The Incredible Evolution of Computers

0:00 In this video we’re going to explore

0:02 the evolution of computers over the past 80 years.

0:06 We’ll start with massive mainframe computers,

0:08 move into early personal computers and gaming systems,

0:12 go through the turn of the millennium and into the age of smartphones,

0:17 wearable devices and graphics cards,

0:19 and finally end at the advent of AI and the future of computing.

0:24 In essence, this video will explore how, in a span of a few decades,

0:29 we moved from room sized mainframe computers

0:32 that weighed as much as an elephant,

0:35 cost millions of dollars and were used to calculate

0:38 rocket trajectories for NASA and the Apollo Missions,

0:41 to smartphones that are a fraction of the size,

0:44 cost and weight, are around 16 million times more computationally powerful,

0:49 and are used for a range of activities,

0:53 none of which are calculating rocket trajectories To make this video,

0:58 we built accurate 3D models of over 60 different computers,

1:03 gaming systems, smartphones, and integrated circuits.

1:06 In fact, we purchased and physically disassembled

1:09 a number of these computers and devices,

1:12 then desoldered all the chips from the motherboard, took hundreds of pictures,

1:17 and built accurate 3D models of the interior components.

1:21 But to be clear, we’re not just going to list the different

1:25 computers and devices and show you the 3D models along with their specs,

1:30 but rather we’ll focus on the underlying evolution of science

1:34 and engineering that enabled smaller and faster devices year after year,

1:39 essentially answering why and how computers evolved the way they have.

1:44 We have 80 years of computers, technology, science,

1:51 and engineering to cover, so let’s jump right in.

2:00 This video is sponsored by Brilliant.

2:03 Perhaps you’ve heard of Moore’s Law,

2:06 which was the prediction made in 1965 that the number

2:10 of transistors on a chip would double every two years.

2:13 Therefore, you may think that the evolution of computers

2:17 was simply the progression of smaller and smaller transistors starting

2:20 with a few thousand transistors in a microchip and progressing

2:25 until reaching tens of billions of transistors in a microchip.

2:28 While Moore’s Law was rather accurate and profoundly impactful,

2:32 focusing solely on it is both an oversimplification and misleading.

2:37 For example, the Super Nintendo Entertainment System

2:41 and Nintendo Switch were released 26 years apart, and in that timespan,

2:47 the number of transistors inside these two

2:50 gaming consoles increased by 80 thousand times.

2:54 However, when it comes to processing power, the Nintendo Switch is closer to 1

3:00 point 4 million times more computationally powerful,

3:03 and the underlying reason for the exponential increase in processing power is

3:08 a lot more complicated than just simply how many transistors it has.

3:13 Therefore, throughout this video we won’t

3:15 focus on Moore’s law or transistor count, but rather we’ll explore how

3:21 the specific underlying technological developments directly

3:24 resulted in the increase in processing power across the evolution of computers.

3:30 Specifically, we can divide this timeline into 8 distinct time periods

3:35 or ages with each age having a different set of technological advances.

3:40 For example, in the 1950s, during the first age,

3:44 computers were changing from using vacuum tubes to using transistors,

3:48 and then in the second age scientists and engineers

3:52 focused on how groups of transistors could be packaged together.

3:56 Essentially, the advancements made in one age are

3:59 entirely different from the advancements in the next age.

4:03 Additionally, exploring these eight ages will provide

4:06 insights into how some companies were technology

4:09 leaders in one age but then failed to predict or adapt to the next.

4:15 For example, IBM was the leading computer company in the 50s,

4:19 60s and 70s and at its peak controlled around 70% of the entire computer market.

4:26 However, in the subsequent age, during the 80s and 90s,

4:29 the key advancements to computers changed and as a result,

4:33 IBM lost its dominance and now it owns less than 1% of the computer market.

4:39 Or, consider Intel’s decline through the 2010s when it

4:43 was a leader in the third and fourth ages,

4:46 and then failed to adapt and then declined throughout the fifth and sixth.

4:51 You can now see similar trends with Nvidia’s meteoric growth due

4:55 to its focus on GPU architecture and its use in AI algorithms.

5:00 And, since history is the best guide to predicting the future,

5:04 at the end of this video,

5:06 we’ll explore what the future of computing might look like.

5:11 As this video will focus on computational power,

5:15 simply listing each computer’s processing capabilities might get

5:19 you lost in the scale and number of zeroes,

5:23 especially considering today’s AI data centers,

5:25 so instead of just a string of zeros,

5:28 we’re going to use Lego Bricks and equate one operation,

5:32 instruction or calculation per second to a single two-by-four Lego Brick.

5:38 For example, when using Lego Bricks to show

5:41 the computational power of the ENIAC from 1945,

5:45 its processing power of 5000 addition computations a second

5:49 builds a moderately sized cube of 5000 Lego bricks.

5:53 Whereas the SNES from 1991,

5:56 with its almost 1 point 8 million instructions a second,

6:01 creates a Lego cube that fills most of a room.

6:05 And then, jumping to 2007,

6:07 the first iPhone’s 800 million operations a second builds

6:11 a cube of Lego bricks the height of a two-story building,

6:15 and the current graphics cards with their teraflops

6:18 of calculations forms a Lego Cube that gets ridiculously large,

6:23 but we’re not there yet.

6:27 So let’s go back to the beginning of this timeline

6:30 and dive into the first age of computers,

6:33 which spans from 1945 to 1962, and is named Transistorization.

6:38 Here are the five example computers

6:41 that we’ll use and explore in further detail.

6:45 We’ll start with the ENIAC built in 1945.

6:49 It’s important to note that there were a number

6:52 of room-sized computers that were built before it,

6:54 such as the electromechanical computer used to crack the Enigma code,

6:58 but we won’t explore them as they’re special-purpose computers.

7:02 What made the ENIAC notable is

7:04 that it was the first programmable general-purpose computer.

7:08 The ENIAC was built 2 years before transistors were invented and therefore it

7:14 functioned using 17 thousand vacuum tubes

7:17 making it rather large and incredibly heavy.

7:21 The next example computer is the UNIVAC-1 which was

7:25 built 6 years later and still used vacuum tubes.

7:28 So the question is, 4 years after the invention of the transistor,

7:33 why is this computer still full of thousands of vacuum tubes?

7:37 Well, the issue was that the early designs for transistors were very unreliable,

7:43 sensitive to voltage spikes, and prone to random breaking.

7:47 Vacuum tubes, on the other hand,

7:49 had been invented almost 50 years earlier, were considerably more reliable,

7:53 and if one broke, then the inside would

7:56 become foggy or the glass would be visibly cracked,

8:00 making it easily identifiable.

8:01 In contrast, broken transistors were a lot harder to find

8:05 and, when you have a computer with almost 20 thousand vacuum tubes,

8:09 switching them to unreliable, highly sensitive transistors that, when broken,

8:14 were a headache to find, was not ideal.

8:18 Therefore, during the age of Transistorization the key

8:22 technological advancements were focused on building reliable transistors.

8:26 To do this scientists and engineers first

8:30 switched from using Germanium to Silicon-based transistors,

8:33 and second the physical design evolved from point

8:37 contact transistors into individually packaged junction transistors.

8:41 By the end of the 50s the RCA501, NEAC-2203,

8:46 and the IBM 7000 series, such as this 7090,

8:51 were some of the first publicly available commercial computers

8:54 to use transistors and were advertised as transistorized computers.

8:59 Throughout the next few years transistors evolved from relatively

9:03 delicate-looking structures to more reliable flat planar transistors.

9:08 Then in 1959 the MOSFET was invented,

9:11 which is a family of different types of transistors.

9:16 Next in 1963 the CMOS circuit which is built from 2

9:20 different types of MOSFETs was invented and by the mid-1980s,

9:24 it became the fundamental type and arrangement of transistors

9:27 used in every computer even to this day.

9:30 While the design of the transistor was, and still is, being evolved,

9:35 we decided to have the first age of computers end in 1962,

9:40 with the last computer to use vacuum tubes.

9:44 It’s important to briefly note that while

9:47 we’re focusing on the computers and processing power,

9:50 computer memory has also had its own evolutionary timeline.

9:54 Early methods for storing data included rather interesting techniques

9:58 such as using sound waves in tubes of mercury, called mercury delay lines,

10:04 or utilizing cathode ray tubes and the glow of phosphor.

10:09 Eventually small magnetic toroids with hand-woven copper wires between them

10:13 called magnetic core memory and large magnetic drums were used,

10:17 and then after that, spools of magnetic tape,

10:21 and then magnetic disks, DRAM, and so on.

10:25 The evolution of computer memory and storage is incredibly complicated,

10:29 so we’ll explore it in a separate video.

10:32 There are many more details we cut for time,

10:34 but here are the additional notes you can read if you’d like.

10:39 Before we move to the next age of computers,

10:43 it’s important to understand that scientific breakthroughs

10:46 and technological progress typically don’t happen by accident.

10:50 But rather, every step, whether it’s from vacuum tubes to transistors

10:55 or from Graphics Cards to AI Processors, was made by many hard-working people

11:00 who had a fundamental and deep understanding

11:03 of science and engineering and applied

11:05 that knowledge through critical thinking and problem-solving.

11:09 That's where this video's sponsor, Brilliant, comes in.

11:13 Brilliant is an educational platform with thousands

11:16 of interactive and engaging lessons in math,

11:19 science, engineering, programming, data analysis, and many other subjects.

11:25 However, Brilliant is more than just a catalog of courses; rather,

11:29 it factors in your prior education and knowledge and then

11:33 starts you off on a journey with personalized lessons,

11:37 practice sets, in-depth instruction,

11:39 motivating challenges, and thought-provoking questions.

11:43 In essence, Brilliant is like a personalized tutor

11:47 that pushes you to deeply and intuitively understand complex topics,

11:52 such as college-level calculus, algorithms, circuits, and more.

11:57 Brilliant’s mission is to create a world of critical thinkers

12:01 and problem solvers and, with technology evolving faster than ever,

12:06 Brilliant continuously updates its lessons to anticipate

12:09 what you need to know for your education, career, and the world to come.

12:14 For example, they have a new and in-depth course on AI and Large Language Models

12:20 that explains how Generative AI works far

12:24 better than any textbook or video currently available.

12:27 To learn for free on Brilliant, go to Brilliant.org/BranchEducation,

12:34 or by scanning the QR code on screen, and you’ll then have free 30-day access

12:39 to the wide range of courses throughout their catalog.

12:42 If you enjoy their content and decide to stay,

12:45 the link in the description below will also

12:48 save you 20% off an annual premium subscription,

12:52 giving you unlimited daily access to everything on Brilliant.

12:56 So, let’s move on to the second age,

13:00 spanning from 1964 to 1977, which we’ve named “Packaging Transistors”.

13:08 Here are the example computers we’ll use to demonstrate

13:11 how computers evolved from room-sized mainframes to fridge-sized minicomputers,

13:16 and then to personal desktop computers, which were called microcomputers.

13:21 It’s funny to consider these rather large computers as mini and micro,

13:28 especially when compared to today’s smartphones and smartwatches, yet,

13:33 when matched up with room-sized mainframes, these names made sense.

13:38 Regardless, for most of this age, the IBM System 360 mainframes,

13:44 which were released in 1964 and discontinued in 1978,

13:50 were the most popular and successful family of computers.

13:54 These mainframes were so dominant that, from the late 60s to the early 70s,

14:00 around 3 out of every 4 computers made were IBM System 360 mainframes.

14:06 These computers were revolutionary for many reasons,

14:10 but one key concept they established was that the entire

14:15 family of 360 mainframes used a standardized computer architecture,

14:20 allowing the same software to run on any model.

14:24 Essentially, the IBM System 360 laid

14:27 the groundwork for device and software compatibility,

14:30 as well as the expectation that customers buy and upgrade

14:34 to newer models every few years with minimal disruption.

14:39 For example, using the System 360,

14:42 it was possible to upgrade the CPU, and to do so,

14:46 all you needed to do was swap out the CPU cabinet with the latest one,

14:51 each of which was technically called a frame or unit.

14:56 Alternatively, if you wanted to increase the computer’s

15:01 memory capacity from 256 kilobytes to 512 kilobytes,

15:05 which is like adding sticks of DRAM to a desktop computer,

15:09 you simply needed to install additional

15:11 cabinets containing arrays of magnetic core memory.

15:14 Due to the System 360’s impact on the computer industry,

15:18 it’s considered the Model T of computers and, as a result of its success,

15:24 IBM had a monopoly over the computer market,

15:28 eventually releasing a total of 14 standard

15:31 production models and selling over 33,000 mainframes,

15:35 with the cheapest ones costing more than a million dollars.

15:41 One question you may have is, considering that these computers were built using

15:47 transistors around the size of a sesame seed,

15:50 which are small compared to vacuum tubes,

15:52 then why were these computers still so big, taking up the area of a room?

15:59 To answer this, let’s explore the IBM System 360 Model 75,

16:05 which cost around 36 million dollars,

16:08 could execute a million instructions a second, and was used by NASA to calculate

16:16 rocket trajectories during the Apollo missions.

16:19 Let’s start with a quick tour of this computer and begin

16:23 at the system Control Panel with its hundreds of switches,

16:26 knobs and lights that were used to operate the computer,

16:30 read the data in the registers, quickly debug, and many other things.

16:36 Attached to the control panel is the CPU, along with the CPU memory.

16:42 Each memory cabinet can hold 256 Kilobytes of Magnetic Core

16:46 Memory with a blindingly fast point seven five microsecond access time.

16:52 If you needed more memory, you would install additional CPU memory cabinets,

16:58 each holding 2 megabytes with an 8-microsecond access time.

17:03 For long-term data storage,

17:05 this computer used cabinets containing reels of tape,

17:09 each reel holding 45 megabytes and using 2400 feet of magnetic tape.

17:16 Alternatively, for faster access times,

17:18 you could save data to large magnetic plates or even larger magnetic drums.

17:25 To load a program into the computer,

17:27 you had to write it onto thousands of punch cards

17:31 and then load them in via the punch card reader,

17:34 which could process anywhere between 200 and 1000 punch cards a minute.

17:39 Also, there was a printer, a keyboard, and a cathode-ray tube display.

17:45 All these modules were connected using thousands

17:48 of wires running through the raised floor.

17:52 We don’t have time to go through the inner workings of each of these cabinets,

17:57 but let’s take a look at the CPU and see

18:00 how it works and what makes it so large.

18:03 Inside this cabinet, we find rows and rows of circuit boards.

18:07 Upon examining one of them, we can see around two dozen square metal components,

18:13 known as SLT or Solid Logic Technology.

18:17 Inside each SLT package,

18:20 we find several individual sesame-seed-sized transistors,

18:24 along with diodes and resistors, all connected on a ceramic substrate.

18:30 This package was built by first printing wires and resistors on the ceramic,

18:35 and then mounting a separate set of discrete transistors and diodes on top.

18:40 These components are protected by a metal cover,

18:43 and the SLT package has thru-hole pins sticking out

18:47 the bottom to mount it to a printed circuit board.

18:51 One of these SLT components, for example, is an AND logic gate;

18:55 another is an OR gate; and here’s a set of inverters.

18:59 Also mounted to the board are capacitors,

19:02 inductors, and other electrical components.

19:04 By connecting all the SLT packages and other components,

19:08 this board, for example, was able to add two binary numbers.

19:14 CPUs are incredibly complicated with many different sections and therefore,

19:19 to build a complete CPU,

19:21 you needed an entire set of cabinets filled with circuit boards,

19:26 each covered with SLT components.

19:28 Therefore, to build a working mainframe computer you

19:32 needed a room of cabinets filled with computer hardware.

19:36 So let’s move on from mainframe computers.

19:41 You may be familiar with early desktop

19:44 computers and microchips that look like this.

19:47 Inside this microchip is an integrated circuit or IC,

19:51 so then, where does the IC fit into this timeline?

19:56 Well, you may think that SLT packages containing just

20:00 a few transistors each were the precursor to integrated circuits.

20:04 However, the first integrated circuit was actually invented in 1958,

20:10 which is six years before the first

20:14 System 360 mainframe and SLT packages were released.

20:18 Therefore, integrated circuits were an alternative method to manufacturing

20:23 and packaging transistors and were developed in parallel to SLT packages.

20:29 So let’s discuss Integrated Circuits or ICs.

20:33 To start, they’re called integrated circuits because multiple transistors,

20:39 resistors, capacitors, and diodes are built simultaneously using a complex set

20:44 of manufacturing steps and then connected together using layers of wires.

20:50 Despite ICs appearing relatively flat,

20:53 they’re actually complex, multilayer, 3-dimensional structures.

20:56 In contrast, discrete transistors have only a single transistor in either

21:02 a package or are found as multiple separate ones in an SLT.

21:07 In the early days of integrated circuits, between 1958 and 1965,

21:13 scientists and engineers were able to combine only a few transistors

21:17 into an IC in order to build the basic logic gates.

21:22 These chips had at most 10 transistors along with a few

21:26 other components and were categorized as small-scale integration or SSI.

21:31 They were smaller and lighter than the square SLT packages,

21:36 but also incredibly expensive, initially costing over 10,000 dollars each.

21:42 Therefore, early integrated circuits were only used in specialized applications,

21:48 such as the Apollo guidance computer.

21:51 By the mid to late 60s,

21:54 through substantial improvements in manufacturing techniques,

21:57 integrated circuits had progressed to medium-scale Integration or MSI,

22:02 which ranged from combining or integrating 10

22:07 to 500 transistors onto a single silicon substrate,

22:10 and they started to be more affordable

22:13 and find their way into commercial computers.

22:15 An example computer using a mix of SSI and MSI

22:20 circuits was the commercially successful PDP-8 minicomputer from DEC,

22:25 which could do 385,000 instructions a second,

22:30 and greatly expanded the accessibility to computers

22:33 with its smaller size and semi-reasonable price tag.

22:37 Other computers that used MSI chips were for example, the Xerox Alto from 1973,

22:44 which contained 4 separate 74181 chips, each consisting of 170 transistors,

22:52 which could do parts of simple calculations.

22:56 The next step was Large Scale Integration or LSI,

23:00 which ranged from integrating 500 to 20,000 transistors in a single chip.

23:07 In 1971, Intel released the 4004, which is widely regarded as the first

23:14 commercially available CPU packaged in an integrated circuit.

23:19 This chip contained 2,300 transistors, processed data in 4-bit chunks,

23:25 and could execute 92 thousand instructions per second.

23:29 While this was indeed the first CPU chip,

23:33 it was incredibly weak in terms of processing

23:36 power and due to its 4-bit calculations,

23:38 and therefore its main applications were

23:41 in things like simple calculators and cash registers.

23:45 By the end of 1974, two more CPUs, the Intel 8080 and Motorola 6800,

23:52 were released, and, while they powered some microcomputers,

23:56 the chips alone cost around twenty-five hundred dollars.

24:01 These high prices propelled the development of the MOS 6502,

24:06 which was released in 1975, cost only around 150 dollars,

24:11 contained 4528 transistors, and ran an 8-bit architecture.

24:17 Due to this chip’s affordability and decent processing

24:22 power at 4 hundred 40 thousand instructions per second,

24:26 it kicked off the age of personal computers with the 1977

24:31 release of the “trinity of personal computers” consisting of the Apple 2,

24:37 the Commodore PET, and the TRS-80, two of which used the MOS 6502.

24:44 The 6502’s architecture and low cost were so impactful

24:48 that it was used in dozens of computers and devices,

24:53 including the Atari and the NES, and many others.

24:57 Let’s move on to the next step in integrated circuits.

25:01 With further improvements in manufacturing,

25:03 Very Large-Scale Integrated Circuits or VLSI chips

25:07 containing 20 thousand or more transistors were released.

25:12 For example, the Intel 8086, introduced in 1978, is one of the first 16-bit

25:19 CPU chips and contained approximately 29,000 transistors,

25:23 but we’ll discuss the progression of these chips more in the next age.

25:30 We decided to end age 2 in 1977

25:33 as it’s considered the start of the mass-market personal computer,

25:37 but before we move on to the next age, let’s explore a critical question.

25:43 If IBM had a monopoly on mainframe computers throughout the 1960s and 70s,

25:49 why didn’t it stay on top and lead the development of integrated circuits?

25:55 Well, as we discussed, IBM established its market dominance with the System

26:01 360 using a significant investment in SLT packages.

26:05 At the time, IBM was vertically integrated and built entire factories

26:11 and production lines focused on mass-producing

26:14 cheap and reliable SLT components,

26:17 as well as the entire mainframe and every step in between.

26:21 In contrast, in the early 60s,

26:24 small-scale integrated circuits were incredibly expensive and only

26:28 used in niche applications like rocket guidance computers.

26:32 It’s important to note that IBM did develop

26:36 integrated circuits for memory storage and, in fact,

26:41 patented the first DRAM integrated circuit in 1968.

26:46 Additionally, following the SLT,

26:48 IBM developed and transitioned to MST or Monolithic

26:52 System Technology for the 370 family of mainframes,

26:56 which was essentially a small integrated

26:59 circuit packaged within an SLT form factor.

27:03 However, their delay in developing ICs

27:06 allowed companies like Fairchild Semiconductor, Texas Instruments,

27:10 and eventually Intel to build up a considerable lead in IC research,

27:17 development, and manufacturing.

27:18 Essentially, SLT made IBM the market leader, however,

27:23 at the critical transition point to Integrated Circuits,

27:26 it didn’t make sense for IBM to siphon money away

27:30 from the successful and established technologies and pour it into ICs.

27:35 One additional note is that between 1969 and 1982,

27:40 the US Government conducted anti-monopolistic investigations into IBM,

27:45 which likely inhibited IBM’s interest in using its market

27:49 dominance to be the frontrunner in the integrated circuit market.

27:53 On a separate note, in the 60s and 70s,

27:57 the potential of the personal computer market was unknown.

28:01 Many experts were familiar with large,

28:04 expensive mainframes and minicomputers used in businesses and universities

28:08 and therefore couldn’t foresee computers being scaled down

28:12 to the size of a desktop or the role

28:15 the personal computer would play in individual households.

28:19 In fact, in 1977, the CEO of DEC claimed that "There is

28:25 no reason for any individual to have a computer in his home." Furthermore,

28:31 despite the successful release of many popular personal

28:35 computers in the late 70s and early 80s,

28:39 it wasn’t until 1984 that the personal computer

28:42 market surpassed the mainframe market in terms of revenue.

28:46 There are two lessons to take away.

28:49 First, it’s impossible to know for certain which technology

28:53 will be the frontrunner from one age to the next.

28:56 And second, while it may seem like

28:59 the personal computer replaced the mainframe computer market,

29:02 it’s more accurate to say the personal computer

29:06 market simply had greater growth potential in subsequent years.

29:10 In contrast, the mainframe market’s expansion was in the 60s and 70s,

29:16 and stagnated in the 80’s, and declined after that.

29:20 A similar trend can be observed today.

29:22 The demand for AI data centers is rapidly growing,

29:26 but this doesn’t mean that AI data centers are taking over desktop computer,

29:32 laptop, and smartphone sales, but rather the market for these devices is

29:37 simply already mature and thus not growing as much.

29:41 There are many more details on the second age of computer evolution,

29:45 and here are additional notes that you can read through if you'd like.

29:50 Now that we’ve finished the age of Packaging Transistors,

29:54 let’s move on to the third age, spanning from 1975 to 2000.

29:59 You may think the third age is on Personal Computers,

30:04 but because we’re focusing on technological changes

30:07 and what drove the increase in processing power,

30:10 the third age is on increasing CPU clock frequency,

30:14 and we’ve named it The Frequency Race.

30:17 In this age, we’ll see how computers

30:19 go from a one-Megahertz clock frequency to one-Gigahertz,

30:23 yielding a speed-up of at least a thousand times more processing power.

30:27 But before we get there,

30:29 we want to mention that making these videos is incredibly complex,

30:34 and while we use some historical pictures,

30:37 we tried to fill this video with as many

30:41 intricate 3D models and details as possible.

30:44 Specifically, each 3D model with its interior components

30:48 takes hundreds of hours to build in Blender,

30:51 and therefore, all these models have taken

30:54 more than 3000 hours to build and animate.

30:57 We love having the opportunity to provide this video for free,

31:00 and while we do get some money from our Sponsor,

31:04 Brilliant, and from YouTube Ads,

31:05 it would help a ton if you could support us on Patreon.

31:10 The main perk is that we include the names

31:13 of every contributor at the end of each video.

31:15 We’re also working on exclusive behind-the-scenes

31:18 videos and considering other perks as well.

31:22 Also, if you enjoyed this video,

31:24 please take a moment to leave a comment or question below,

31:28 and we’ll do our best to respond.

31:30 Make sure to subscribe to the channel if you haven’t already, like this video,

31:35 and then, most importantly, share it with a friend, family member, or colleague.

31:40 All these things help us to create more videos, and we thank you for your time.

31:46 Finally, as you may notice, we’re near the end of this video.

31:50 The evolution of computers is incredibly complex,

31:53 so we’re planning to divide this video into 4 episodes.

31:58 This is the first episode and covers the period from 1945 to 1977,

32:05 encompassing the first two ages.

32:07 The second episode will explore the third and fourth ages,

32:12 spanning from 1975 to 2011, and called the Frequency Race, and Multicore CPUS.

32:18 The third episode will cover the fifth age with FinFets, Smartphones,

32:23 and System on a Chip or SoCs, and then the sixth age will be on Graphics Cards.

32:30 And then the final episode will explore age seven and AI processors,

32:36 or Tensor Processing Units,

32:38 and their use in AI Algorithms, and then, in the eighth age,

32:43 we’ll explore what the future of computers will look like.

32:47 Once we’ve released all these episodes,

32:50 we’ll combine them into a feature-length movie about the evolution of computers.

32:54 Stay tuned for the next episode, which, to be honest,

32:58 will probably take at least 3 or 4 months to make,

33:02 as we’re a small team, but we hope it’ll be worth the wait.

33:07 Until then, we’re thankful to all our Patreon

33:10 and YouTube Membership Sponsors for supporting our videos.

33:13 If you want to support our work financially,

33:16 you can find the links in the description below.

33:20 This is Branch Education, and we create 3D animations that dive deeply

33:25 into the technology that drives our modern world.

33:28 Watch another Branch video by clicking one

33:31 of these cards or click here to subscribe.

33:35 Thanks for watching to the end!