All right, welcome to the stand up.

Today we have asked legendary game

programmer Casey Muratori to give us an

amazing presentation and what he's going

to do is going to give us a glimpse of

the root of the root of all

evil is pre-optimized performance. Okay,

I can't even I can't even remember the

title. It's too many too many crushes.

>> Thank you. I'm crushing it. I was also

going to say prime I mean I think we

need at this point probably add computer

historian

to Casey's repertoire.

>> At this point maybe.

>> That's true.

>> Yeah,

occasional computer historian perhaps.

>> Yeah. Okay.

>> Yeah.

>> Today we have with us legendary

[laughter] game programmer and competent

and very insightful his computer

historian Casey Muratori.

Anyways, I'm sorry.

>> I love the legendary like the this not

legendary at all but now you have made

it a thing and under my videos people

NOW PUT LEGENDARY

>> [laughter]

[applause and cheering]

>> YOU GOT TO CHANGE YOUR USERNAME.

>> WAIT. WAIT. CHAT, can you guys search

Casey on Google? Okay, [laughter] I

don't know what the legend There's no

legends. There's not a single legend.

Anyway, um

>> Chat, we have the funniest thing we can

do. You know AI overviews? If we keep

this up

>> Oh, no.

>> AI overviews will have a

>> legendary

>> Yeah.

>> [laughter]

>> game programmer Casey Muratori.

>> Yeah. That's what I want to see.

>> All right, spin up your bots.

>> Yep.

>> So, it's time.

>> The legend is about how how I started

being called [laughter] legendary.

That's the actual only legend.

And that will be an actual legend since

people will have to explain it cuz it

doesn't make any sense. Anyway,

yeah, so

a lot of people last year saw that big

oops talk that I did because for some

reason it was very popular which was

just luck I guess cuz it's very long.

>> Um it's very long and I wasn't expecting

that. So, just like I don't know, people

were in the mood for a longer thing, so

that was cool. And so, a lot of people

watched the big oops. That was my Better

Software Conference 2025 talk. So, for

Better Software Conference 2026, I just

wanted to do uh one more of those.

They're very hard to do. I think I've

talked about this before. You have to

just read a tremendous amount and you

have to condense things down so

dramatically that like you're leaving

most of the things out. Uh you know, a

lot of people ask me like, "Why didn't

you talk about X or this that?" And I'm

like, "Because, bro, there is so much

stuff you can't talk about even spending

2 hours. There's just I mean, it was

thousands of pages of documentation.

Everything had interesting stuff in it.

I mean, computer history is just

chock-full of things. It's hard to get

through a single paper without being

like, "Wow, I this would be cool to talk

about, but it just doesn't fit here or

whatever."

So, anyway, um

So, this year I wanted to do one more of

those and it was uh on the phrase

"Premature optimization is the root of

all evil."

And the reason I wanted to do that one

is cuz I before I started researching it

a while ago, I actually didn't really

know the circumstances around the sort

of like first times it was committed to

page, if you will. Like written into an

article and or had uh examples shown

about it. And I kind of wanted to go I I

knew about the original sort of sighting

in the Knuth paper.

>> [clears throat]

>> Sort of the first paper anyone wrote

that has it in it. Knuth actually said

it uh

according to a transcript. He said it in

a lecture slightly before that paper was

printed, but the paper's manuscript had

already been submitted at the time. So,

this is like the first time it would

have ever really been committed to paper

as far as I know.

Obviously, people probably were saying

it before that, but anyway.

So, I wanted to find out more about like

why it got said, like why would someone

like Knuth write it down, you know, what

was going on at that time period. I just

kind of wanted to get it more than I did

because I just didn't really know, and I

obviously know Knuth's like research. I

know he has the Art of Computer

Programming series and a lot of other

like he has his discrete math book and

all these like publications, all that

stuff, but I never really like knew much

about his history.

And so, I kind of started like looking

into all of that and and kind of found

this whole story about structured

programming and what was going on at the

time that I found really educational.

So, really the talk is just an attempt

to kind of share that point in history.

It's like, here's kind of what they were

thinking about, what they were tackling,

what professors like Knuth were looking

at at that time. And once you kind of

know all that, I think it puts the

phrase in much better perspective about

what like it actually means. It's like a

really much more tangible

>> Yeah.

>> way you can relate to the phrase. So,

anyway, that's what the lecture is.

>> If you're anything like me, you have a

bunch of ideas, but perhaps you're a bit

disorganized. So, often I find myself

having a long brainstorming session, but

at the end of it, I just have a big

markdown file and then of course at that

point I've lost everything. Hey, stop

doing that. You got to check out Linear

and use the Linear MCP. I have a big

conversation, I come up with a bunch of

ideas, and at the end, I get a set of

tickets that are actually all the things

that I need to do. And when I don't know

what to work on, I go, "Hey, Agent, go

to Linear and tell me what I need to do

next." And it will find exactly what I

need to work on. This has been so

successful for me that I'm in actual

serious danger of completing a project.

Go to linear.app/terminal

and get 6 months for free on Linear.

Thank you for supporting the channel,

and thank you Linear for sponsoring the

video.

>> It's exactly like the big oops where I

you know I use pretty much exclusively

excerpts from source material to kind of

like give you a sense of everything that

was going on and who is involved and all

this other stuff.

So um that's up now. The VOD is up. You

can get to it easily if you just go to

computerenhanced.com/theroot.

All one word, the root. It will take you

directly there. And I have a page where

it's got the YouTube video of the talk

and then uh references hyperlinked so

you can actually like any of the slides

that you see if you're like I want to

read more of that uh I'd like to see the

original. The they're all hyperlinked on

that page. You can just click through

and get them uh and read the original

references. Sometimes they're paywalled

cuz they're at like you know it was an

institute uh like you know it's IEEE's

like you know publication that isn't uh

available for free currently or

whatever. So but most of the time there

there's just a freely available PDF that

the author has posted or or whatever or

or that journal is is out of out of

print for so long that it's just on you

know they just put their archives

online. So

>> Casey, you know that you could uh

Oh, never mind.

Never mind my joke. I just realized

completely doesn't work. Keep on going.

>> Brian, we could fix it in post.

>> an actual question.

>> Josh, definitely leave it in.

>> Brian, go again.

>> No no no I like I can't because I

thought computer enhanced was CH for

abbreviating it or for initializing it

but then I realized there's the EN piece

then it could have been CH root. It

could have been charoot.

Yeah yeah yeah yeah that doesn't work

but then I realized I'm an idiot.

>> [laughter]

>> I need to capitalize that. It used to be

like computer enhanced with like a big H

and then it can

Uh okay sorry what were you going to say

Trash?

>> Trash I had a trash

>> Uh real question. What does um for these

long talks like how many hours does it

take to

like like what's the level of effort

that you're going through to make these

long talks?

>> Way too much I guess is what I'd say.

Typically, I do like typically the

the idea or basis research comes from

something I happened to do earlier, so

like a year or two ago or something like

that. And then I spend

it's just like I don't know maybe like 2

weeks of 16-hour days

like straight just going through all the

material, clipping it out. It's like cuz

I can't afford to spend that much time

on them and I have to go through so much

history that's just It's It's really

brutal

cuz then just going back and and you

know

going through all the papers again,

finding new papers, answering questions

that like are like oh this

what about this part? Like where's that

reference or why did this person do

this? Can I find something?

And so it's it's really really brutal

and I don't recommend it.

I don't recommend it at all

>> [laughter]

>> to anyone.

>> Every My rule of thumb is always been 40

hours per hour when it comes to a

conference talk.

>> Um that's

>> Yours sounds even higher.

>> It's It's higher than that just because

of the nature of the work because of how

much reading is involved. Um and and I

think

you know, one of the things that I uh

sort of talked about with Dimitri cuz we

do that AI podcast and we do an episode

coming out later where I I sent him like

a paragraph describing like the research

and I was like Dimitri, have the have

the AIs do it and we'll compare them

like afterwards on an episode. We'll go

through it and see like how they did cuz

it was going to be like a kind of a

shoot-down episode, right? Where we go

like here's what Anthropic got, here's

what OpenAI got, here's what Gemini got

cuz now I know the thing like backwards

and forwards I can kind of be like who

who did what when and how. Uh

And like

one thing I was thinking about on that

episode, too, was just

I don't know how you speed this process

up. You can't really outsource it to an

AI or something like that or even a

research assistant because the valuable

part is you exploring the text without a

preconceived notion of like what the

questions are that you want to have

answered, right? And so it's kind of

this It's It's this really difficult

problem of making a talk like this. It's

The value comes from you immersing

yourself in the material, not from

answering a particular question. And so

it's very hard also to figure out how

you would accelerate it. Like using

using AI, using assistance, using

whatever. I just

It's It's not really

I don't know. It just kind of is a long

process is I guess all I would say. And

so eh

>> I've been I've been trying to do that as

well, seeing if I can like Can I do

research and all that? And then at the

end of the time after spending, you

know, 3 hours getting research from an

AI, I go, "Ah, I have to read the

paper anyways." And so then I just go

read it like either way. And so now I

just just like, "Well, I can't say this

thing because I don't even actually know

how accurate it is or if it really is

the like the thing I'm thinking." And so

then

>> And And I wouldn't like I don't want to

say it too strongly because since I

haven't tried actually using AI to

prepare a talk like this, I don't know

if maybe like you could come up with a

workflow that sort of works.

>> Yeah.

>> But in my experience, like, you know,

you doing reading all of the papers

yourself, like, all the stuff you find

is so valuable and you never knew you

were looking for that, right? Uh and so

that's just That's the tough part about

it is it's all about that happenstance

of exploration. I really enjoy it. It's

just the time pressure is so brutal. And

I'm never going to get a chance to

rehearse these talks, either. So, I just

have to get up and give the talk just

from the slides that I ended up with

that are were done, you know, at just

before

I get up there, right? Because I'm

constantly making modifications. And so,

it's just it's a it's a really grueling

experience is I guess what I would say.

Uh but so, I probably this is probably

it. Those are probably the two that I'm

ever going to do.

Uh but, you know, I certainly learned a

ton when I was doing it.

>> Nice.

>> I think the stuff for me that I found

helpful uh using AI for it is like

finding original sources for stuff,

which like if you have papers already as

a starting point, then you can use those

and you could literally read them

through. But sometimes you're like, I

don't even know can you like find

anything to read? Like that's very

helpful.

>> Right.

>> Um and to like even just, you know, hey,

can you turn these really badly

formatted PDFs or like weird old web

pages? Like you can just

>> ask it.

>> Like I ask it sometimes I literally say

like, just turn this into a web page

locally in H like just that I can click

through and open it like a book. You

know what I'm saying? Or some stuff like

that, which just helps.

>> Like OCR this for me cuz this is a mess.

Yeah.

>> Literally OCR this.

>> right now that I can't scroll cuz of

just the way it is on the internet. It

just won't scroll. It just is so chunky.

So, I'm just like, make it different.

Thank you.

>> Yeah, yeah, yeah.

>> Or like collapse it into something like

I have this one thing right now where um

I just it goes through uh the the guy

who made Obsidian made a thing that like

turns most web pages into like really

simple like markdown that you can read.

So, like use his little package to like

turn some stuff into like offline

markdown. I can read articles like from

my phone. So, like that stuff's pretty

nice if you're trying to just like get

stuff.

>> All right. Hey, Casey, you need to kind

of keep on going with this presentation.

I realize we're pretty far off topic.

>> That's okay. I mean, it's the standard

>> I already gave the talk. If you want to

see the talk, it's there. Uh so,

>> So, uh yeah, there uh

we were kind of talking about

obviously these talks are not things I

can give other versions of, right? So,

there's not much I could really do uh on

the stand-up, but

what I kind of wanted to do was figure

out something I could show on here that

might be fun because, you know, I like I

already said when we were talking about

this, you basically you you don't really

have um time to put in almost any of the

things you found, right? What you see in

the talk is like a tiny little sliver,

right, of the actual um of the actual

stuff, right? And so,

what I wanted to show here was there's

this um

one of the period- There's there's

multiple periodicals that like everyone

read at the time. Communications of the

ACM

uh was an example of of one of those.

It's in fact Communications of the ACM

was what that original Donald Knuth

Structured Programming with Go To

Statements paper where Premature

Optimization is the Root of All Evil,

where he writes that uh that was

actually in Communications of the ACM.

That was like a very common thing people

would would see. Um

Actually, no, that I I'm sorry. Uh

that was in Computing Surveys.

Um

Communications of the ACM is kind of

like more like has the letter letters in

it and stuff like this. So, those

periodicals from the Association of

Computing for Computing Machinery, those

are like very common periodicals that

lots of people read. But, there was this

actual magazine called Datamation, d a t

a m a t i o n, that was a very prominent

like trade magazine at the time, and a

lot of things are like are in there.

And so, I happened to have to read

through those, as you'll see in the

talk, uh I had to read through those to

like dumpster dive to find this one

advertisement that Edsger Dijkstra, like

like of the Dijkstra algorithm, like

that Dijkstra, uh programmer of the

month refers to multiple times. He

Sorry?

>> Programmer of the month on Prime's

channel. So that was cool.

>> Oh.

All right. Well, yes, programmer of the

month Edsger Dijkstra.

>> Yep. Very important. Very prestigious.

>> Yeah. It's literally like Turing Award.

Programmer of the month.

>> Yep.

>> Um yeah, I was going to say that Turing

Award is kind of like a nice to have

uh resume patter, but programmer of the

month is when you know you've really

made it, right?

>> you know you've Yeah.

>> You know.

>> I get that one day.

>> So someday we're all aspiring to

uh

So anyway,

in this this magazine uh like Dijkstra

actually refers to an advertisement

derogatorily refers to an advertisement

in this magazine Datamation and that

that's in the talk you can you can go

see that. But while I was trying to find

this ad that he sort of, you know,

cryptically refers to multiple times,

I read through all the issues of

Datamation like in 1968.

And because they've been digitized by

this like preservation uh patron called

Bit Savers.

And while I was reading through these,

it kind of like just struck me how

it it's weird, but like the more things

change the more they stay the same. In

1968, like computers were like massive.

Like most a lot of them were size of the

room. They were at least cabinet-sized,

right? Uh they had like, you know,

teletype for for like talking to them

stuff. It was like, you know, they were

like terminals. Occasionally you would

have a CRT display that could maybe like

draw some lines on it for you and things

like this. They also had big plotters

that they would use, you know. It's that

era of computing history. Nobody has a

personal computer. That's not a thing. I

don't even think the term personal

computer would be used at this point,

right?

>> I think '75 was that Apple I? I don't

I don't know, but like uh cuz yeah, that

heaven help me.

>> It was called a personal computer.

>> Yeah, and so you know, we're talking

about 1968. So several years before the

the term was really a thing that people

were thinking about, right?

But anyway, when you read through it,

it's still kind of nuts. Like you get

this really interesting sense for the

time period. The sorts of things that

were kind of different. Like now you can

buy a high-resolution CRT for 60 to $100

in 1968 dollars. So, you know, it's kind

of interesting to see how some things

have have uh

have gotten better. But in a large to a

large extent, the stuff that they talk

about is like the exact same stuff we

talk about today. So I kind of just

wanted to grab some screenshots. I'm

sorry, I just rapidly did this last

night Prime when you said

that I should I should show something

today. So these are not as organized as

I would like.

>> That's perfectly fine. We love to hear

from you.

>> say that's $60,000 in today's dollars if

I did the Google search correctly.

>> it would That's it's around 10x. Yes.

>> Yeah.

>> So I don't know if you can see here, but

this is an adverti- well, it's not

really an advertisement. It's an

announcement for a new product of a 4K

display.

>> [laughter]

>> In 1968.

Right? 4K 4096 by 4096

raster elements.

Right?

>> doing here?

>> I'm [laughter] just I'm just pointing it

I don't know. It's hard to get more

information about some of these things,

but I thought that was kind of cute. Um

Anyway,

>> [clears throat]

>> I thought it was kind of cool that we

could get some salary comparisons. So

again, 10x these. So, you know, here's a

partial list of computer careers.

$100,000 to $350,000 a year.

So

>> Nice.

>> It was already like people were getting

Silicon Valley salaries.

>> True.

>> In 1968.

>> Wow.

>> Um

which is pretty great.

Uh,

AI

taking over the world and us being very

scared about it.

>> That was huge [clears throat] during

that time.

>> Hexapon, a learning demonstration. This

is by some people

from IBM.

And they essentially say that

that

what what was the good thing it says?

The um,

uh,

This arises from the belief that the

computer is a serious threat to man

because it poses and it possesses an

intelligence that dwarfs man's. Many

people do not realize the contribution

that programmers make to computer

performance. Instead, they ascribe the

performance to some inherent mystical

ability of the computer. This naturally

results in their regarding the computer

with considerable awe.

>> [cough]

[clears throat]

>> And it has I think it says there is

therefore a need to communicate to the

public that basic principles underlying

computer artificial intelligence in

order to dissipate some of the mystery

surrounding computers.

And there's another one in here that I

was going to point out where they talk a

little bit more about the

the degree to which people

In fact, this article itself says

accompanying this positive attitude

toward computers is an underlying sense

of anxiety. There's there's more here on

here.

Anyway,

this

These are in no particular order, so

we'll get back to that thing that we I I

just showed in a minute.

>> Is that a COPE 45?

>> That's what I was going to say. This is

This computer is called the COPE 45.

>> [laughter]

>> Brian, you can reply to people with this

on X with this ad when they say

something you're like, "Uh, here's a

COPE 45 for you."

>> [laughter]

>> Okay. The COPE 45 is here, guys, working

now. That is what it says in this ad,

the COPE 45.

It provides 100% backup with immediate

turnaround in a load-and-go environment

for your library of programs. So, you

know,

>> That sounds literally just like today's

computing.

>> When [laughter] people tell you to cope,

okay, when people tell you to cope, you

tell them

I cope I use a cope 45 [laughter]

for all my backup needs.

So,

another thing that's exactly the same is

everyone was worried about

data mining.

In 1968,

there was literally this thing called

the long subcommittee, which they were

reporting on

in Congress, uh where they were

>> though Casey, right? Not like

they're not just saying that

>> [laughter]

>> It's Edward V. Long, a demo- a Democrat

from Missouri.

>> Casey, I thought you were about to show

us something where they were saying,

"Guys, turns out having subcommittees

that put red tape on everything really

makes the program suck." That's what I

thought we were about to get.

>> That's also separately true.

Uh but if you take a look at this, it

basically says

um

"Citizens' names appear 3 billion times,

their age is 2 billion, their marital

status 1.5 billion, and their income

1.25 billion. And Senator Long, who was

like heading the committee, goes,

"Hearings before my subcommittee have

convinced me that whatever privacy

remains for the American citizen remains

because the federal government is

presently too inefficient to pull all

its personal information files together.

It is extremely time-consuming and

expensive to put your finger on any one

individual, but the

but the proposed federal data bank will

make it easier and cheaper to put your

whole life history no further than the

push of a button away. What good will

this technical progress be if we fail to

preserve our privacy?" In 1968.

>> Dude, Long would have lived quite short

knowing [clears throat] about Flock

cameras.

>> Oh, uh Flock cameras, you say? Gosh, uh

I wonder if those will come up in a

clipping in a couple of slides here.

So, um

software as a service, they got you

covered. Now you can lease your software

with no capital investment, guys.

Yep. They had it all figured out. They

were like, "Look, you just pay us, and

then, you know, for this monthly fee, we

will provide the software to you, and

when, you know, when you stop paying, we

take the software back. It's all good,

guys."

>> The benefits are great, though. It says

right in the article. It says the

benefits are great.

>> [laughter]

>> Exactly. Uh,

this was the year, 1968, that they

decided software could be patented,

right? There was a

There was a bill introduced that would

have banned software patents, and they

decided, "All right,

maybe not."

That was this year, 1968. So, that that

bill, had it been passed, we wouldn't

have had software patents.

All right.

Moving on. I'm just kind of going

through these, uh, uh,

rapid-fire rapid-fire here. This is more

about the the patent, like, office

there. If you look at the table of

contents,

those among you

with with eagle eyes, with the sharp

eyes, the razor-sharp eyes of a

sharp-eyed stand-up viewer,

>> Sex Party

>> eagle.

>> Sex Party

>> Yes.

>> Yes.

>> will notice

will notice a particularly

famous publication here in the April

edition.

>> computerandhance.com

>> Nope.

>> [laughter]

>> Can anyone see it? Does anyone Can

anyone out there see it? Do you see it?

>> FCC? No, I don't.

>> Number page [clears throat] 28,

How do committees invent?

>> Conway's Law.

>> Conway. That is right, T.

It is the original publication of

Conway's Law.

>> Whoa.

>> That's wild.

>> That's actually full circle.

>> How did people read it if it was more

than 140 characters, Casey?

>> I don't know. [laughter]

Uh they they were thumbing through this

weird dead tree thing and they I don't

know how it worked. It's weird.

Uh really foreign concept.

>> Mhm.

>> So, uh this was good. 27 cents for 1,000

characters. That was the price of the

>> Oh, I love measuring it in characters.

>> Oh, that's awesome.

>> Yep. Uh so, here's that uh Melvin

Conway, the original.

There he is right there.

Very dashing, I would say.

>> just thinking like hey ladies, that

guy's handsome.

>> Hey ladies.

>> I'll listen to anything he says.

>> Let me tell you about how commodities

are sold.

>> [laughter]

>> Yeah.

>> Uh so anyway, the term firmware was corn

coined uh

just the prior year and they were

talking about the concept that computers

of the future might have firmware in

them.

>> Oh.

>> In 1968.

Boy, did

>> [laughter]

>> do they ever. I I'd like to tell you

about UEFI BIOS. Uh

Anyway, that's not really quite what he

meant by

>> to tell us about UEFI BIOS in one of

these

uh in a stand-up one day?

>> I mean, there's not much to say. It's

it's it's like something like 2 million

lines of code just for, you know, the

reference implementation of the thing

that is the software on your computer

before you use any software on your

computer, right? It's it's nuts.

Firmware here, they were kind of

thinking of it as being um

Well, no, this is kind of what

I won't go into it. It doesn't matter.

That'd be like a whole rat rat hole, but

anyway, uh

I thought it was kind of fun. It was

like wow, 1968 was the first time they

were really talking about firmware. It's

the year before, 1967 actually, that I

think it was first coined.

Now, this article, an experimental 360

40 for time sharing, right? 360

[clears throat] being the the system

360, like the IBM system 360.

And 40 being the model 40 of that uh

that line of computers.

I thought this was kind of interesting

because

they already were talking about like

virtual memory and all this stuff, but

they were also talking about virtual

machines and everyone running on a

virtual machine. Like hypervisor, they

were like, "Yeah, yeah, yeah, yeah,

yeah, like no problem."

You can read up in the corner here, it

says, "If only If only software is used,

the overhead is considered to be larger

than if the supervisor program is

assisted by some additional hardware.

With this added hardware, the user is

not aware of anyone else in the system.

In fact, he cannot reach data belonging

to others. That part of the physical

machine the user sees under these

conditions is called a virtual machine,

which can, but need not, have the same

specifications as the host machine. In

our case, the virtual machine had the

same specifications as the host

machine."

So, it's like

>> That's crazy.

>> They were just running on VMs. They were

like, "Yep, that's going to be cool.

Don't worry about it. We got it."

>> [laughter]

>> Uh did you want like continuous

integration testing? No problem.

They were like, "Do you need to have

like your you know your tapes tested and

things like this?"

>> Is this an ad?

>> Yes, it is.

>> Oh my god. I just can't believe how much

of this is like right for today.

>> No, it's it

>> Based on how much of these fundamental

assumptions are wrong? Like right like

do you need your tapes tested today? We

never test tapes today, but it's like

but it's all right.

>> Yes. The the the the the like we have

done nothing it basically. Hardware got

a lot better and we just are exactly the

same as what was in a Data Mation

magazine from 1968.

Um

So, uh

what was this in like Oh.

>> of OSes. I have to I have to hear about

this because I hear agents are operating

systems these days.

>> Okay, this is a joke. Uh that one of the

weird things about this is the level of

like satire and like just generally

horsing around that happened in

Datamation magazine is kind of nuts.

We'll get to another one of these in a

second. But they would There'd be a lot

of these parody articles that are just

in there. No call out. It's not like,

"Oh, this haha this is a joke." It's

like we just wrote

>> Mad magazine?

>> Yes. They just They would just do that

every Maybe maybe every episode there'd

be like two articles that were just

straight-up humor.

And there was no differentiating them

from the completely dry, 100% serious

like the other stuff that I've been

showing, right?

>> Here and they they literally say,

"A new programming system was announced

for delivery in 2 years to be called

Assembler Stack Support or ASS."

>> [laughter]

>> So it's just like, all right.

Good job, guys.

The ASCII standard was adopted.

So President Lyndon Johnson

I assume I don't

>> LBJ stamped ASCII?

>> I think so. It just says President

Johnson. It could have been another guy

uh President of

maybe President of the Standards

Institute, I don't know. It just says

President Johnson. Uh

>> Oh.

>> Because I mean I have my petsky right

here already.

>> Okay.

>> You know, so ASCII's a big upgrade. I do

have to say.

>> I

and Lyndon Johnson maybe uh agrees with

you.

>> helped me get off the petsky and I'm

very happy about that.

>> So I thought that was kind of cool. I

was like, all right. And again,

this is not Here's all the back issues

of Datamation. This is just 1968. All of

this is just from 1968.

>> That's insane.

>> Um I I like this picture

>> year, Casey, or how often did they

publish one for month?

>> It's a It's one a month. So that's 12

issues of Datamation.

And oh, but but the things I'm showing

you, I only had time to go through the

first six again. So, it's only half that

you're actually So, only the first half

of 1968 that we're looking at here.

>> Mhm.

>> Um

I like this this piece of artwork. I

think it's very moving. That's all. The

mournful fisherman, it's called.

>> Scary.

>> Uh anyway, um so, here's more about the

privacy. Statistical data bank said to

be privacy threat.

Protecting individual privacy requires

more than controlling access to personal

dossiers, a lawyer and management

consultant told the Boston Bar

Association's Committee on automation

recently. Aggregate statistics can also

provide intelligence about individuals.

And they go on to talk like it's all

here.

>> That's crazy.

>> Like it is all here. Uh everyone they

people knew about everything that was

going to happen, but it happened anyway

might be the way to say it, right? Uh

which is kind of a bit sad uh or or

disheartening, I don't know.

Uh anyway, these are just more salary

examples.

>> If you're blocked in your career, COBOL.

>> Yeah.

>> Still good advice. Still good advice

today.

>> beyond old limits. It's an acronym,

Prime.

>> [laughter]

>> Oh, career opportunities beyond old

limits.

>> That is real good. I didn't even notice

that.

>> Jeez.

>> But hey, it's still actually good advice

today. If you learn COBOL, you can make

hundreds of thousands of dollars because

they really need people to figure out

how to, you know, maintain and move

forward these systems.

>> At at at this point in time, no one is

sure if there will ever be a time when

you when learning COBOL isn't a still a

valuable marketable skill. Like it

should have been [laughter]

that that no one knows it today.

That's not what happened.

Uh okay.

Uh using computers to try to win

elections. 1968,

right?

>> Sick.

>> Political pros applying computers to

campaign. The Democrat National

Committee recently replaced a 1401 with

a five tape 360-30 is working with the

uh with the equipment in two and a half

shifts or more just about every day and

plans to add at least one disk pack.

>> Damn.

>> The RNC is renting 25 hours a month on

an SBC mod 40 and Lester from another DC

service bureau Integrated Business

Methods. So, it's like

yup, they were already doing that, too.

Lawyers see computers as aid in handling

cases.

So, this was someone going, "Look, we

could have the computers read these

legal opinions and then they could

predict what the judge was going to do."

It's It's just kind of nuts.

>> Now, my computer is my lawyer, which is

sick.

>> [laughter]

>> Yeah.

I just hold it up. I hold up uh my phone

in the and have it speak for me at at

the thing.

>> Your honor, can you talk to ChatGPT

voice for me, please?

>> Objection.

>> [laughter]

>> Objection.

>> Repeat that. Repeat that for ChatGPT

voice.

>> It feels like a husk. This is like

turning into a husk thing with husk in

[snorts] the uh in the courtroom.

>> This

uh

writing from a reader

and it's an op-ed page.

Literally.

Like

literally, it just sounds like

an AI stan from today wrote in.

They say, "With each passing year since

the development uh called ENIAC, man has

been increasingly making computers and

automation the whipping boy for his own

deficiencies.

>> [clears throat]

>> I have been working in this field for

over 15 years and have watched in

silence as unemployment, dehumanization,

Big Brother, invasion of privacy, and

finally the very destruction of the

fiber of our society have been blamed on

my chosen field. I no longer can remain

silent. It is time we put our critics

and their criticisms in perspective.

It's not our computers to err is human.

Let me cite some examples. Several years

ago, I was associated with a project

which attempted to provide computer

assistance to medical x-ray

interpretation. Most, if not all of the

doctors to which the project was

ascribed, offered the universal argument

against computers, but you cannot

automate the years of study and value

judgment which must be based on years of

experience. Does it sound familiar? A

series of x-rays from the case history

of a patient who died was presented to

trained physicians and the computer. In

all cases, the physicians missed the

vital interpretation until the malady

could no longer have been arrested. The

computer sensed an irregularity early

enough So we were we were advised that

corrective action could have started,

right? He just goes on and on talking

about all these ways that like people

have demonized about the computers are

better. It's It's nuts. Like this is

1968.

>> This is wild.

>> It's totally nuts. Here

Sorry, yeah?

>> Uh so now I feel like I can't decide if

it's a good tweet or not, so I need your

approval before I send it. I typed it

while you were talking.

>> [laughter]

>> All right, yeah, yeah.

>> Can I So it said to err is human, you

know, cuz it's talking about how the

people make mistakes. But now I kind of

feel like we should be saying to err is

agentic. Should I tweet that or not?

What do you think? Is [laughter]

>> I'm the wrong one to ask.

>> Is agentic.

To fable is divine.

>> [laughter]

>> All right. I'm just going to say to err

is agentic. I'm just going to tweet that

only. There you go. Okay, sorry.

>> only one way to find out. You just got

to Sometimes you just got to put it out

there to you just see what happens.

>> yeah, I did. Okay.

>> Uh somebody asked about flock cameras.

I think Prime, you were saying flock

cameras.

>> Oh, I did, yeah.

>> Cuz you were like saying that other

piece, oh, sort of like flock cameras,

but it was kind of distantly related.

How about actually just flock cameras?

New York State has a development program

underway to use a video camera to

capture the image of the rear of motor

vehicles passing a vehicle presence

detector. This image is processed by a

specialized device which locates the

license plate and identifies its

alphanumeric code. The code is

transmitted to the central computer at a

remote location for comparison with

stolen vehicle and traffic warrant

files. A match results in the output of

the license number, vehicle type, and an

indication of the charge. The police

vehicle is monitoring the output device

a short distance from the detector and

is able to enter the stream of traffic

and apprehend the wanted vehicle two or

three seconds after the initial

detection. J. J. Paley of the New York

State Identification Intelligence System

expressed optimism as to its potential

contribution in the stolen vehicle and

wanted traffic violator areas.

>> Wow.

>> Do you just have to read 1968's

data whatever, the animation, you can

come up with modern-day billion-dollar

companies.

>> Just read through this and find

something someone isn't doing right now

and there's your next Y Combinator

pitch, folks.

>> Casey, this reminds me of Have you ever

seen Oh, go ahead. Go ahead, Casey.

>> No, continue.

>> Have you ever seen Colossus: The Forbin

Project?

Have you ever watched this movie? 1970.

>> That's right. I have not watched it, but

that sounds very I've heard the name,

certainly.

>> They probably wrote and, you know,

produced this during this time. Here,

Casey.

Little I'll give you a homework

assignment for once, which is watch this

movie. You'll love it. I promise you. I

guarantee it.

>> Say it one more time about Colossus: The

Forbin Project or something?

>> Colossus: The Forbin Project, yep.

>> Okay. Sorry, continue.

Don't eat in the computer room.

All right. You eat these donuts.

>> I'd be thrilled to eat these donuts.

I actually tell this to my kids. They're

not allowed to eat in the computer room.

Yeah, this is good. This is actually

really good advice.

>> Okay.

>> Residual static charges attract crumbs

to the tape surface and the crumbs cause

dents and dents create permanent errors,

guys. So, that one not as applicable to

today.

Uh there's some more software as a

service. I thought uh you guys might

like 1.5 cent uh per bit

storage cuz I was thinking pretty soon

with the AI uh data center build out I'm

imagining this is roughly I'm imagining

this ad will be accurate again.

>> [laughter]

>> I want to say Casey, I do have to say it

seems like it cost a little bit to get

memory in that time.

>> It cost just a little bit.

>> Is that a little bit Lockheed Martin?

>> Yeah, that's what I was going to say

Lockheed.

>> Yeah.

>> Um so Lockheed Martin was just Lockheed

Aircraft Corporation at that time

because they hadn't merged with the

Martin Marietta

>> But with Martin Marietta.

>> Got it.

>> Uh interesting. Yeah, that that that

wasn't true in 1968, yeah.

>> They were just making generic also like

memory system stuff too like outside of

>> Well, yeah, because

>> No, it's a you can actually see uh you

probably can't a minute it says Lockheed

Electronics Company a division of

Lockheed Aircraft Corporation. So they

kind of had like this you know what I'm

saying?

>> Yep. Yep. Yep.

>> Um so yeah.

Uh this is well o- okay, I I guess I

don't know.

I assumed this was another one of their

comedy ones. It's called a hexadecimal

pronunciation guide.

I thought it was really funny. I I

>> It has to be.

>> I just it has to be. I don't I I don't

know.

So don't take my word for I can't prove

it uh because like I said they don't

really um

they don't really say whether the

article is comedy or not. But but here

are the way that they suggest

pronouncing the the digits.

A for Ann, B is Bett, C is Chris, D is

Dott, E is Ernest, and F is Frost.

So when you encounter things uh

>> [laughter]

>> you're supposed to

>> Oh my gosh.

>> That this Yeah. Yeah.

>> Yeah.

>> [laughter]

>> Uh I'm loving that.

>> Yeah, me too. Anti, Betty, Christie,

Dottie, Earnesty, right? Uh Frosty.

>> Giffity.

>> And it's not off.

>> Yeah.

>> [laughter]

>> These are just so good.

>> Anti 07? Oh, that's crazy because on

Silicon Valley F

F5 was Flevity 5.

>> Yeah.

>> Anti Christine, that's like a that's

that's hitting home.

>> So anyway, I was like, we should do

this. Like I should try to learn this

pronunciation thing.

>> 2?

>> Yeah.

>> Oh my god.

>> A FROSTY 2 IS SO GOOD.

>> It is good. I thought maybe they were

going to be working up to like a hidden

message inside where they like pick a

hot or pretty high number and then it

like says a whole sentence and it's like

you're an idiot for reading this article

or something like that.

>> I don't I don't think it just it's

written very tongue-in-cheek so I just I

just assumed it was comedy but but I

don't I didn't have time to really read

through the entire article to like see

if there was any thing where they

literally say it but I don't think they

ever do. They These The humor was like

wicked dry in that way.

>> Can you send me a link to that one cuz I

think I'm just going to make a YouTube

video on that one. That was just really

funny.

>> Yeah, just send me send me a message

>> Send me a DM so

>> I'll remember as soon as we're done

here.

>> Okay, I'm going to send it now.

Yeah, yeah.

>> Otherwise it will be forgotten.

>> Yeah, otherwise I'm also forgetting your

forgot your forgettings.

>> Um

So again, we kind of already went over

this. Here's an entire book called

Privacy and Freedom by Alan F. Westin

that they're reviewing here that's

entirely about like

the you know, it has an entire section

on computer like what computers are

doing to privacy in that era.

>> All right, Tom. Why does it say $10 like

after the heading?

>> That's the price of the book. The price

of the book.

>> Oh, got it. Got it.

>> Which would have been $100.

>> Uh Yeah, this is a review of the

>> It's a good book. It's a $100 book.

>> Mhm.

>> Honestly, its logo does kind of look AI

generated though. If you go back to it,

notice that it has like it's an anus

with kind of arrows sticking out of

things and then one part doesn't have

arrows. It's literally claws. It's a

claw logo. You're right.

>> It's the claw logo.

>> One of It's a claw butthole, yeah.

>> [laughter]

>> They stole the logo when they read this

and stole the data. That's makes sense.

>> [laughter]

>> Claude created his own logo by stealing.

>> Um I I like the idea of asking the

boldly asking the question, how do you

gruntle the buyer of memory sacks?

I

because I've always wondered how you

gruntle the buyer of memory sack.

>> Can one define gruntle?

>> Uh yes, you can and I I'm going to use

it because it's not what it sounds like.

I'm going to use it in an inappropriate

ways in the future as well.

>> Oh, [laughter] I love this definition.

>> Me, too.

>> Oh my gosh.

>> Well, cuz disgruntled, right, is to be

upset, to be made upset, right? So,

gruntled

>> Oh, so if you gruntle you're not upset.

>> Yeah, is the opposite of disgruntled,

yeah.

>> Right, like the Eagles. They became very

gruntled.

>> Yes.

Uh Communist China

>> post-sex spiral gruntled

>> [laughter]

>> Yeah, yeah, exactly. Communist China

still posing a threat to our computer

supremacy.

Very important [laughter] to track. They

They said within 5 years they would have

uh

computing technology.

>> They didn't know China

>> That was what this estimate said. They

were 15 years behind currently and then

they would pull to 5 years behind in the

next something. And it was yeah.

>> Trash, what did you say?

>> Sorry, what?

>> What did Trash say? Trash say it louder.

>> [laughter]

>> It was a whisper joke. I was just saying

I didn't know China existed in '68.

That's crazy.

>> [laughter]

>> That's a good one, Trash.

>> No, one wasn't actually supposed to hear

it, but then you started laughing.

>> [laughter]

>> So, uh this one is not funny or really

interesting to anyone but probably me.

And that is that if you look up like on

Wikipedia, this is one of those ones

where I'm like, "Ah, man, I'm not doing

another one of these talks, but

if you look up on Wikipedia, what the

first ray traced image was,

it will tell you that the first ray

traced image was like published about in

April uh 1968

by uh a guy named uh what what was his

name?

I don't remember. Sorry. [clears throat]

Uh

but randomly going through the 1968

Datamation, here is three ray traced

images in an advertisement for a service

that does ray tracing for you.

>> Right.

>> And this is February 1968.

So,

I guess Odin being noteworthy is not the

only thing Wikipedia maybe doesn't quite

have correct in there. But I was like,

oh, that's

I wonder

where where the actual first ray traced

images came from cuz they probably

aren't the one cited

uh in that in that Wikipedia, but I

don't know. So, I I wanted to go through

the

>> very first time in an ad for a company

that's been doing and selling [laughter]

the services.

>> Uh

so, I don't know. Uh it's pretty it's

pretty hard to say, but you know,

made me want to look into it more. And

that's it. So, that's what we got, which

is good cuz I think we're basically out

of time anyway.

>> That was awesome.

>> And that's that's just the first 6

months of 1968. There's so much more.

Uh it's it's crazy. And by the way, if

you like that Coke 45

>> And two zigzags.

>> Um if you like the Coke 45, there's a a

another ad for Coke later on in

Datamation, later on in that year, where

the Coke logo is huge. So, I can try to

get that I can try to grab a screenshot

of that one for you. Just this big Coke

computer.

Um I'll I'll go dig that one up later

today.

>> Obviously, we'll have a link to the

whole

talk.

>> to your

>> Sorry.

>> I was just Go ahead, Prime. Go ahead,

you take it.

>> Okay, I was just going to say in all of

this kind of research, this is all the

side research. This is how Unbroken got

discovered. This is the same process

where you're just doing a lot of side

research for the root of the root of all

evil is premature optimization.

>> Yeah, because you know, like I said, the

to to me the interesting part of these

history talks is not like I don't really

want to tell people what to think about

the topic because that's more of like an

op-ed kind of a thing that's like, "Oh,

let me tell you what I think about, you

know, optimization or something like

that." Which is fine, you know, and

that's a different kind of article. But

when you're doing these history talks,

what I kind of want to do is I want to

go feel like I kind of know the time

period. I want to as much as possible

like put my brain back there by reading

all of the things that were going on at

the time and and stuff like reading

their trade magazines and all this sort

of stuff. Uh and then try to give you a

feeling for like this is when this thing

was said. Like this is actually what was

going on at the time. This is where

these people were in their lives. This

is what they were thinking about. This

is how their personal lives were

whatever, right? Just to kind of like

put you back there. And I to me it's

really really cool to have that

experience. And unfortunately,

I only can get a little bit of that

across in the talk, right? Like I I get

some of it across as best I can, but you

know, if I go back and look at the talk,

I'm just like, "Yeah, I I only got a

little fraction." Like I tried my best,

but you can never really give the full

impression of it. Of when you actually

go back and like read all this material,

you get such a much more palpable sense

of like, "Oh, okay. Like I see what what

life was like and what they were

thinking about." Right?

>> Hey guys, if you like this episode, you

can watch the rest of it on Spotify. And

don't forget to like AND SUBSCRIBE. WOO!

SEE YOU LATER.

>> [music]

>> Fire code, errors on [singing] my

screen, [music]

terminal coffee

and heaven

living in a dream.