[@jordanjacobson6046]

What is this Deepak Chopra nonsense your're saying? They couldn't be any more unlike each other. The actions of the transistors in a computing device are deterministic, and are on discrete binary values. Waves and particles in quantum mechanics represent a duality of quantum systems, the ones and zeroes of binary are not some duality, they are just symbols is a formal system that has been arbitrarily defined. Quantum systems typically dont deal with any kind of binary values, but instead have a (possibly infinite) set of discrete values the system can take on. They arent deterministic, they are random .

[@npm1811]

Jordan Jacobson lol my thoughts too

[@youfurz2]

Thats not how it works dude...

[@ArlanKels]

Light based computers sounds cool. But does that mean you'd need to have a flawlessly sealed container to keep external lights out?

[@doesntlooklikeanythingtome]

April fools

[@azmeriliza3788]

arnold lowe it's 31th March 😁

[@truthexposed7167]

Azmeri Liza oh....he fooled u in march...

[@sideoutside]

nope, just a fool.

[@ToriksLV]

I am a really slow learner but visual grafs made it very easy to understand it. nice

[@samsharma3216]

Your Awesome Man

[@johnkaboly5313]

Earth is flat flat

[@chris_tzikas]

Yea, his awsome man.

[@mantisnomo5984]

The uninitiated are incapable of distinguishing garbage from haute cuisine.

[@SIMKINETICS]

You missed describing the most important attribute of transistors; that they are amplifiers! The action of a diode gate in a transistor is initiated by power much lower than the switched or amplified power transmission in a circuit. A macroscopic model of this action is found in an electromechanical relay that uses a low-power coil to close (turn on) or open (turn off) an electrical circuit with electrical contacts with much more power transferred in the separate [secondary] main circuit. A transistor gate can be controlled with a small rise or descent in low current and/or voltage to either vary much greater power as an analog device or to simply switch such greater power as a digital signal device. The amplification & low-power switching in transistors allows miniaturization of logic circuits with commensurately lower values of power consumption, latency, noise, eddy-current interference & heat generation. The employment of quantum tunneling allows even greater miniaturization, but it doesn't alter the essential function of amplification in transistors.

[@cahidijoyoraharjo7833]

Is time travel scientifically achievable?

[@nashs.4206]

To the future, yes. To the past, no.

[@oari1150]

Nash Shrestha eh, if you talk about time traveling to a different universe, than maybe

[@ski88les]

No

[@benharrykirk2824]

Normally: Future yes, past no. With wormholes/timeholes: Future yes, past maybe. With negative mass: Future yes, past maybe. With black holes: Future yes, post-time yes (though you'd need to enter the black hole to do so, so you'd be dead anyways), past probably not. With relativistic speeds: Future yes, past no. With superluminal speeds: All bets are off. With a DeLorean: Go whenever the hell you want.

[@yayjuiws4224]

I hate how everyone is so sure in each and every of your comments when all those claims aren't even in a theoretical level.

[@plc3653]

we have to think the reason why we need to shrink the transistors at all. It is because of we need small and more powerful hardware to run our mobile devices. Consider this, if one day a new battery technology comes out that drastically shrink the volume of the batteries we use, there would be a lot of space left for bigger chips. There won't be a need to shrink anything else

[@maximalgamingnl9954]

NONESENSE

[@TenTenzo]

How is this nonesense? You obviously don't know how computers work.

[@maximalgamingnl9954]

Julian Stassen No, you're a computer

[@oari1150]

UNACCEPTAAAAAABLEEEEEEE

[@f.a3202]

Well at n-channel mosfet or fet current flows from drain to source. And electrons flow the other way around same as current at p-channel mosfet or fet

[@sekharmenonk8462]

yes 180th !!!!!

[@sekharmenonk8462]

highlighted!

[@carrotzombi]

I have a big question, when we hit that limit, why don't we expand the cpu chip area.

[@wa54]

YES. 50th

[@irfanulkarim4992]

WA earth is flat.

[@evananderson1455]

Karim Great Britain is the only reason anybody cares about India Luk i haz trollz 2

[@irfanulkarim4992]

Evan Anderson Earth is rectangular. Have you seen rectangular planets?

[@evananderson1455]

Irfanul Karim all the best shapes are rectangle

[@irfanulkarim4992]

Evan Anderson Great

[@madjimms]

"end of silicon!" We've been hearing this for a long time and nothing new has ACTUALLY replaced it in CPU's...

[@KalaMiDeviL]

The 1 and 0 thing is actually wrong. The only time you will find a current in a transistor, is when it changes its state. The real 1 is the voltage of 5V the transistor has after switching and that 1 is interpreted is by the effect the voltage has on connected transistors as it encourages their gates to react. And 0 would be no voltage. If the current was flowing all throughout the on-state, transistors would need much more space or room-filling cooling systems due to the heat generated by the current. Also, the chips intel makes nowadays would need massive power supplies to run and be very shortlived.

[@ThatOneScienceGuy]

When Trace first started out I thought he was a dork who wouldn’t make it. But I’ve grown to really like him.

[@sol666]

Molybdenum is already an extremely rare metal and is used in absurdly minute quantities in smartphones. It is unlikely to resolve the shrinking transistor conundrum when it cannot even be obtained on commercial scale.

[@grayswandir47]

The first transistors were based on germanium. Silicon turned out to be easier to purify and the crystals are more stable which made it more reliable. Is any research being done with germanium these days or is it just an entry in the history books. What became of gallium - arsinide? I read about the this material years ago as a potential successor to silicon.

[@claymccormick1203]

the short answer to what happens when we can't make the transistors any smaller is that we will get much more 3d but this causes problems with heat so we will also have to get more efficient there is a long way before we run out of ways we know to get more complex before we run out of ways to do it.

[@carlocarnevali7790]

3 nm is the minimum that can be achieved. Samsung and TSMC have fully developed 7 nm architecture, almost finished 5 nm, and samsung is starting to work on the 3 nm silicon. For that, 3 nm sets the limit of shrinking to date. But I think that even smaller architectures can be achieved.

[@和平和平-c4i]

Photonics could replace the actual processor technology but the title is really misleading since photonics also uses silicon. Molybden is way too expensive and silicon is by far the most present chemical element in the lithosphere (after oxygen of course). The point is no more to shrink the lithography, it is to replace the actual kind of chip for another one and photonics could be one of them.

[@meajur]

I once worked at a silicon wafer factory (edge rounder, acid etcher, lapper, polisher, and inspector). I'm not sorry to see it go.

[@thishadowithin]

When we get advanced quantum computing can we focus on the more important matters like creating a RPO (Ready Player One) game??

[@sparksthe_pegasus3978]

The Gate Drain Source architecture is actually that of a CMOS transistor or (Crystalline Metal Oxide Semiconductor) whereas, modern chips use Silicon! and the three pins are actually named the Emitter, the Base, and the Collector! More than that, the individual Logic gates are actually constructed in such a way that throws the basic transistor out the window, and is much more complex... not to mention the way they are put in a matrix together... but that's getting a bit off topic... main point: we use Silicon today! not CMOS variants!