– English Lyrics

Lyrics & Translation

[English]

I think the like dystopian view of AI or

AI as like taking all our jobs and all

of that is like not correct. And I think

the future of work is more human, is

more interactive, is more multimodal.

Once the making of things gets easier,

the bottleneck goes back to to to like

how many ideas you can have. I wouldn't

put learning to code as the top thing. I

would put learn to make things. Learn to

make things with code. Learn to make

things with video. learn to make like

anything with AI.

[Music]

>> Welcome back to the breakdown. I'm Dave.

This is Tom. And today we're joined by

Amjad Msad. Amjad is founder and CEO of

Replet, which was a YC 2018 company. Uh

welcome to the show.

>> Thank you.

>> So today, uh we want to talk about kind

of your beginnings in helping people

learn to code. what you're doing now in

replacing a lot of that work and then

where we see the future going in terms

of how people make things. So, Replet

started um I guess in 2016, you were in

YC in 2018. You started out as a tool to

make it really easy for people learning

to program to set up development

environments purely on the web. Um but

now you're really taking off in AI

assisted coding. Tell us tell us the

latest. Yeah. So, um the mission has

always been to make programming more

accessible. Uh then we sort of uh you

know after YC we sort of updated it to

be a little more ambitious. We started

talking about a billion software

developers and it just sounded absurd at

the time and

>> this is all prel right.

>> Yeah. Prelm. There was a moment of time

I don't know if you remember that but

like 20 in the like 2015 AI hype. There

was like a bit of a NLP natural language

hype as well. There's a bunch of AI

companies that ended up being just like

humans behind the scene. They all went

broke, but um there was like a a glimpse

of potentially doing NLP on code. So, we

had an idea that, you know, it's it's

probably coming. I I even had it in my

in my seed deck that like at some point

like we'll collect enough data to like

train models. It wasn't until like GPG2

in 2020 that I felt like it was was

going to be possible. So we you know we

had we had built all these um primitives

like the development environment the

sort of hosting environment all the

other stuff around it. Uh and I felt

like if if we just like added um like AI

agents like it will be able to

orchestrate all the stuff and it'll be

it'll be really great. agents were like

every time we tried them. Uh I think the

first time we tried it like 2021 didn't

work, 22 it didn't work and then 2024

like early early in the in the year we

just felt like it's getting close.

>> Even GPT40 could like be coherent for

like um 2 minutes somehow just ended up

doing a big bat. The company was not

doing super well.

>> Yeah. was I was about to ask was there

like a bet the company moment where it's

like we were teaching people to code now

we're no longer doing that now we're

going to enable everyone to build apps

>> yeah yeah yeah we had grown the company

quite a bit um and we were burning way

too much money uh and we decided to do a

layoff

>> and so we we um cut like perhaps 50

people and then like another like 15 20

people left and so we were like less

than half of the company and so I just

put everything on on replet agents I

just felt like this is the thing it has

to work. It felt like it was close.

Yeah. I don't know. I was just like burn

the boat boats kind of moments. It's

just like we got to do this. I mean,

it's I felt like this was the this was

the the thing that would make the

company work. And honestly, like if claw

3.5 hadn't come out mid US building

agent, we would have probably failed.

>> Yeah.

>> Because um like I said, like GPT40 would

like stay coherent for 2 3 minutes. Quad

3.5 was the first one that like could

work for like you know five to 10

minutes and actually actually work and

the code generations

>> and that's sort of the approach that I

think is most successful for startups

kind of building on the very edge of

what is possible you know you start out

on a mission it's the techn is not quite

there yet you start building and you

sort of you you skate where the puck is

going you know it sort of catches up

with you

>> yeah and you you want to keep keep doing

that

>> totally

>> yeah I remember you came to speak at YC

like 6 months ago and right after you

had launched this and at the time you

said we're very far from fully automated

software development.

>> Do you still feel that way?

>> No, not at all. I mean, I I feel like

I've, you know, every time I make a sort

of some kind of prediction, it feels

like bold and like, you know, but I've

been I've been consistently wrong and

like uh the way things are moving uh is

a lot faster. Just the level of

autonomy. Uh like I said, like maybe 3.5

was like 5 to 10 minutes. Like 3.7

perhaps can get up to like four five

minutes to an hour. The uh 4.0 No, in

the in the system card and opus, they

said they made it work for seven hours.

>> Wow.

>> 7 hours.

>> That's incredible.

>> That's always been the limiting factor

for making agents is like, can they stay

coherent? Can is the context length

actually useful to reason over. If an

LLM can work for for seven hours, that's

basically like a human worker.

>> Yeah. And presumably they're working at

much faster speed than a human would do.

So they're kind of completing perhaps a

week's work in those seven hours. The

only thing that I think is is missing

and a big limiting factor for actually

automating a lot of work is computer

use.

>> Computer use kind of sucks.

>> You know, I'm sure you've you've tried

it. And this is the difference between

Replet agent really uh giving it one

prompt going all all the way in an app

versus having to kind of babysit it a

little bit and like test with it.

>> And we had a company in the last batch

called browser use that's um that's

working browser automation.

>> Browser Use is great. And another one

called pig which is doing the same for

kind of windows desktops.

>> And uh I think advice I would give

founders today is um taking either

browser use or windows automation with

pig and trying to apply that into

enterprise into vertical

>> into a vertical industry.

>> The moment this technology worked those

two companies are just going to

>> totally and I think I think we're like

weeks or perhaps single digit months

away from it working really really well

and so now is the time to get started

using those technologies.

>> Absolutely 100%. And this is this is

what we're focused on. So you know

replet agent v1 to v2 was a huge jump in

autonomy. Uh v3 is the most autonomous

thing. So we're already working on it

and um the interesting thing for us is

the underlying technology and how it

would enable autonomy. There are like a

few important things. One is uh uh you

know being transactional. The ability to

roll back is is very important. So you

you would want you'd want it to be safe

for agents in the same way that gits

made it safe for human programmers to

kind of experiment and create branches

whatever you want the same thing for for

agents like if an agent kind of messes

up a DB migration it should be able to

roll back and also it should be able to

like sample

>> uh across different different paths. Uh

I think this is very very important for

autonomy. If you look at uh like for

example when when uh Anthropic publishes

their Swedbench score they publish one

with that sampling and one with sampling

and it goes from 70% to 80%.

>> And so this is the idea that you sort of

you spawn multiple agents each one will

take a a shot at it and then you figure

out which one works and and choose that

that branch effectively. We built a

infrastructure that is um uh fully

transactional and moves in lock step

meaning the file system is a snap

snapshot fail based file system the DB

is a snapshot based uh u DB and so a as

you're we're making commits uh to the

entire system including the virtual

machine as you're going uh and so what

we can do is also we can we can fork it

and branch it out. So if if we had

computer use that's working well you can

you can just like right now what they do

with the sampling is they do they have

some kind of judge that does ranking

>> which you know it's not a true verifier

a true verifier is a is a is a test

right and a computer use test so you you

can sample out and really pick the best

branch that's actually working

>> that's incredible

>> and and repeat that on and on and

reliability gets really really high

>> so fast forward six or 12 months you're

not spawning one agent you're spawning

is Is it five or 10 or a million? Like

how

>> I think this is where it's going to get

interesting, which is you'd want to give

the user the ability to set compute

budgets. I I think we're starting to see

that with um with the O style models of

like here's here's how much budget. But

but look, I mean, if if you give us

$1,000, we'll spend them if you know.

So, uh

>> that's really cool. I've not thought of

that before. Um but the idea of like

spawning multiple branches then just

picking the best one and being able to

do that in parallel. I just the human

brain doesn't naturally work like that.

We think sequentially. We can't.

>> It reminds me of this legend. I don't

know if this is actually true, but at

Apple when Steve was running it, he

would intentionally have teams doing

basically the same things and then see

which one did the best job. It's like

>> I've heard OpenAI does that now.

Interesting. Uh I I've heard that like

the Codeex project like had multiple

different different teams. So in the

literature, you'll see that uh small

small models sampled uh will beat larger

models. So Sonnet sampled is probably

better than than Opus companies. uh

haven't tried that where instead of

hiring one senior engineer you hire like

10 junior engineers and like give them

the same task and then

>> it's just very expensive with humans but

with LM it's relatively cheap you can

just do that do a hundred of them and

just pick the best one every time

>> so how are people using replet agent

like and and who are the people using it

>> you know it goes back to sort of our

early vision that if you make

programming easy uh then uh then more

and more people would want to would want

to do it actually this is one of the

first thing uh that that PG PG and I

sort of connected on um before we got

into YC actually PG um found us on

HackerNews and so we started this email

relationship um and so he he told me

there's like a super linear relationship

with how easy programming is versus how

many people would want to do it. The

optimizing function of rapid has always

been like just keep lowering the the

barrier to entry and and that's how you

grow users and you grow uh customers.

right now. Um, basically people from

kind of every walk of life, we've seen

users, product managers tend to be tend

to be a very a great um use case and

users for us and we've had customers uh

product managers uh who uh are able to

make significant impact on the business

without talking to engineers at all like

you know running AB tests or

optimizations or or things like that. I

mean it's empowering. I mean it gets us

to think about the really the seams

between these roles like you know

product manager, designer, engineer. We

actually just created a new new product

group and and and typically you know

product uh you know head of product is

have has a bunch of like product

managers reporting to them but we're

actually having uh this this group have

has designers, engineers and and PMs and

the the idea is they're all using AI all

the time to to prototype in some cases

go all the way to production. there

isn't this sort of waterfall model where

you know there's a lot of inefficiency

you know like communication problems

between these different teams they can

move incredibly fast and I think that

that starts to change how tech companies

work my experience of this was when

whenever I was working in a startup the

list of ideas and the backlog would

always would be infinitely long

basically and the bottleneck was always

engineering time and now I'm doing my

own personal projects I write my to-do

list or my idea list and then I get to

work and my the ideas just get done and

suddenly the bottleneck is like my

ability to have ideas

>> and it's just such a weird experience

looking at a to-do list. It's just empty

being like what do I do next? I've heard

from a team that has like a really large

uh replet deployment in their company

that their founders using Replet and

that's stressing the engineers out

because oh I did this in a weekend like

can't you do it in a

>> what do you guys show for yourself and

so are these previously technical people

or non techchnical people they're

building the first version are they

deploying that production or are they

giving to engineers and say hey build

this like what what do you see typically

>> we advise to like work with with

engineering but that doesn't always

happen I think It's um understandable

that a lot of PMs and designers want to

go straight to users. What we're seeing

is they go to like beta users and test

users and I think this works really well

but in some cases they just like put it

in production. Uh and so right now we're

having discussions with all these

companies especially the engineering

leaders are unhappy about this like

who's on call for these services you

>> who's responsible for the bug

>> who's responsible for it. There's a lot

of questions that are coming up. The

obvious answer to all these questions is

like agents are responsible.

>> So what's the limiting factor today? I

buy code a thing and I'm like yoloing it

into production. Like what are what do

the engineers typically object like?

What what's going wrong?

>> Security is is is a big one. Like um

LLMs are fallible like like humans are.

They tend to write some uh there's some

components that they do terribly at like

for example off

>> uh like they all kind of suck at it. all

use like you know old methods of of

assaulting and hashing and so that's

been that's been a big sticking point

and we've seen a lot of examples out

there right now of some catastrophes

luckily it hasn't been like there hasn't

been a major catastrophe I think it's

coming but uh there's been uh like solo

founders who would leak um you know API

keys or would make it really easy to get

around the login security protections

and there are a lot of tools out there

that are like really not trying to take

responsibility for that and saying, "Oh,

it's it's the user's problem, the user's

fault." For us, we think that as a

platform that is marketing uh for the

non-developer, you actually have a

responsibility. We're trying to take

away

uh some of the things that we think LMS

should not do today. Uh so O for

example, we have a built-in O. So if you

go to wrap and just say add o, it will

pull in an o component that we built

from scratch. It has capture on it. It

has all the security you know bells and

whistles. So you don't have to worry

about all of that.

>> We integrated with your database. You

have a user management portal on on your

site as much as possible. Building in

these components that are tricky. I

think payments is the other one. I don't

think you'd want the LLMs to kind of do

do the payments.

>> And they're not that different, right?

like the sort of checkout. You know, you

might have a one time checkout, you

might have subscription, pay as you go,

but there are not unlimited versions of

payments. There's like two or three or

four.

>> And this is the analogy is today humans

don't write their own version of

payments or their own version of O. They

use providers that have built these

components already. So, seems like we're

seeing the same thing play out and

that's the best way to do it.

>> Yeah, 100%. And then the other thing we

added uh we partnered with is a a great

security company called Samrip. And so

right now uh when you go to deploy a

replet app, we run a security scan. We

run a code security scan and we give you

like a report of warnings and errors and

things like that and the agents can try

to fix them for you.

>> Okay. And so security is one of the big

kind of bottlenecks to fully deploying

to production. I can imagine there must

be other things that coming down the

line like sort of scalability, you know,

looking for like N plus1 database

queries, uh performance bottlenecks.

What do you have in your mind like a

clear set of of blockers that need to be

overcome before this is truly like one

click to deploy?

>> Yeah, I I mean the big the biggest one

is just going to be humans like social

like you know there's just going to be

mistrust and I think it's just going to

have has has to play out. So you know

leaving that aside like enterprises has

to adapt to all of that stuff having

some way to also scan for scalability

like figuring out you know uh like doing

fuzzing or whatever it is or having some

kind of adversarial agent that's you

know trying to to to break your app I

think um you know that's one big thing I

think integrating with a company's

ecosystem so one thing we're adding is

the ability to bring in your design

system

>> that's cool

>> so in addition to kind of us providing

these components if you go to any

company they have a lot of these

components built out as raplet is

getting deployed into these large

companies how can we hook into their

their internal systems

>> this makes me think about the kind of

spectrum of these different coding

tools. Um, on one end of the spectrum,

you have kind of what I would call the

the power tools, the cursors, the wind

surf that let developers use this as a

way to amplify their efforts. And on the

full other end of the spectrum, you have

more of the like consumerf facing, hey,

if you want to make an app, like you can

now make an app. And it sounds like you

guys are kind of in the middle. Um,

you're helping companies get stuff done,

but you're doing it for people who don't

look like the tra traditional developer.

How do you see this world playing out?

Are there going to be like n different

tools or will we converge to one place

on that spectrum?

>> AGI is convergence obviously but but

leaving that aside, it's really hard to

plan for that world. I think that the

battle for how to incrementally make

product uh engineers more productive is

just it's an obvious one. The market

there is obvious. There's a lot of

companies going after it. Both the

application companies like cursor, the

underlying model companies are trying to

go there. I mean um you know claude code

is competing with cursor. Cursor uses a

cloud. I think it's a bit of a blood

bath there but it's the market is

obvious and the market is really large.

I would guess that there's going to be

more of a consolidation there.

>> Maybe it's not you know one but it's

probably two or three uh at best. I

think the sort of the the market we're

in is a lot larger. Uh so the

addressable humans is a lot more. You

know, we're talking about a billion, but

it could be more. Like really any

knowledge worker should be able to like

solve problems with software. My

conception of replet right now is the

what we want it to be is a universal

problem solver. It'll solve problems in

your personal lives or you know solve

problems in uh in your work and and all

of that. And so I think that that market

is uh will probably be like a little

more diverse and I think companies will

kind of figure out where um where they

slot in trying to solve uh autonomous

programming with the focus on the

non-engineer.

>> Uh we want you to not worry about

security, not worry about systems, not

worry about any of that. We want you to

really come in with your ideas to to

Replet and be uh an agents manager. And

we're trying to do it in a way that is

like as human as as possible and kind of

fits into into into the workflow. One

big difference between engineers and

sort of non-engineers, be it executives

or product managers, you're not on your

desk like 8 hours a day. Mobile is a big

part of that. We have like a really

great mobile app. And we're thinking

about this way of like ambient building.

Maybe you start an app on your uh in

your desktop. You go away with your

phone. You're in boring meeting. You get

a notification from the agent saying you

I'm done with this. Do you want

something else? You got texted. And so

we're trying to kind of build that

thing.

>> Yeah. A question I had um related to

that point is is about the user

interface. So for something like cursor

or wind surf, it's pretty obvious. The

primary UI element is like the code. You

you see code, you have a little chat

window, but primarily it's about diffs.

It's about kind of changes to code. And

for a tool like Replet, the primary

interface is like the graphical user

interface. You know, it's the it's the

buttons and the the wizzywig like you

see what you're building.

>> Exactly. And that's great for building

user interfaces, but when you're trying

to build more complicated sort of

logical flows, I found it a little bit

difficult cuz I couldn't there's no way

I couldn't see the code. I can't

visualize what's going on behind the

scenes. And it's sort of it's almost a

black box. Fast forwarding here, if

you're trying to build more complex

internal workflows, how do they how does

a product manager or an operations

manager at a big company visualize that

workflow and the kind of logical

branching that happens?

>> Yeah. So, if if you look back in the

history of computing, there was always

this vision of visual programming, it

never worked uh very well because

ultimately it's about Turing

completeness like these systems are not

universal computing devices. And now we

go to codegen. Obviously codegen is is

turing complete. Um but you're

interfacing with it primarily via

natural language. Natural language is

fuzzy. It's really hard to know whether

it's doing the right thing. I think the

synthesis of these two things is

probably coming where you are

interfacing with natural language but

you can instead of like just staring at

code uh there's maybe an interface or

like a different view on top of code.

You can imagine being able to do you

know small talk.

>> Yeah. So small talk is this uh

>> it's the first object-oriented um uh

programming um system and you know Alan

K would say it's like it is actually OP

where everything comes after it is not

um but the interesting thing about it in

small talk you can actually the way you

interact with code is not via files but

via objects via like logical objects and

so there's some kind of prior arc there

and I think the world we're headed in

where there's some kind of abstraction

over code that allows people to like

understand it.

>> Yeah, I think that's really interesting.

I there's some there's like an open

space there whether it's like pseudo

code is like looks like English but it's

a little bit more structured or it's a

visual drag and drop. I don't know. It's

>> Yeah, I think back to building products

with a team of engineers, designers, and

other folks. The interactions that I

would have as the like product lead with

those teams was verbal. It would be

written like abstracted ideas. We would

draw stuff on the whiteboard together.

We'd make system diagrams. We'd look at

the results of the app and we would test

it and point out like oh this thing is

broken that's too slow and it feels to

me like that sort of interface which is

very multimodal and very flexible is

probably the best end result like I

think we will get to something like that

where the author of products will be

doing that but the teams they're talking

to are not other humans they're agents

doing these things. Has has there been

an attempt in in sort of PM land to have

a little more formalism around

communication or

>> Yes. And I would not say it's been good.

The main thing is just like the PRD,

right? The product spec, right? And it's

just become, in my opinion, uh

oftentimes a performative

>> work artifact that you create just so

that you can have a thing to get your

promotion, right? If you're at a big

company. Uh but they're not actually

that useful. To me, the most useful

interactions are just these like

whiteboard conversations, right? What do

we wanted to do here? Oh, we got to

think about that. Oh, we didn't consider

this. Okay, let's re rethink this whole

thing. Like those sorts of

conversations.

>> AI can play a role in that as well. So,

I um you know this startup granola uh

that allows you to kind of record

meetings and they released this like

team version that like all the meetings

uh gets transcribed and and go there and

they have a mobile app right now where

like you can put on the table and like

it's trying. So I I was thinking maybe

we should go gr granola maximalism where

look you shouldn't fight the trend in

which companies are becoming

increasingly oral as opposed to like

written um and because like you know

people are talking on Slack people are

in meetings people like are

communicating via prompts with with

agents uh but you you would want a set

of AI tools that is actually like

creating that that that record and in

background that's like searchable and

organizable and and all of that.

>> Yeah. I wonder when we'll have our first

AI in like oral meetings. You know, you

you're jamming with your designer and

you the AI like chips in and says,

"Well, how about this idea?"

>> Yeah. And and this is where I think the

like dystopian view of AI or AI as like

taking all our jobs and all that is like

not correct. And I think the future of

work is more human, is more interactive,

is more multimodal, is more fun in my

opinion.

>> YC's next batch is now taking

applications. Got a startup in you?

Apply at y combinator.com/apply.

It's never too early and filling out the

app will level up your idea. Okay, back

to the video. All right, so so last time

we chatted, uh, you had launched Replet

agent and things were growing really

crazy. I imagine that has continued.

Any, anything you can share there? Maybe

soon we'll we'll share some numbers but

uh since replet agent launch we're

growing 45%

compound monthly uh average.

>> These are like metrics that we tell YC

companies during the batch when they

have a base of like no users to try to

achieve and you're doing this at larger

scale.

>> Yes. But you know it put a lot of strain

strain on the company and our systems

were still relatively relatively small.

I I feel like it can get to to your head

and you can start optimizing for for the

wrong thing. It's very easy in AI to

increase AR while users are not happy

>> because they're spending a lot more and

like not getting the results and in some

cases maybe shouldn't grow that that

fast because like you'd want users to

get a better experience for less less

money and so it's one thing that we try

not to obsess we actually don't have AR

goals at at Replet we have like more

product goals retention goals just like

other methods

>> yeah because the the sort of bad pattern

with some AI companies you grow topline

revenue very very quickly the churn is

like approaching 100% and eventually

that just catches up

>> and the gross margins are horrible too

and so it's like

>> you know the more growth you have the

worse the company's doing financially

and so

>> so how do investors see this space you

you must have talked with a bunch can

they tell the difference

>> it's all kind of a blur for them because

investors when they I mean I'm just

going to generalize here but when they

start looking at the space they'll use

everything for three minutes and

everything for three minutes looks the

same so I I think I think it will it

will start to clarify and these products

will start to uh I suppose to converge

diverge more with the different focuses

in the areas we're talking about. I

think over the next year it'll be like a

little more clearer but I think a lot of

them are are just like you know super

when we talk to them they're super

confused. They don't understand these

systems. They don't understand where

they're going.

>> I think that's a great segue into um the

the sort of technology underlying some

of these tools. I'd love to dig in a

little bit deeper. We've we've seen some

announcements from cursor and wind surf

about how they're kind of layering

whether it's um Claude or Gemini or um

the OpenAI models with then their own

layers on top the fast apply APIs stuff

like that fast apply models I guess can

you give us an overview of how it works

at Replet?

>> Yeah, a lot of it is you're patching

problems with the underlying frontier

models. Yeah.

>> So the reason fast supply was important

because none of the models uh were doing

very well at diffs. Can you explain what

what supply is?

>> When you're trying to like edit a file

as an LLM, the best thing to do is to

like uh create a create a diff. Uh but

these models are actually not very good

at creating diffs. They're actually not

very good at countering the lines in the

source code. So they get confused about

a lot of these things. So for a while, a

lot of companies were just like

rewriting the entire entire file.

>> And so diff for for nontechnical users

is sort of like remove these three lines

and inject these other three lines

instead. like find and replace almost.

>> That's right. That's right. And so you

needed some way to um make the model

generate some something of a diff,

>> but it's actually not good enough to

merge it. And so you need another model

to actually do the merge. So

>> instead of rewriting like it's a

thousand line file, just like output the

entire thousand lines,

>> it's very slow. It's very expensive. So

you prompt the model to be as lazy as

possible.

>> But in that case, it's really hard to

apply. Okay, so you need another model

that's like doing the application. Okay,

>> you can train a model or you can use

Gemini Flash or some of these smaller

smaller models. And so in some cases,

you need to to train a model or

fine-tune a model to do a better job at

it. In other cases, you just also patch

a bunch of other models to to do it.

It's engineering. It's engineering as

opposed to like research, I would say.

>> And I noticed um you guys don't expose

the underlying model to the user. With

other tools like cursor and wind surf,

there's a drop down. It's like I want to

you know let's see what Gemini thinks of

this problem. You don't do that. Why is

that?

>> A a big part of our research efforts is

in eval

>> and I think this is like an underrated

part of like you know co AI coding. So

we spend a ton of time evaling new

models writing evals generating evals

just data crunching trying to figure out

what the users are are getting or are

feeling. We built a lot of systems try

to understand um how these systems are

are performing the moment a new frontier

model uh lands we are evaluating almost

immediately. Jim and I for example like

uh you know couple months ago were

making the rounds really great model

really awesome at like oneshotting in

some cases better than claude agentic

work wasn't at like a tool calling at

you know some of the things like that

but users just see the hype and they're

like oh like yeah give me Gemini that's

>> how much notice do you have do they like

drop it on you and you like scrambling

or do you have like days or weeks before

that to to try to you know test it out

>> well we we have good partnerships with

these companies we We have a great

partnership with Google. We have a great

partnership with uh with Anthropic, even

with OpenAI. We have a close

relationship. So, a lot of them give us

give us a heads up, give us some early

checkpoints and we we try we we kind of

play with all of them. And in the case

of anthropic, we're always like the

first day kind of launching because we

end up building. A lot of times we're

sort of anticipating where they're

going.

>> Yeah.

>> Because you kind of like you can tell

like 3.5 3.7 there there's some

direction you can tell like where 4.0 is

going to land. So you start architecting

the systems but you know ultimately a

lot of our engineering efforts still

infrastructure like the the distributed

network file system that snapshot based

network file system like it took us 2

years to build like there's nothing off

the shelf to do that a lot of the

security stuff is really really

difficult like replet is one of the few

places in the world where you can get

like a virtual machine in the cloud um

you know by just like creating an

account and that's like a it's really

hard to kind of run this and protect

against this. You have crypto miners,

you have all sorts of stuff like that.

And Replet also uses um uh Nyx OS under

the hood. Nixos is a like fully

declarative also transactional uh

operating system generator as it were.

Uh we have like this multi- terabyte uh

hard drive in all the different regions

that we have compute. um that that that

cached all the packages in the in the

world and that gets attached to every

container and and again all of that I

mean I keep coming back to this idea of

transactionality like you would want the

system to be fully functional. You would

want it to be um you would want it to be

safe in order to like experiment with

something go back do the sampling with

with agents. So really a lot of the a

lot of the work is just like the

engineering of this infrastructure and

it's like not as a parent not as sexy as

sort of like we're trained this model

here's the um but but I think this is

where you can build like a uh a bit of a

lead.

>> Yeah. I when VCs talk about moes uh the

thing that I translate it to in my head

is what is the compounding advantage

that you might be able to achieve.

Right. It's not really a defensive mode.

It's more just I'm ahead and by being

ahead in some vector it allows me to

continue to move faster. That's right.

>> And it sounds like this is a great

example of that.

>> True modes are often not obvious until

like decades perhaps many decades into

into the company. Like you wouldn't know

what like Netflix's mode. Uh but

obviously they have one like Disney

tried to like compete with them and

everyone was like you know down on on on

Netflix but turned out they have a mode

that is this uh content production

system that they that they built.

>> So Amad you you started with the mission

of like making it easier for people to

learn to code. You've accomplished a lot

of that. Now you're pushing the envelope

of what it even means to code. I've got

young kids. I want them to be productive

creators in the world. What should I

tell them to do? Should they learn to

code? What does that even mean?

>> Look, I think if you want to go the like

professional software developer route, I

think getting a computer science degree

and like learning fundamentals makes

sense. But if you want to be a a

creator, if you want to be a generalist

in this world, I don't think it's

necessary anymore to learn to code in

the more like traditional ways of

learning to code. I think you pick it up

by osmosis almost like like go go go to

replet and you know and start using it

and at some point you're going going to

run into some issue where you're going

to have to look at code or you're going

to have to look at logs just by being

resourceful having to Google around and

all that. You'll start picking it up.

And by the way, this is how our

generation kind of learned how to code,

right? when you were talking I'm like

that's what I did.

>> Yeah. And somehow uh it just became very

industrial and very formal over the

years like the way we made web apps is

like you just like start a notepad with

a you know HTML file and now you have to

like learn like webpack or whatever you

know it's like so um I think the future

of work is not really clear what is you

you know we can sort of like have some

like you know idea of like where where

the world is headed and so with with my

children I want them to have like as

broad uh based knowledge as possible. Uh

I want them to be as generalist as

possible. I want them to be as

generative as possible like being able

to like create a lot of ideas because

once the making of things gets easier.

The bottleneck goes back to to to like

how many ideas you can have.

>> Um and so I I I wouldn't put learning to

code as the top thing. I would put learn

to make things. Learn to make things

with code. Learn to make things with

video. learn to make like anything with

AI.

>> And so what do you think happens with

SAS generally if we're very soon able to

say create me a version of Google

Calendar or clone Docyign? What happens

to SAS? Do you think

>> we have stories today of um a lot of

people replacing hundreds of thousands

of dollars worth of SAS um with Replet?

The other day I heard a story from

someone who um you know they got quoted

$150,000 for a piece of software. He

went and made it in replet sold it to

his employer for $32,000

cost him $400. Companies that have a

platform developer community around it

and and plug-in ecosystem and things

like that. I think those are safe.

>> You're not going to be able to vibe code

Salesforce. I think the the vertical SAS

is is in trouble and I think it's my

guess it's already probably showing in

some of the metrics.

>> Okay, last question. What advice would

you give to um founders starting out

right now?

>> You know the best advice is what we what

you what you actually pointed out

earlier is work on the edge of what's

possible.

uh because uh one evolution of of AI or

models will make your business valuable

and suddenly you're first on market. I

find it rare to see founders that are

actually like sitting down trying to

actually predict the future and and and

maybe that's something that was you know

ill advised in the past but I think

right now trying to actually figure out

where things are headed is very very

important uh skill to have. So like make

some prediction uh figure out like how

to create like a crappy product that

would get better immediately as you

switch switch the model. I mean computer

use is a is is a great example.

>> Oh it's been an absolute pleasure. Thank

you so much for coming on.

>> Pleasure. Thank you.

>> See you next time. Thanks.

[Music]

Key Vocabulary

Vocabulary	Meanings
code /koʊd/ B1	noun - source code or computer program instructions verb - to write computer code
build /bɪld/ A2	verb - to construct or create something
learn /lɜrn/ A1	verb - to gain knowledge or skill
make /meɪk/ A1	verb - to create or produce
idea /aɪˈdiə/ A1	noun - a thought or concept
agent /ˈeɪdʒənt/ B2	noun - an AI program or software agent
model /ˈmɑdl/ B2	noun - an AI or machine learning model
program /ˈproʊgræm/ B1	noun - a computer program verb - to write a program
work /wɜrk/ A1	noun - employment or task verb - to function or operate
human /ˈhjuːmən/ A1	adjective - relating to people noun - a person
interactive /ˌɪntərˈæktɪv/ B1	adjective - allowing interaction
future /ˈfjutʃər/ A2	noun - time yet to come
easy /ˈizi/ A1	adjective - not difficult
access /ˈækses/ A2	noun - permission to use verb - to reach or use
possible /ˈpɑsəbl/ A2	adjective - able to happen
grow /groʊ/ A2	verb - to develop or increase
need /nid/ A1	verb - to require
change /tʃeɪndʒ/ A2	noun - alteration verb - to make different
problem /ˈprɑbləm/ A2	noun - a difficulty
way /weɪ/ A1	noun - a method or manner