Welcome to the stand-up. Today, we have

a very special episode. You may notice

that it looks lower quality video. That

is because we have nobody here to

record, and I'm just raw dog recording

the screen straight from Riverside.

Uh anyways, I'm sorry. With that, we're

going to be talking about Microsoft

layoffs. Now, Microsoft has just

recently released an entire kind of

exposé of exactly what happened. This is

by Asha right here, the new CEO of Xbox.

She has let everybody know that they are

laying off a whole bunch of people, a

bunch of reasons in here, things that

they want to accomplish, including a

bunch of crazy pieces of information I

did not think we were going to learn,

along with also crazy management style

of Xbox itself, and a bunch of studios

and people have been hit today directly

cuz 1,600 were laid off and 1,600 more

throughout the fiscal year of 2027.

Also, really weird way to measure time

is the fiscal year. I'm just throwing

that out there. I think that's just

weird personally.

Uh but at the end of the day, that is

what has happened, and we're here to

discuss it. And I figured one of the uh

big burning topics that I see all over

the internet is about id Software. id

right now is kind of like the uh

the star of the layoff, if you will.

Approximately 10% of the total layoffs,

132 out of 1,600 of the layoffs, I

believe it's 132 or 135, came from id

Software. Most of the engine team has

been gutted, a bunch of designers, a

bunch of people running it. Uh

effectively, the the heart and soul of

id has been ripped out, and now what is

left is just whoever's remaining there

right now, and it would probably be

impossible if uh or improbable, if not

impossible, to be able to make anything.

And so, I first wanted to talk a little

bit about that because I am not in the

video games industry. I don't know what

it takes to create uh an engine. I don't

know what it takes to maintain an

engine. I don't know what it takes to

move forward an engine. And we do have

with us a legendary game programmer,

Casey Muratori, who does engine work in

fact. And so I would just like to hear

some of your thoughts maybe from the

engine perspective of like how does it

is it just doomed to become yet another

Unreal 5?

>> Was that a pun?

Was that a Was that a pun?

>> It actually wasn't a pun. I didn't even

realize it was a pun until you said it,

but is it doomed? To become another

Unreal Engine game?

>> Bef- Before we attempt to to answer that

question, I wanted to talk about

something you mentioned in that lead-up

as well, which is you were pointing out

like fiscal year 2027. I

I don't know about you guys, but in my

mind, I imagine like Satya Nadella and

Asha Sharma and I don't I mean I guess I

don't know who runs the other divisions

there. Do they like get together for a

fiscal New Year's party? Like whenever

this year Like do they all Do they like

stay up till midnight and then they pop

the champagne and they're like, "Happy

New Year! Happy fiscal New Year!"

>> Like when it rolls over, they're like,

"New filings, SEC!" And they press it

and everybody freaks out and

>> And then they sing a song that's like

Old Lang Syne, but it's about, you know,

uh accountability margins or something

for so you know, or whatever.

>> Teach, production's down.

>> I don't concern myself with such

matters.

>> What do you mean production doesn't

concern you?

>> I've shipped 37, nay,

38 features today.

Teach always makes the [music] mess and

I always clean it up.

>> I'm better than that. I don't have to be

a janitor.

>> Someone has to be the adult around here.

>> I

>> Who Who is that?

>> It's me, your Westlake Godfather.

>> John John Carmack?

>> What? No, it's me, Trash.

>> Oh.

Wait. Are you here to finally get

management to understand software?

>> What? I'm a Westlake Godfather, not a

miracle worker.

>> Oh. Then why are you here?

>> I'm here to tell you about Sear by

Century,

>> [music]

>> the world's first AI debugger that

actually has access to all of your logs,

stack traces, code, commit history and

more. Code breaks, but you can get the

fixed [music] faster with Sentry.

>> Who are you talking to?

>> What?

>> Build with AI.

Fix with Sentry.

I think there's a couple of problems

here with trying to answer it. Uh we can

talk about some abstract stuff, but the

the biggest problem or the thing that I

don't know is we have like no like I

don't know any inside stories or

anything from the people there and what

happened, right? So everything is on the

table. Like everything from

uh literally they just got handed layoff

notices. So like, you know, a

substantial portion of the engine team

just wakes up in the morning to like a

thing where like someone from HR told

them

that you don't have a job anymore.

All the way to like they were approached

and said, "Look, we need people to do

programming on stuff because we want to

build out these other properties like

Minecraft and Forza or what Well, not

Forza, that playground has their own

engine team, but um

you know, these other properties we have

King like we want to put put you on

these other things and you'll be a

centralized engine division. Um and they

were like, "We don't want to do that. We

would rather take voluntary layoff or

something." Like, you know, I literally

just have no idea. So I don't feel like

I can really comment on anything

specific to the id software layoffs with

layoffs with no with no data, right?

Because there's there

it's not good. So it's not like there's

any out There's not there really isn't

probably anything where I would be like,

"Yeah, great." Unless it was they were

approached with something and they all

decided, "Screw this. We're going to go

form our own company." and took a

voluntary and and now we're going to

have like a new like id thing somewhere

else. I suppose that would be good. Um

but short of that, there's there's no

real like way to spin this as a good

outcome probably, but how bad this is

depends a lot and how

um

how like frowned upon I would say that

the behavior by Microsoft was is is a

sliding scale based on what really

occurred and I just don't know. Is that

fair?

>> Yeah, yeah, yeah, that's fair. I wasn't

really trying to look for like the the

moral side of things. It was more of

like is it possible for id to continue

being id and being, you know, its own

game engine producing these kind of

technical masterpieces as I often see

people claim the Doom series is. Uh is

it possible with the gutting or is it I

mean just in your experience as a as an

engine creator, how hard is it to move

forward when you have a significant

portion of your team gone?

>> I mean it really depends on what you're

going to be doing and like what that

what the moving forward

requires, right? So, the first thing

I'll say is that in sort of modern

in modern engine development

usually the the bottleneck, if you have

good engine people, the bottleneck is

not like that we had to sit down and

write some code for this engine.

Um in other words, the stuff that Fable

does or whatever, right? Uh usually that

is not going to be your super the hard

part of the problem, right? The harder

part of the problem is usually on two

other on like two other

um

in two other domains. One is just

figuring out what actually should be

built at all, right?

>> Yeah.

>> And this is basically active research

into like what is our strategy going to

be for like dynamic lighting? What's our

strategy going to be for like level of

detail? Like those sorts of things.

Uh those technologies sort of they

change shape over time

because the underlying capabilities of

the hardware change, and so every so

often there are like you have their dis-

discontinuous. So, you know, previously

we were doing things without ray

tracing, and then we get enough hardware

to make ray tracing make sense, or we

make enough software progress to figure

out ways to do ray tracing in ways we

weren't doing it before that make it

economic

uh to do so.

And now we have to like re-architect the

way their engine is working. And that it

again

well, sometimes and certainly for

novices it can be the case. Uh

sometimes, you know, you look at that

and be like, "Oh, well, so it's all

about writing this code." No, like

that's not the hard part for experts.

The hard part is knowing what you are

supposed to do. What are the algorithms?

What's all the math have to be? How does

it all work out cleanly? How does it all

move through the system at a at a fast

enough speed? All of that stuff requires

a lot of experimentation, a lot of

testing. And if you look at stories from

people who have done like the sort of

big technical pushes, there's often just

a lot of that. It's like, "Well, we

tried this, and we tried this, and we

tried this." You know, you could go look

at the um presentation by the guy who

made Nanite, for example, and you'll

hear that. You can go look back at John

Carmack's stuff, or Michael Abrash's

stuff

uh that he put out uh when they were

working on Quake. There's just a lot of

like, "We had to try this, we tried

this, we tried this. We finally found

that this was the right thing, right?"

And so,

um

that requires currently uh a lot of just

people. Because for every one of those

that you want to try to do

experimentation on, somebody's got to be

sitting there doing it. Uh and even if

you are buying into the new like AI

future, whatever, uh AI still need a ton

of guidance for that kind of work. They

need somebody there to like tell them

what to do constructively, and to build

test cases, and to say, "I need you to

try doing this." And then to look to see

if they actually did that right, or if

they they screwed something up, you

know. And so

>> Let me jump [clears throat] in there for

a second, too. Even if you could do all

the looping and all the fancy words that

they say, let me give you some AI speak.

Even if they could do all that looping

and all that, the actual getting the

data out and all that could take some

amount of time. And so that loop itself

could for each one of these experiments

could cost you 10, 20,000 dollars worth

of stable time plus

>> Sure.

>> two weeks of actually running,

gathering, testing, running, gathering,

testing and test. So it's like it's not

a it's not just like a oh, just put AI,

it just makes it free. It's like no, it

could actually also cost a large chunk

of money plus the engineers' time.

>> And and I just like I haven't seen a lot

of evidence that there's any kinds of

significant speed-ups happening yet in

that domain where it's like oh yeah,

like we can just now you know,

some studio got it to the point where

like one engine coder can just kick off

tons of agents and run all the

experiments and it it works out. Like

maybe that's a future that can exist,

but like no one has that at the moment.

And so this So just to answer your

question, so

So

depending on how much of that you

envision doing

that changes pretty dramatically the

kinds of budget that you're going to

need for the technical side, right? And

that's like prong one.

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

you can watch the rest of it on Spotify.

And don't forget to LIKE AND SUBSCRIBE.

SEE YOU LATER.

>> BOOT up day

five [music] code

>> [singing]

>> errors on my screen

terminal coffee

>> [laughter]

>> and hair

living the

dream.