Welcome to the standup. Today is a

extremely special episode. We have Adam

Ununice. We have Nolan Eie, EIO, and of

course, we have the master vibe coder,

Tee. What up?

Yeah. Yeah.

Uh anyways, sorry.

>> You guys mentioned offline not part you

know before we recorded just null and

some of how like you got in introduced

to making games and part of that was

through Adam which is part of the fun of

like we wanted to have you both on at

the same time. So I thought maybe you

could talk about that a little bit and

just uh give a little bit of that

backstory and just uh so you know so the

listeners out there could hear that side

too. Uh yeah, like back story to like

how I started making games or like and

any complaints about Adam and how he

teaches a sprite. You can throw that in

there if you need.

>> Both of those.

>> Uh let's see. So I started making games

maybe like three three years ago, almost

exactly three years ago. Um and uh

originally like this was not really a

thing that I was planning to do for

particularly long. Um it was mostly a

way to get back to coding. I love games.

I played tons of games my whole life,

right? Used to write about games. um and

started making games like really small

games in like GDAU. I think like one bit

of advice that you see a ton on the

internet. I think it's really hard to

give good advice about about game

development. But one I think very

broadly applicable often repeated bit of

advice is just make a ton of games,

right? Like make you know don't like the

classic mistake is that somebody makes

an MMO that takes them four years and

nobody plays it and you know they learn

so much along the way. Um, so I was

making these tiny little games in GDAU

and found Adam stuff because I was try I

like made all this pixel art and it all

looks terrible and I was like so like

there's this weird thing when you're

you're bad at something where you don't

understand why like you don't even have

the the like faculties to understand why

it's bad. You're just kind of like

looking at it and you're like this is

this is terrible and I don't know the

first step towards fixing it. Um, and I

never got I never got good at at pixel

art. I ended up kind of moving away from

making more more typical games into the

weird stuff that I do now. Um, but one

thing that I remember really enjoying

about Adam's channel, uh, and sorry Adam

to to compliment you live, is I think a

lot of the stuff that you see on YouTube

about making games is super

edutainmenty, right? like the audience

of people that kind of want to imagine

themselves as game developers is much

larger than the set of people that

actually have the time to sink into

making a lot of games. And I think also

probably the audience of people that

wants to talk about games is larger than

the audience of people that actually

makes games and can talk about those in

a compelling way.

>> Um, and so I think early on I was

watching all this other content, right?

And I would watch this video and I'd be

feeling like I was learning a lot and

then I'd try to think about like what

did I actually learn from this? And the

answer is like nothing, right? like

nothing applicable to making stuff.

>> But when you watch people that are

actually doing the work and talking

about their decisions, I think you learn

a lot more there.

>> Um, so that is that is how I ended up

finding finding Adam's work and I made

slightly better pixel art. Learned

things like your pixels should be the

same size and to use a consistent color

palette and other dead simple things

that I think are in literally my first

blog on the site. Um, but uh yeah,

that's that's that's the background. And

I think none no one played any of the

games that I made in that era. Not

literally nobody, but like

>> like I got like a hundred players for my

second game and I was like, "Oh, this is

this is incredible."

>> Mhm.

>> Broadtown, what happened? I don't know.

Did you push on a Friday?

>> Never. How are we going to figure this

out? How are we going to figure this

out? Did

>> I hear Broadtown? Do you guys need the

wheel?

Nev config.

>> That makes sense.

>> I'm going to spend a couple hours

refactoring. I'll leave the plugins.

>> Don't guess where your issues are. You

can see exactly where they are happening

with Century. Get all the context you

need to debug any problem because code

breaks. So fix it faster with Sentry.

It's funny because that's also basically

the way that I found Adam was I was just

playing around with making some games.

Uh just like the concept of making a

game and then I found Adam's channel and

I also really liked the art and then me

and Prime somehow found ourselves in a

scenario where Cursor was like, "Yeah,

you guys can make a video game. Go ahead

do it in a water tower.

>> Um just put the game in the bag, bro."

And I was like, "Well, I don't know what

to do." And we were both like, "I don't

know. We should." And I was like, well,

the coolest person I know that makes art

for games and makes games is Adam. I

don't know him at all. I've only watched

his YouTube videos. And he happened to

be in the US. What? In the first time,

since when? Adam?

>> That That was a year ago today, by the

way. That was literally today. A year

ago.

>> No, one year.

>> Oh, yeah. Yeah. I was right in this

area. Oh, yeah.

>> Oh, that's crazy.

>> That's cool.

>> Yeah. I was in I was in San Francisco

for GDC. Um, and you just you guys just

reached out and I was like, "Oh, I'm

going to be around." And I hadn't Yeah.

I hadn't been in the US since 2011.

Yeah.

>> Yeah. And then it was like, "Okay, well,

do you want to fly down and make a game

in a water tower with us? We're doing a

24-hour stream." And

>> and uh and that's how the friendship

started, which was awesome. So, Ken

second to Adam being really good at

making YouTube videos. If you guys

aren't following him, you definitely

should. and he makes really good pixel

art. It's addictive to watch.

>> Yes. Very true. Very true. He's the

pixel king. Okay. So then Nolan, so you

were starting off, you found some stuff

there and then like how did you kind of

transition out of like making some more

like GDO based stuff into like making

crazier things?

>> Yeah. So uh do you all know what the

Rehearsse Center is? Feels like

something you might know about.

>> I don't think so.

>> No. Okay. This is going to sound like

almost like an ad. I swear it's not. Um,

so early like middle of that year, I

went to this place called the Recurse

Center, which is like a thing in

Brooklyn that's kind of like a writer's

retreat, but for programming.

>> Oh,

>> like you show up for like 6 or 12 weeks.

It's free. They like make money with

like placing people in jobs or

something. It doesn't really matter.

>> Uh, but the point is like you're in this

room with other people that think it

would be fun to program for 6 or 12

weeks on stuff that isn't for their job,

right?

>> And I'd known about it for a while. It's

like 20 minutes from me. Um, but I was

there and I started kind of

experimenting with like doing weirder

things. I remember the first thing that

I did uh that was stranger was um do you

know the open search spec?

>> No.

>> Um this is okay. So you know like in

Chrome if you type in like a website URL

and you hit tab.

>> Yeah. Yeah. Yeah.

>> And then it like lets you search that

website directly, right?

>> Yep.

>> Uh so like you need a spec for that.

Like the browser like how does the

browser know what endpoint to hit in

order to like search that website? Open

search is that spec. It's just an XML

document you throw uh you mention it

like the the head of your your index

file

>> and um yeah no I mean it's old right

there's also like 14 things called open

search but this is the open search spec

>> so I happen to know um or I kind of like

remembered that there's this extension

to the spec that lets you offer search

suggestions

>> uh so like you can offer an alternative

endpoint and you like hit that endpoint

and uh the endpoint returns back like

suggestions for what you might be

searching for which makes a lot of sense

right

>> did you make a dungeon dungeon crawler.

>> I didn't make a dungeon crawler. Um, but

I was like, you know what? Like all you

need for a game is like you're typing or

like you have some input mechanism and

some output mechanism and this works,

right? Like you type a character, it

hits this endpoint, the endpoint returns

suggestions. The suggestions are

automatically displayed in the browser.

Like you can make a game there. And so I

made Wordle

>> um which I called I called Wardle uh not

knowing that that is Josh Wardle who

created Wle's last name. Um,

and then like met him like a year later

and that that was funny. But um, so I

made that and I was like in this

environment where instead of just kind

of making it and shipping it off into

the void, there are all these like nerds

around who are like, "Wait, this is like

this is kind of funny. Uh, this is this

is compelling." Um, and someone posted

on Hacker News. It's the first time that

I got on HackerNews.

>> And I think like that is how I started

kind of narrowing in on like, oh, I like

like it's funny to make computers do

weird things. It's funny to make jokes.

Um, and so I started noticing like I

went back and I made some more

traditional games

>> um, while I was there and I remember the

last thing that I made it was like a a

tiny little uh, Vampire Survivors clone.

>> Oh, nice.

>> Um, like you know, people were like

doing all sorts of things like Right.

Cuz Vampire Survivors had just popped

off. Yeah.

>> And I was showing it at this gameplay

test and people were giving me feedback

and I could like feel myself cuz I had

made some more weird stuff by then. I

could feel myself like not like being

like I don't want your feedback. Not cuz

like it's not good feedback, but because

like I'm in I don't want to make games

like this and like you're giving me

advice on like a thing that like I wish

I hadn't made cuz like everyone around

me it was at this uh this the bar that

we were talking about earlier.

>> Yeah.

>> Um had all these like weird interesting

experimental things

>> and I'm just not like I'm much better at

that than I am at building like a

beautiful wellthoughtout you know kind

of within genre game. So, that's kind of

where I uh I think that that was the the

thing that really pushed me towards

getting weirder and weirder and weirder.

>> I do have to say I thought it was the

most programmer description of a game of

all time was well, you just have some

inputs and then you have outputs. So,

it's like that's a game. You know,

>> it is. So, this is I think about this a

lot with like the the stuff when I'm

making a game that is running in another

application. So, like that or like I

have like Flappy Bird running in Mac OS

Finder or whatever. Um that is how like

the terms that I'm thinking of you like

find like a surprising way that an

application gives you input and output

like for example in Macfinder I can

query to determine that you have clicked

on a file

>> um and like that is that is sufficient

input right and then output can be

renaming files

>> um this is a terrible way to think about

like real games that people want to play

but it's a great way to invent games

that you can write blog posts about that

people read

>> which hey if That's uh that's a whole

new genre of games, but it it does work

and they are funny, so it makes sense.

>> Yeah. I think I think if I made if

that's all I did, I would get a little

sick of it. But it's really fun

sometimes.

>> Yeah.

>> Is there like a do you ever have like a

like a a personal win condition for a

game? Like is it, oh, I need for the

player to feel this, or is it more like

a personal expression thing? Like how do

you when do you know when something's

finished?

>> Oo. Oh, that's a good question. Um, and

I would I would love to hear hear your

answer uh to to this as well. Um,

I think it depends a lot um for like the

bigger multiplayer experimental stuff

thing. I think it's kind of vibes like I

think a lot about like how are people

going to ruin this? And I want to be

really confident that it's hard for

people to ruin it, right? because that's

just like

>> you get kind of get one shot and if like

a bunch of people go to the website and

it's just like slurs and porn or

whatever, you know, depending on how

people want to ruin something, it's just

like not that's just not fun.

>> That's very typical internet, too.

>> Yeah.

>> And it's Yeah. And it just like it just

happens, right? Like this and this this

sucks, but um if you don't do something

about it is what people do for kind of

the smaller like

>> the way I think about this is that I

make games and then I also make things

that aren't really games. Like the the

Wordle thing, no one's going to play

that. that exists so that I can write

the blog about it, right? And teach

people something interesting in this

cart mechanism.

>> Um, and there

I think it's kind of when there are no

more problems left to solve. Like I get

to a point where I'm like this

technology is not meant to run again.

Like it's, you know, Mac OS I pushed Mac

OS Finder to four frames a second. I had

to add like double buffering to it. It

was a lot of work and like this is like

where we call it. um I can't make it any

any faster. And so that's normally what

I'll do. And sometimes I get to that

point and like the game just isn't

interesting enough and so I just don't I

don't post about it.

>> Do you have like a lot of like ideas

that you started that just didn't just

didn't go anywhere?

>> Uh not that many. Like I'd say like more

often than not I finish what I start.

But

>> that's great.

>> I think it's like close. You know, it's

probably like 6040 or something.

>> Wow. That's a much better ratio than

anybody I've ever met in my entire

lifetime.

>> I was gonna say though, Adam, you can't

get away from the reverse question

though because Nolan wanted to hear your

answer and I want to hear your answer

too. So,

>> yeah. I'm also I'm curious. I know like

you've both done like game jam game.

Like I remember you you placed really

well at a at a Ludum Dar a while ago and

then you also have these really long

running things.

Yeah, I mean I I am such a like a uh

boring case of a gamer because like I

literally just have that the exact story

everybody else has of like you know I

grew up and I had a Super Nintendo and I

was obsessed with that and then I got a

PlayStation and I was obsessed with that

and you know by the time I was maybe

like 11 or 12 I was like I want to make

games. This is what I want to do. So, I

think I just for the big game, it's

really just me trying to channel the

feelings that I got playing like Final

Fantasy 10 um into a project and wanting

other people to feel that. And so,

because it's such a conventional shaped

thing that I'm making, it's done when

it's finished, you know, when I can get

to the credits and you can, you know,

it's got 20 hours of content and side

content, blah blah blah. Um, for the

smaller games that I feel like comes a

lot more from

a more like I I studied design at

university, so the idea of like hacking

stuff together and prototyping and

wanting to explore a concept um that get

that itch gets scratched by the smaller

projects. And so yeah, I feel like they

are two very different modes. Uh, and

I've tried sort of in the middle as

well. I had like a couple of six-month

projects. Um, those can be fun too. I

think the the tough part is then like

when a project needs to become a product

and you have to like support it as a

piece of software or something like

that, it becomes there's like a

threshold there you have to cross over

uh when you release something and the

more time you spend on it, the more it

has to sort of justify the time you

spend. So yeah, I tend to stick to one

end of the spectrum at a time. Yeah.

>> Yeah. I mean, I think like game jams are

so nice for the like I'm just going to

express this one idea, right? And then

like maybe a lot of people find it

compelling and you do something else

with that, but like having the freedom

to do that without the like

>> I worked on this for six months and so I

feel like I should get something out of

it

>> type of thing. It's really Yeah.

>> Oh, speaking of just as a fun aside, our

buddy Cakes, we should link it prime in

the description of this video, just he's

a longtime Twitch streamer friend. He

just released his tower defense game.

He's working been working on it at

Twitch for like four years. We should

just give it a shout out. I I am giving

it a shout out right now. The link will

be in the description for other people.

This is not an ad. It is just he's a

friend of ours who raids us a lot on

Twitch and I want to make sure we're

shouting him out everywhere cuz he

finished a project odd stream after four

years. So, it's awesome. Uh so people

should go check that out. That just

reminded them.

>> Cupcakes.

>> Yeah, cupcakes.

>> Yeah. And yesterday I was uh I opened up

Steam to play uh a game that you may

have heard of, Slay the Spire 2. And

when I opened it up, me and my son are

doing a little co-run together. And when

I opened it up, it said it said, "Hey,

Tower Defense Mania or something like

that." And I was just like, "See more."

And it said, "Here's a bunch of games of

Tower Defense." And it was like

balloons. And then Tangy was number six.

>> Nice. That's sweet.

>> Yeah.

>> Yes.

>> Stoked. So anyways, that was just a

quick aside that I wanted to make sure

because I was thinking, "Oh, we have a

friends on who make games. We also have

to talk about our other friend who just

released a game. So, um, anyways, Adam,

the question I was going to ask you

though was from any of your like smaller

projects where you're working on like

Game Jam or something else, you take any

of those ideas, put them in Insignia,

like any good examples.

Uh, actually, there was one recently

that's not from a game jam at all. Uh,

we we had like an onstream game we used

to play uh called Well, we'd call it

just clinkers. It's like a there's a

food. It's like a candy in Australia

that is covered in chocolate and there's

only three there's like three kinds in a

packet but you don't know what you're

eating because until you eat it because

it's covered in chocolate.

>> Uh so we we just use it as like gambling

in the stream. So people redeem points

and I bite into one and they guess and

if they guess correctly I put an Easter

egg in the game and we've we've we've

decided that uh we're not doing it

anymore because um one I don't think I

think they're phasing it out but also

like I can't stand the taste of them

anymore. Like I

over like six years now it's like

killing me.

>> So um it's in the game now or at least

it will be. we're gonna we're going to

put like an in-game uh it's like a one

in three chance thing.

>> So that that happened. Uh there were

times when I would do a game jam like

we've thought about specifically working

on like an insignia in-game miniame

through a jam.

>> Mhm.

>> Certainly like yeah a few different

ideas have sort of crept in actually.

Yeah. Like one of the main mechanics in

the game came from a jam.

>> So

>> Oh nice.

>> Yeah.

>> Say more about that.

>> Yeah.

>> Okay.

uh a game I made called Trauma back in I

think it was 20 22 maybe or 2020. One of

those one of

>> I think my therapist told me about this.

>> Yeah. Yeah. That's what it's it's kind

of like it's a platformer where you play

as a a a girl who's in therapy and the

way that you play it is it's a

platformer and you get like um abilities

that send you off in different

directions in the screen.

>> Uh so like to clear a jump like she

can't jump on her own. she can only move

around and you have to touch these like

specific like trauma points to like read

about or like to like prompt

conversations and then once they're kind

of like little puzzles and once you

solve a puzzle

>> uh and you kind of like heal the trauma

>> um you get to like you get like a card

and then you play that card to use the

ability. So, like at first you'll look

over a little node and it'll like throw

you away, but then like after you've

solved the little puzzle, like you'll

use that node and the card you get from

it to like this is a jump card or this

card thrusts you forward or and by the

end you get like six or seven cards and

you can chain like a really long jump

sequence just you're just pressing space

to play the cards,

>> but you're starting at one point and

sort of like clearing this crazy like

Celeste style gap. Um, so yeah, that the

ability of like like the concept of

chaining like directional style

abilities into like one big platforming

leap. Uh, that's kind of like what

Insignia's um ability system

>> kind of Yeah, that's how it works. So,

>> all right, I got I got I got something

for you, Nolan. Uh, so I my favorite

blog that you wrote was the 1 million

checkboxes blog. And the reason being is

that as you started off, you started off

with this like really simple concept. I

want to send a million checkboxes down

that people can like look at and and

click and it's a very like approachable

and consumable idea. You're like, okay,

I I get everything you're about to say.

But then you start going, okay, well,

how do I send the data down? Okay, like

this is actually a really big problem

because a million of anything is a big

is a big number, right? Like it's just

it's just a big number to send down. And

so I can't just send down like a JSON

map or an array that's encoded in in

JSON because it's just going to be huge.

And so you start kind of going through

this process of like working through and

sending down the data through binary

stuff through, you know, I forget if it

was websockets that you sent it down via

or whatever it was. And you ended up

having to pack everything down into

bits. And then you talk about, okay,

well at first my infrastructure was in

Python and then that just was not going

well. So, I switched to Go and things

went way better for whatever reason.

Magic, I guess. The magic of compiled

>> language. I don't know. I don't know

why. Yeah,

>> it's crazy how compiled languages are

just so much better at that kind of

stuff. And so, I just I love that

process of going from I'm going to build

an idea to this is actually how an idea

can realistically execute.

>> Uh how how often at this point do you

still run into a situation where you've

learned as much as you did say on a

million checkboxes? Cuz now you've seen

a lot. You you've seen some [ __ ] And so

now it's like, okay, well, I'm not going

to start off in Python on a single box.

I'm going to, you know, you kind of like

you pre-plan for some things to come

down.

>> Yeah. So, I've been thinking about this

a lot like the firefighting like that,

right? Like I launch one million

checkboxes. It's like a single small VM.

It's running in Python. It's like, you

know, people are going cra you know,

there are hundreds of thousands of

people trying to check boxes.

Firefighting like that is just really

fun, right? like you also learn a lot

but like it's super motivating right

like you learn exactly what you need to

do

>> to keep the site up for the next like

two hours

>> whereas these days I'm you know just

thinking a lot more about like okay like

let's imagine a number of people like

for the snake game right

>> I made like a game that you play over

SSH it's like a massively multiplayer

snake for something like that I'm like

all right what's a number that's insane

you know 2500 people are not going to

concurrently play my stupid SSH snake

game so let's make sure that we support

that out of the box and then like I did

that and like it we could I could

support that. Uh but it might have been

more fun to like launch that game

without really thinking about

performance at all and then seeing what

what broke. Um so I think like I don't

I don't want to say I don't I don't

learn anymore or anything like that. Um

for for the chess game that I made last

year, I still I still learned a lot. Um

but it's harder, right? like

in a way I'm pre prematurely optimizing

right and it's like hard to say because

I am putting things out on the internet

there are a lot of scary thing like I

don't want like a massive bandwidth

build that I wake up to or something

like that so there are things that like

you actually do really have to think

about

>> but

I don't know I kind of feel like I

should be moving a little bit faster

especially since sometimes the things

that I'm preparing for just like truly

don't matter or like they are just

content right like I think the decisions

that I made for the snake game that I

made recently were primarily because I

just wanted to learn a lot about how to

write a super efficient terminal

application. But I had to like make that

choice. My users didn't make it for me.

>> It was very it was a very good blog by

the way the SSH stuff. I really enjoyed

>> how you put together especially the unic

codes and the vertical the vertical

movement classic first problem of any uh

terminal game.

>> Yeah. Yeah. Yeah.

>> I thought you were going to say you

really liked how we gave Terminal a

shout out. That's what I thought you

were going to say for that was why you

really liked the blog post.

>> It was a good shout out. It was It was

real.

>> It is It is funny though like cuz you

would think, you know, from just like,

"Oh, I'm going to make something that

runs in the terminal." You're like, "Oh,

it can't take a lot of bandwidth or it

can't take a lot of data." But we ran

into this too in some of the early stuff

for terminal shop. And we ended up not

having to fix it to the same extent

because it's not like updating as often.

It just updates when people take action

so we don't have to redraw.

>> Um

>> I think at one point we were sending

down like the first hour it was like 60

frames a second no matter who was

coming. The thing that we did initially

that made us send an insane amount of

data was we had a big a huge React Miami

logo and it was saying like coming soon

and so you SSH in and it was just like

you know coming soon react but then we

had something in there that was like

spinning or some something that made it

so we had to like send a refresh or no

you know

>> it might have been my coffee cuz I did

build a particle system.

>> Uh no pretty sure we never put that in

the application. Um, so, uh, I actually

think it wasn't even we didn't even

redraw it. It was just something with

bub, uh, with one of the like charm

things where we had a timer running

somewhere. So, it just forced a refresh

of like their framework. So, it was like

counting down the seconds until we said

we would release. And so every second it

would send down or like a bunch of times

a second. I think the default ticker

would send out this huge I don't know

because we made you have to zoom out to

look at it because we wanted the picture

to look

>> it was like 10,000 by 2,000 something

like that. It was like huge

>> and then and well we took a picture and

we did the like picture to Ansie thing

right. So we took a physical picture I'm

saying right and so it had like a

million terminal escape sequences in it.

Right. Cuz it was like every every

character had its own to like make it a

different color and make some of them

bold.

>> Slowly going from pink to more pink.

Like it was just it was the worst

possible case.

>> So every every character had probably

like at least you know 10 to 15

characters we were sending, right? And

then we made you zoom out to be like,

"Oh, we can't show this at, you know, 40

by 80 or something like that. You had to

be at like 100 by 800." And then we just

were sending that like 60 times a second

for no reason. It was a static image.

>> I just remember that like Dax comes out

like, "Guys, we sent like a 100

gigabytes in the last hour." And we're

just like,

>> "Huh, how do we do that?"

>> At that point, we were like, "Is the

whole selling coffee over SSH thing, are

we just screwed?" We were like, it's

like maybe we can't even build this as

an application.

>> It's the coffee that's the problem.

>> It's the coffee that's the problem. I

thought for sure the problem with our

margins was going to be trying to get

good coffee to people and physical

products, but it turns out it's just

bandwidth.

>> Bandwidth.

>> Just bandwidth. It's just egress cost

from our container. Um, we did

eventually figure out, oh, when the time

when there's a timer running, it resends

this. So, we fixed that eventually, but

I was cracking up about some of it

because you just can't even imagine that

you're going to have this bandwidth

problem over SSH because it seems

absurd.

>> It's text. It can't be that big.

>> It's text. Turns out it can be. So, I

don't know, Nolan, if you want to talk

about uh for people who didn't uh who

haven't yet yet read the blog post uh

some of the like fun stuff you did for

the the snake game.

>> Yeah. I mean, my favorite thing that I

did was that I like patched the Go SSH

library.

>> Uh which I wrote like a whole a whole

separate library or a whole separate

blog post about because this was so wild

to me. Yeah.

>> Um, but like the gist of that, right, is

I'm working on the game, working on the

game, doing all this profiling because I

want the game to be super fast. And I

like on

this change and my CPU usage drops by

like half and like at this point I'm

running like I have like a harness with

like 500 bots that are connecting,

sending moves. Um, and uh, yeah, so CPU

usage drops by half and I'm like

thrilled, right? Like I'm like, "Oh man,

like this changed like I've done

something awesome."

>> So I start I'm exploring that. And I'm

trying to figure out like why cuz this

was not a change that was supposed to

make things faster. Like why why did

that work? And I realized like actually

all I had done was I had broken my

harness. Like instead of my bots getting

data, um they just got a single message

that said your screen size is too small.

>> Um I had like broken cuz like SSH

there's like TTY like whatever. You

know, the details here don't really

matter. But I had broken my testing

harness and that's why performance

improved because I wasn't really testing

my game. But then I'm like thinking and

I'm like, wait, if my bots weren't

receiving any data for my game, why did

performance only drop by half,

>> right?

>> Like why did performance not drop by

99%.

That doesn't make any sense.

>> Mhm.

>> So I start capturing packets and like

looking at all of that and eventually

like see that if you're just SSH to a

box and you press a single key, SSH

sends, you know, like about 90 packets.

uh and what's going on there is that um

there's the the SSH is doing something

called keystroke obuscation. The basic

idea is that if I see what letter or

what keys sorry the timing at which you

are pressing keys but nothing else like

all I have is like key press key press

key press pause key press well that

pause maybe that means that you're

reaching for like uh you know the number

pad or something like that like you're

pressing something that's harder for you

to get to.

>> So SSH just sends junk data all the

time. um to like obfuscate the actual

key presses that you're making. But it

sends like if you the way it works is

that it perpetually sends this junk data

as long as you whenever you press like

one key.

>> And that's really bad for games because

you're pressing like one key a second.

The junk data gets sent for about one

second. And so it's like one key press

turns into 101 packets. Um and this is

not something that you're supposed to be

able to disable. And I just like I found

the patch and I like read through the

the SSH code and was like if I just

pretend if I like claim to the client

that I'm just running an old version of

SSH that doesn't know about or of SSHD

the the you know serverside component

>> that doesn't know about this feature

maybe they won't send these these junk

packets to me. And that totally works.

>> Um so I did that and then I posted about

it and a bunch of nerds got really mad

at me because I wrote this whole blog.

Yeah. And I didn't acknowledge like the

blog starts out with like I am working

on a high performance game over SSH.

>> And I just like deadpan say that without

acknowledging that that's deranged. Um

but it like broke and it like broke

containment like it made it to people on

like Hacker News or Reddit or whatever

who don't know the type of stuff that I

normally do.

>> So all the right comments are just

people being like this, you know, this

dumbass vibe coder has no idea what he's

doing. Like what stupid he is. Yeah,

because people are typing in their

passwords to very secure uh boxes over

your connection to the snake game.

>> I don't even know. I I don't know what

they were mad at me about.

>> Well, we Yeah, it's funny cuz we had the

same not Well, we had this problem and

we chose to keep it in because we were

worried about someone actually some

somehow finding something you could

exploit over that way. But we had a

bunch of people complaining too of like

>> they're they're harvesting people's

public keys which was the funniest. That

was a real complaint when we lost. Yes.

>> Yeah. It's like Yeah. I'm going to What

am I going to do? Public key.

>> I know. It was the funniest complaint

and like tons of votes and people like

this is this coffee site is a scam to

from big data to get all of your public

keys. And you're like, what am I going

to do with h with only the public key? I

don't, you know, it was like,

>> do do you know what the first word of

public key is?

>> I know. It was like it was so funny.

>> Yeah. So I similar similar thing of

people getting very upset uh about this

when you're like, I think it's okay,

guys. Like it's it's fine. Yeah. Yeah.

Also, it's not like you tried to

upstream it back to go and get them to

remove it or like it's you know.

>> No, but I did ping. So the recurse, the

place I mentioned earlier, Fippo, the go

one of the go crypto maintainers is also

a recurse alum. And so I pinged him on

like the recurse chat to be like, "Am I

like is this really stupid? Like

>> am I fired if I do this?" And he was

like, "For your game, it's fine. Don't

worry about it."

>> Right.

>> Thank you, Fippo.

>> Had the official author's approval just

reply with a screenshot to all people

getting upset like, "I've got receipts.

I got receipts."

>> By the way, I do I can I make a quick

shout out for your SSH blog? So, if you

haven't checked it out, you should go

check it out. But this was my favorite

joke of the entire thing, which is why

does the snake game render strangely for

me? Snake snake snakes.Run makes use of

uni block unicode block elements. Since

these were added in 1991, your terminal

may not fully support them yet. For

example, Mac OS's terminal app renders

them poorly. Try using a different

terminal. It's like my favorite thing in

the universe is this thing right here.

This was I cannot tell you how many

issues I got in Neoim from people using

terminal app and being like Neoim's

broken. It can't do colors. And you're

like, I don't understand how. And then

turns out, oh, this the default terminal

Mac can't figure out how to update.

>> No, you can't update it. The technology

the the the private key to the repo was

lost and so they just have the same

thing since 1989.

>> It's so frustrating.

It's actually stuck in 1989. It's

>> It's old.

>> Prime that uh that site uh that that you

had up, I just want to note that um in

keeping with the SSH aesthetic, the

whole site uses tables for layout and

only HTML that was available in 1995

with no CSS.

>> Oh,

>> which I was very proud of. Um that was

tables for layout actually kind of sick.

Like kind of

>> tables for layout actually really good.

It's actually

>> it was like honestly you know what

really easy to think about.

You guys heard it here first. It's

coming back after this podcast. You're

going to start seeing tables for layout.

The Google search trends big. It's going

>> What about frames? Remember frames.

Bring those back.

>> Like first off, I frames are all over

the place. You go to any news. Okay.

Okay. Now we're Now you got me. Now

you're going to ruin my life and now I'm

going to be very upset. You go to any of

the news outlets we're talking about. It

doesn't matter what it is. Techrunch.

>> The only news outlet I go to is the

prime time. Okay.

>> Prime Time. Well, I synthesize this for

you, but if you go up on your tabs and

uh you you look at you can highlight and

hold something over and it'll tell you

how many megabytes. Like Twitter right

here, while having an actual video right

here of Doom being loaded and being

played is taking up about 211 megabytes.

It says it right there. If you go to any

news place, it will be 1.5 gigs every

single time. And then you go and you

start perusing their code and it's all

just these big videos being pushed into

iframes and it's just insanity going on

in the website. I frames the devil. I

just want to throw that out there.

>> But they're very fun to play with. You

could totally do that. I made an entire

business

>> game on uh Facebook, the old uh Facebook

apps they used to have way back in the

day. I'd load an iframe, do

authentication through an iframe and

through uh

>> the URL.

>> Quick uh quick proposal for a game. I

don't know what it needs to be, but we

need something about eye frames, but

it's like both the eye frame and the

internet way, but also the Dark Souls

way.

>> The Yeah. Yeah. Yeah.

>> Like both.

>> Oh, yeah.

>> You know what I'm saying? Something

about them makes you invincible. I don't

know how. I trust Nolan's brain to

figure that out in 6 months.

>> That's inputs and outputs right there,

Nolan.

>> That's inputs. Outputs. Inputs. Outputs.

>> Inputs. Outputs.

>> Input. Out. I frame. Output. I frame.

Think about it.

>> I had a friend, I don't know what

happened to this project, who was

working on a thing because Blue Sky

still has a fire hose, right? like you

can get all of the the data from Blue

Sky.

>> And so he had a site

>> um it's more than you'd think, right?

No. So it was he just made it was a site

that had an individual iframe for every

image and every video that anybody

posted to the site like scrolling like

as fast as it could.

>> It was like actually pretty sick. Like

it was fun to watch.

>> Yeah.

>> Um but I don't know. Maybe it's too

scary to put online. Spencer, if you're

listening,

>> there you go. New blog post as well.

>> Get that up.

>> Mhm. Well, there's definitely something

in uh like the game jam community. I

don't know why, but it's kind of like,

you know, that that thing about uh how

everything evolves into crabs. In game

jams, there's this thing where

guaranteed at least one entry in every

game jam is a game where the game is

being played across multiple like

viewports, like multiple windows or

multiple. It's like a platformer and you

your character jumps between the so

maybe there's an eye iframe idea there

where you have like multiple eye frames,

but it's one game being played across

them.

Oh, dragable eye frames to solve

puzzles.

>> That's something like that.

>> I was just

>> with eye frame souls like

>> it. It has to be. Yeah.

>> He's cooking.

>> Still pro. Yeah. Yeah. Yeah. All right.

No, he can't cook on stream. Can't cook

on stream. I have to think about that

one after.

>> That just reminded me.

>> You got Yeah, that like hit me. All

right.

>> It hit him. It hit him.

>> You just nerd snipe him. You just

changed his next six months, TJ.

>> Yeah. Yeah. Sorry. I apologize. I did

not mean to do that to you, but I am

excited for frame. I frame I frame

squared.

>> Yeah.

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

can watch the rest of it on the Spotify.

And don't forget to like and subscribe.

Woo! See you later.

>> Boot up the day

five co errors on my screen.

Terminal coffee.