Tokenmaxxing: How Top Builders Use AI To Do The Work Of 400 Engineers

I think that's like the defining question.

Like, will you have control over your own tools

or will your tools have control over you?

Using open cloud these days is like driving a Ferrari and it's like exhilarating.

It's insane.

Like, you get to do things like it it figures things out you would

never think a machine could figure out

and it does it

so quickly,

but then it's also like a Ferrari and that you better be a mechanic.

Like, it's a Ferrari

that will break down on the side of the road you know,

when you most need it

and you need to get out with your wrench

and pop the hood

and like fix it,

you know, you're going to have to fix it yourself.

And so, this is a very exciting time >> >> in uh computer science and technology.

>> Welcome back to a special episode of the Light Cone.

In this episode, we're going to talk about how Garry Tan got back to

building.

If you follow us on Twitter,

you'll know that after a multi-year hiatus to become an investor,

Garry Tan is back to being a builder

and in the last couple months he shipped hundreds of thousands of lines of

code and built popular open source projects

that have gone from nothing to more than 100,000 stars on GitHub.

And he did all of this

while having a very demanding job running YC full-time.

A lot of people on the internet don't even think

that this is possible

and are somewhat like in disbelief,

but it actually happened.

We know because we were here to see the whole thing and so today,

we're going to talk about how he did it.

>> Well, I'm relatively uh shocked myself.

>> >> I'm amazed as well.

It was 13 years of not coding and then suddenly boom,

I'm doing about 400x the amount of work

that I was that year the last time I was even sort of like

2/3 of the time writing code.

>> Baby, to start things off,

how would we go back to the project that started it all off,

which was Garry's List?

>> Oh, yeah.

And just like talk about a few months ago how you powered up Cloud

Code and like started to get back to coding.

>> And it was right after one of the Light Cone episodes, right?

>> Oh, yeah, definitely.

I realized that I wanted to bring together all the people who believed what

I believed,

um particularly for California.

And so, I started a uh 501c4,

and now it's a C3 and a PAC,

which is sort of what a lot of political groups do.

Um it's a very common way to bring people together.

You know, everyone focuses on the money,

but we're trying to bring together smart people.

Um you know, what I learned in the years of working in San Francisco

politics is that bringing together people is

so powerful.

And uh that's what a mass social movement is.

And I said, "Okay, well,

why don't I just make a website where we start doing that?"

And it would just start with um why don't I start writing about the

issues that I'm worried about.

It's like, I want children in school, you know,

people watching this from all around the world might find it very, very strange,

like I find it strange, that uh it was not possible and still very,

very hard for a seventh grader

or eighth grader in middle school in San Francisco public schools to be able

to take algebra.

And that was, you know, a math education thing.

Like, I you know,

if I didn't get to do

that when I was in public schools in the East Bay of the Bay

Area,

there's no way I would have studied engineering at Stanford.

I never would have written code.

I never would have been able to do any of these things.

So, it was close to my heart, and I realized like, "Hey,

it's time to write code."

And I ended up building Posterous, my first YC startup from 2008.

>> What what was Posterous for people who don't remember it?

>> Yeah, Posterous was uh that simple blogs by email.

It grew to be a top 200 website on the internet,

and then Twitter ended up buying it for about $20 million.

So, that was sort of like my first bag, really.

I actually built it again uh as Posthaven when Twitter um you know,

bought it for the amazing people that we had hired,

and uh they shut down the startup.

It would have cost a couple million dollars to buy it back from Twitter

and at the time I had no money in the world,

so the next best thing was why don't I write it again?

And then in January of this year,

I ended up writing it a third time.

Only, you know, the first time it took about, you know,

$4 million and six or seven people and about a year and a half.

And then the second time it, you know, took about, I don't know,

100 grand and two people, me and my co-founder Brett Gibson,

who now runs Initialized.

Um, and maybe like 3 months or so.

And then in this case it took about $200,

which was my Claude code Max account, and probably 5 days.

Full featured blog platform, does everything you want, and then on top of that,

like full rag, full um, agentic retrieval, like be able to, you know,

sort of go out and read all of the internet,

like every tweet I've ever done, recursive crawl, deep research of any topic.

The algebra thing is just one of a whole lot of different issues

that we really,

really care about, and to be able to go ingest the internet, you know,

see all the arguments for and against, and then to craft incredibly detailed um,

reports on the back end about um, what are all the quotables?

Like I think people who are big followers of the Light Cone might remember

one of our first episodes about agentic uh,

systems with Jake Heller.

Actually, so Jake created Case Text

and he described exactly what I ended up building for basically journalistic uh,

long form articles about any, you know, sort of issue or uh, you know,

piece of news that was happening.

And so, you know,

anyone can go to Gary's list.org today and you know,

we do about two or three relatively, you know, researched, all fully sourced um,

articles about what's going on in California, in San Francisco, and LA,

and like how do we build a better government?

>> is the thing I feel like people missed about Gary's list they don't

fully get is that it's like the classic thing we've been talking about here,

which is like software was you build software to let people use it.

So, it was like you build a blogging platform

and people like write blogs

and maybe like they start their own Substack eventually

or they write articles.

But, Gary's list is both blogging platform,

but it actually does the work of a high-quality investigative journalist.

It's not just something that a journalist uses to publish their articles.

>> Yeah, I mean basically the for the equivalent of like $5

or $10 of Opus calls,

I mean I would estimate that it does the work of like, you know,

a real human being

that would have to like go painstakingly through dozens of articles,

read entire books about certain subjects, uh annotate them.

I mean, going back to the case text example,

like the thing that Jake taught me was

that you need to think about what a human would do with the context

given.

Like what would it retrieve?

Like does it go to the library?

What kind of book would it look for?

What does it search on for search, you know, on the web?

I mean, the great thing now is like you don't have to just do

that.

Like you can get Perplexity's API and you can do deep research there.

You have X's API, you can do deep research there.

You know, Grok's API

if you need to like do research on X using the Grok API is

actually very,

very good.

And you can just grab all of the context.

This is sort of going back to the philosophy of boil the ocean,

which is one of my essays.

It's like particularly when building a Gentex software now,

you don't have to settle for um what we did

when we were humans writing the code.

Like and that goes for research as well.

What if you absolutely boil the ocean?

Like what is, you know, the total completionist,

like if you were a human this would take you about a month to

do this research,

you can just, you know, zap the rocks harder.

Uh you know, it you pay more money and you might be token maxing,

but you should token max.

Like, basically, if there is incremental work that makes something more complete, more awesome,

more uh, you know, in the case of um, this type of writing, like,

we want it to be more representative of reality.

Like, you know, we don't just settle for one source

when we can get 20 sources

and we can cross-reference them.

We can figure out like, well,

these 13 sources say this and seven sources disagree with that.

And then, you know,

you want to feed all of

that context into like your core prompt

and then you can basically make a better decision than what you would like

just,

you know, a human being clicking on a link, reading a headline,

and that's all you understand.

And I think if you token max, like,

that's actually the coolest thing you can do now

and it's not just in you know,

generating articles.

It's not, you know, it's clearly in uh, writing code, right?

I think now it's it's going to permeate every part of society.

Like, every thing that we would call knowledge work could be token maxed.

And um, I don't think

that it means that we're going to get rid of people.

I think it means that people need to still supply uh, the agency.

Like, I need this.

Like, I'm the one who's sitting here caring about algebra.

Like, I want kids like me who couldn't afford private school.

You know, San Francisco is the one city in the world

that has the highest rate of private school attendance,

um, probably in the entire country, actually.

And that's not okay.

Like, you shouldn't have to be rich to have a good education.

And, you know, I don't know why that's controversial.

And so, for me, it's like this, you know,

mass sort of a shift in technology was happening and then, uh,

I had a need

and a want and a desire

and it was a burning desire.

Like, I it hurts me and pains me to think about 10, 12,

13-year-old kids who don't know algebra and like could have,

but uh some bureaucrat or, you know,

some virtue-signaling person in power says like actually I don't want

that kid who wants to learn algebra to learn it.

>> So, I think in this process of basically solving your own pain

and need from the young Gary

and building Gary's list,

you sort of discover a lot of patterns on token maxing

and this new way of building

that led you to the next project,

which was uh G stack.

>> Like I actually did not plan to make G stack.

All I did was like I uh realized

that I was doing the same things over

and over again,

and then I got sick of typing the same thing,

so I went into my Apple Notes.

I typed in all the things

that I find myself writing over

and over again into Claude Code.

And it was pretty simple stuff.

It's like here's the plan review.

One of the things I started doing is I really love asking Claude to

make ASCII art diagrams.

One of the things I discovered is um sometimes Claude would just get confused

and like write bugs

or not be complete.

But once I started saying, "Actually, before you start your work,

make an ASCII diagram of all the data flows, all the inputs and outputs,

what are the user flows, what are the error messages."

And you can see this.

It's like data flow, state machines, dependency graphs, processing pipelines, decision trees.

Once it did that, it loaded all of the context in,

and then it just did the work more completely.

Like it boiled the ocean better.

And it broke down into a bunch of different sections.

Like here's architecture review, code quality, test.

I mean, one of the things I learned building Gary's list was

that when I was writing the code myself,

I would always do the minimum amount of testing cuz it's just like not

very fun.

I knew I needed to have it, but I'm here to write, you know,

fun new code.

I, you know, did not like write to write tests.

And then honestly, like I hit all the things

that everyone else hits

when they start vibe coding,

which is like this is slop.

It's not working that well.

Like, it works fine for the 80% case,

but if any users actually touch it, it starts falling over.

And then that's when I realized, "Oh, I can get to 100% test coverage."

I've since learned that 100% is probably too much.

Like, hitting 80 to 90% is usually the best practice at this point.

Um but yeah, this this is basically the first version of plan-eng-review.

I know everyone knows the office hour skill, uh which is, you know,

what people can use

and I still use

when I'm trying to make a brand new product

or a brand new feature.

It uh simulates what a what we do when we're working with a company.

It's like, "How do you know that people want this?

You know, who's it for?

What does it do?

And what's the impact, right?"

But this is like the proto skill.

Like, this is I didn't even know skills existed,

and I posted this and it went viral like, you know,

200,000 people saw that.

And then I made another version of it

that was a much more uh expensive version.

I called it the mega plan,

and then I ended up um renaming it to the CEO plan.

We've probably talked about meta prompting before.

I used meta prompting here.

I took the other review plan that we had, and then uh I said,

"Okay, well, let's do a version of this,

but like imagine Brian Chesky sitting with you, right?

Like, Brian Chesky has this great line about uh what is a 10-star experience?"

So, and you know,

the point of it is everyone thinks about hotels in terms of like three

This is a two three-star experience.

This is a four-star experience.

And he like goes, you know, through the list like five stars.

It's like everyone, you know, yeah, cool.

Like, he's like, "What's a six-star?

And what's a seven-star?

And what's an eight-star?"

And like, he goes all through that entire list.

And um that's one of my favorite like product

and design exercises to go through like

as a mental exercise.

And then the cool thing is like you can do

that every single time now.

And so, that's what this is.

You know, this prompt basically tries to figure out what is the platonic ideal

of what this is.

These are sort of like three the two things that are pretty awesome.

One is what is the 10x check?

What is more ambitious and delivers 10x more value for only 2x the effort.

Right?

And so for whatever reason coming out of latent spaces helps the model like

really visualize.

Like so I'm plan CEO skill I actually really enjoy

because I'm an ADHD CEO

and I love potential,

like pure potential.

And so this is like the one like I can't believe this is just

literally two little sentences.

But like this unlocks an incredible amount.

And so that's how G G stack started actually not as you know,

I didn't want it to be anything other than like, well,

I just need to make some skills

and I had heard

that people were making like skill repos.

But then the third thing I did was I started using these two skills

so much that my conductor instance was getting very backed up.

So this is how I use conductor.

This is actually my real setup like >> Okay.

So this is your like daily workflow.

This is how you've been shipping hundreds of thousands of lines of code a

month.

It's all it's all in here.

>> Yeah, that's right.

So I dropped like 13 PRs in the last 48 hours and then,

you know, I you just queue them up.

Like anytime I come up with a new idea,

I come in and here it is.

You know, I love using the CEO skill.

I love using the engine skill to like really make it super well tested.

I did that all in plan mode.

And then I'd click approve here and then, you know,

Claude would go and do all the stuff.

And then I did

that so much that I ended up having like 15 different features

that were all queued up waiting for me to manually test it.

Like it passed it you know, it passed end-to-end testing.

It passed integration.

It passed unit tests.

But like at the end of the day, I still need to, you know,

for Garry's list, it's like pop open the Rails server and like, you know,

load that user and like make it into

that configuration for that particular user

and like manually just make sure it works.

And I got sick of doing

that and I was trying to use um Claude in code MCP

and it was very,

very slow. 2 to 3 seconds for every turn.

I was like, this is not usable for QA.

But I had heard that Microsoft had released Playwright,

which is sort of um an alternative testing framework.

In retrospect, it's like actually there was like agent uh there like agent harness

and like all these other like tools

that I could have used,

but the upside and downside of Claude code is it's

so easy to just start something

that I just popped open like I literally went in here

and this is probably what I did.

It's like, I'm so sick of using Claude.

Claude in in Chrome MCP, it's too slow.

Let's go ahead and wrap Microsoft's Playwright.

Can we do that?

And then I just pressed enter and then, you know,

one of the things

that emerge with G stack is it like this is how I create new

features now.

Of course, you know, what it's going to do now is like, hey dude,

you already did that, which is hilarious.

You know, I have bug fixes right next to giant features

and then um the way G stack works,

there's a CEO, there's a designer, there's actually a developer experience person in there.

There's a number of design tools uh

and then plan eng is the last one

and then I actually usually run {slash} codex

and I recently added a {slash} Claude in codex.

So, one of the cool things that I actually learned from uh YC alums,

I came to an event and brain totally frazzled, but you know,

went to one of our batch events and we're just, you know,

shooting the about what's going on with Claude code versus codex

and at the time I was a total Claude code only guy.

And uh I realized, oh, a lot of people actually prefer Codex.

Why is that?

And I discovered that Claude code is ideal for the ADHD CEO.

But once in a while, there's a you know,

Claude code will just BS a bunch of stuff.

Like Claude models are very, very good, but like they are not the smartest,

it turns out.

And so a lot of people, you know,

explained to me that if you have a problem that's much crazier,

you need the 200 IQ nearly non-verbal CEO.

So you can just call in a friend,

and then that's what like {slash} Codex is.

It's a you know,

G stack skill that takes whatever plan it your plan is,

or if you're out of plan mode and you already implemented,

it'll take your repo,

and it'll run Codex in a command line prompt with the prompt

that says find all the problems

and all the bugs,

and it reports it back to Claude code,

and then you and Claude code can work through those that feedback.

And then I have since added,

if you use Codex as your main coding agent,

you can actually go

and type {slash} Claude

and have Claude come

and be the CEO briefly,

if you want, as well.

The cool thing about G stack is when I run it through this program,

like I always I do I start with office hours CEO review,

like I do design if there's UI.

If I know a developer needs to use it,

which is like practically all of G stack and G brain stuff,

I run the developer review, and then I do end review, and then Codex.

Once that plan is done, I've worked through all of the issues.

The G stack relies very heavily on ask user question.

So cuz you know, and that's that to me is like really important.

That's where the human, you know, vibe coder operator agentic engineer,

needs to supply their understanding of what's going on, what are we building.

There's not really a substitute to that.

It would surprise me very much

if someone really truly did manage to make a thing

that could just make software without the human in the loop.

Like that, you know, it's controversial take, I think.

But um, I never want to be entirely out of the loop.

I just want the machine to do the stuff

that I don't want to do.

And so, you know, basically QA is a good example.

And you know, I mean, that's hilarious.

Coming back to the demo,

it's like I type something into the modern version of G stack,

and it's like, "Dude, what are you doing?

Like, we already built that.

We have browse.

Browse is a long-lived HTTP demon with 70 commands as a CLI."

And then QA is just browse, but um, in the prompt for QA,

it says, "Look in your context.

What did we do on this branch?

If there's UI or any mutation of data,

go and use the browser to test that thing."

Which is cool.

It's like having a black box browser.

It blew my mind when it first worked.

It's like, "Mini AGI is already here."

You know, I you know, I realize this is not true AGI.

A true true AGI would be like, "I'm not even here."

Um, and actually, that's fine.

In this respect, like as a builder, you know, selfishly, uh,

I hope that we never have to stop.

>> >> I hope

that the machines never figure it out cuz

that would be really cool.

Like then, you know, humans are really important,

and like engineers who know how to do this, who have taste in design,

and product feedback, and um, you know, the real customer in mind,

like we're going to be like we basically have wings for

as long as we do.

>> YC Startup School is back.

We're hand-selecting the most promising builders in the world

and flying them out to San Francisco for July 25th

and 26th to discuss the cutting edge of tech.

Apply now for a spot.

Okay, back to the video.

>> I think you crystallize a lot of these thinking in this post on

X about thin hardness

and fat skills.

>> Oh, yes.

>> Which actually encompasses all of this philosophy on how to token max.

>> Yeah.

I mean, some of it came out of being trolled on the internet relentlessly

about markdown.

And like I you know,

I'm just like peddling a markdown instead of markdown. and it's like, you know,

I guess my lived experience at this point is that markdown is actually code.

It's just like this compiled in a different way,

but like you can get the computer to do really astonishing things.

Like, you mean even this,

it's like could we have imagined

that I would be talking to something

that has replaced Visual Studio for like I I don't use Visual Studio at

all.

Like there's no reason to,

like when I can talk to my agent and my agent can do this,

right?

The article actually name The name actually came from uh our partner Pete Kouman.

We have had to build an internal agent, and you know,

we call that the harness over and over again.

And then at some point using Claude code all day,

we realized like you know,

why should we rewrite a version of that over and over again?

Like, you know, we should just use the things that are really awesome as,

you know, harnesses.

Like a harness is the core loop that takes the user input,

gives it to the LLM, runs what the LLM does.

Like it can do tool calls and things like that.

I mean, why would we build that?

Like what we should be spending all our time doing is thinking about what

markdown should there be.

And the way to think about markdown is

if you were an event planner

and throwing a wedding

and you were trying to write down a checklist of how to throw a

wedding again.

Like what would you what would you write in plain English to teach the

next person who had to do it what to do.

All of that should be in the markdown.

Whereas um all the things that should, you know, be deterministic,

like um I mean, or is is a real action.

Like a a wedding planner might have to call like 20 venues, right?

But you wouldn't use markdown for that.

Like you would make a, you know, a call to Twilio for instance, right?

There's like a you know,

sort of all of the difficulty in in agentic engineering today is

when people try to do things

that should be in markdown in code,

and it fails because code is brittle, it doesn't understand special cases,

it doesn't actually, you know,

code literally doesn't understand what you want or who you are.

It is like, you know,

executing deterministic zeros and ones in a Turing complete loop, right?

Like, it doesn't know.

But then, now we have LLMs

that have latent space

and they know who you are

and uh it knows what your motivations are

and it can handle generic cases.

And then, you know,

a lot of the the magic right now

as an engineer is like figuring out,

okay, how much of it is over here in LLM land

and how how much of it is over there in um code land.

And then, you know, if you combine that with the other thing I learned,

which is like, get to 80 to 90% tests.

Like, if it's not tested and you're just throwing users in there, like,

it's slop.

You know, 10x worse than like human-written code cuz like,

you just have no idea what's going to happen.

Um and so, that's like one of the things that people have to do.

It's like, all right,

not only do you need to figure out what's going on in latent space

and deterministic space,

you also have to make sure that like it's, you know,

unit individually tested and then the integration is tested.

And then, going back to uh boil the ocean, like, the machine doesn't care.

It'll just do it.

It's amazing.

Like, just zap the rocks more

and you can get to 90% test coverage

and then you can have a system

that,

you know, is not quite perfect.

Like, you know, Openclaw right now, um there are lots of like failure cases,

but it's 95% there.

You know, it's uh I feel like using Openclaw these days is like driving

a Ferrari and it's like exhilarating.

It's insane.

Like, you get to do things like it figures things out you would never

think a machine could figure out

and it does it

so quickly.

Uh but then, it's also like a Ferrari in

that you better be a mechanic.

Like, it's a Ferrari that will break down on the side of the road,

you know, when you most need it

and you need to get out with your wrench

and pop the hood

and like fix it,

you know, you're going to have to fix it yourself.

And so, this is a very exciting time in uh uh computer science

and technology cuz it's like this is Homebrew Computer Club uh you know,

the moment when the Apple 1 came out.

Like the Apple 1 created by Steve Jobs

and Steve Wozniak was a breadboard inside like literally a wooden case hammered together

with like nails and duct tape,

you know?

And uh if you wanted a personal computer, that's what you had to do.

And that's where we're at right now.

Like you have relatively, you know, smart technical you know,

people who had to study computer science have to spend like two

or three hours and like maybe like $500

or $1,000 in both tokens

and cloud to actually get something like

that running.

But like once you get it,

it's like we're sort of in the kit car Ferrari phase.

It's like then you can drive and you can go anywhere and you know,

you want you want to shout to the hills like, "Hey,

I got a Ferrari."

>> Even the part about fixing yourself,

I feel people um is

that one of those things until you've like pushed through,

you just don't quite get.

If I really zoom out, it's almost like things have moved so quickly.

Like if you think way back,

just having Stack Overflow

as a website that you could consult

when you got stuck on a programming problem felt like amazing.

And then it's like like ChatGPT launches like, "Oh,

now I've got this like interactive thing that's way better than Stack Overflow."

But you're still sort of doing the same thing.

You're like asking questions

and you're copying pasting code

and you're running the code

and seeing what happens

and copying pasting it back

and then you sort of with Claude Code,

you sort of push through and you realize, "Oh,

you don't need to do the copying and pasting anymore."

It just like actually like executes and runs the code.

And then even with Open Claude, I found out when I set it up,

yeah, it's annoying cuz it can like effectively brick itself

and it does a bunch of annoying things.

But if you actually have like Claude Code like >> So, it'll fix it.

>> Yeah, like you just have Claude Code running, it will just like fix it.

And I it's clearly not the way things will be long term,

but there's this mentality shift of it doesn't actually matter

if it's brittle and requires fixing cuz you can actually just have another agent

like sat there like fixing it all the time.

>> Yeah, I feel like this evolution I was like completely Claude code pilled

and still am but like probably only like 50%

or 60% of my time like building product

or agentic engineering is in Claude code now.

At some point basically >> Almost half of it is through open Claude now.

>> Yeah, which is very interesting.

I mean then again I'm also spending a lot most of my time working

on G brain itself.

So G brain came about

because I met you know obviously we had Peter on the show.

And then I finally got around to it.

It was like one weekend I said I got to check this out like

what's going on with the open Claude.

Let's get it going

and this was about the time Karpathy wrote his ex post about knowledge LM

wikis.

And so I was like okay well I have a repo full of markdown

all my you know I should put all of my context into

that markdown and then at some point I realized oh shoot it's just using

grep.

And grep is not that good.

Like it's you know wasting context

as loading a lot more into context than it needs to

and then I sort of fell into a rabbit hole.

I just went into conductor click quick start

and then I had G stack built into conductor already

and then you know basically this was how I started.

I you know.

>> >> It was actually much more interesting than that.

So I didn't start off from nothing.

One of the things I've learned

as you write like a larger

and larger corpus of code is like you have it loaded in your brain.

You're like oh well in order to build an agentic newsroom for Gary's list

I actually had to learn about vector embedding

and hybrid RRF and chunking like

when you're in there trying to make it work you're just like very applied.

It's like I have an output that I want.

I want the article to look like this.

It needs to be of this quality.

It needs to have these citations.

Like you start building up uh you know your tests

and integration tests and like you end up with like a product that's like

battle-tested from like the output

that you want.

And so, I sort of put two and two together and I you know,

and this is something that you know, anyone can do actually.

It's like this this is why I think we're entering the golden age of

open source.

Uh I could just open, you know,

this project in conductor

and then the first thing I write is like you know,

go look at, you know, tilde/git/gerry'slist.

Like look at how we do chunking, embedding, uh you know, hybrid RRF, rag,

like all of this

and then just like extract it

and then I want to use Postgres with PG vector

and like I want a a you know,

full rag system for my open claw.

And then sort of like one thing led to another.

It's like then I have, you know,

10 windows in G brain and I'm just like at it.

What's cool about open claw, I mean maybe this is a good example.

This is actually my open claw.

I did go ahead and ask.

It's um how you know, how did I actually get into it?

January 23rd.

>> Also, all your emails. >> a tweet

that was like Claude code this week has awakened my 25-year-old self,

the one that checked Red Bulls and stayed up till dawn coding.

We're so back.

>> >> The builder identity resurfaces.

>> Yeah, and you know, I'm basically back to, you know,

sleeping 4 hours and you know, coding 20 hours a day.

You know, this is also

when I started getting myself into trouble like talking about lines of code.

I still believe this by the way.

>> Yeah, this might be like a good quick aside to talk about like

this this idea of like lines of code being important measure has been like

controversial on the internet.

There's obviously the counter argument like oh,

lines of code doesn't like measure developer prod productivity, but >> It doesn't, right?

But but it also does.

So >> It it also kind of does, right?

>> Yeah.

Like it does it's clearly and you know,

what's interesting is you can actually um there's well-published get repos out there

that you can run to uh strip away

and and standardize what is actual logical lines of code.

And so, I actually did go ahead and do that.

Um you know, and I got into trouble for saying like, "Oh,

I'm coding at like 100x uh the rate that I was in 2013."

And then after I did the logical lines of code stripped down,

>> It actually went up.

>> It actually went up.

So, it turns out that I was actually doing 400x the amount of code.

But, you know, obviously I wasn't writing it.

I was directing, you know, 15 agents at a time to do so.

And then by the numbers,

like it was not

that it did like knock down my lines of code from Claude code a

little bit.

But, uh the surprising thing to me was

that it knocked down the amount of lines of code

that I was writing in 2013 by like 70%.

And so, I think that that's sort of the mismatch here.

Like, people get very upset

because it's easy to like pad the lines of code

if you're a human writing code.

Whereas like unless you direct Claude code to literally like pad the lines of

code,

it doesn't necessarily do that.

Like, it'll maybe build the wrong thing.

Like, you might not steer it very well.

It might not do the right thing.

But, like it's not trying to optimize for lines of code the way a

human working a job would,

right?

Which is, you know, that's just life.

And then I guess the really surprising thing is

if you look at the literature about software engineering going back to like 2000,

1990, I mean, it's pretty clear

that the average number of lines of code

that a professional software engineer that's like tested

and production ready,

it's not like 100 lines of code.

It's like 50.

It's like 30.

>> Yeah, a day.

>> Yeah, a day, right?

Like, for me it was like 14, but I was like part-time.

I don't know.

It's uh So, that's where the 400x actually came from.

You know, the other thing I know is like I should have said

that instead of just trolling people more on the lines of code.

So, I you know, if I trolled you on the internet,

I'm very sorry for that.

Like, there you know,

there is a deeper understanding of this

and I did end up releasing a blog post about it

that um explains this quite a bit more.

I mean, and I think it's not a little bit significant.

It's very significant for people who are technical

because it actually raises the bar on like what you are capable of doing.

Like, all the people who are attacking me about lines of code,

they particularly are the people who are most likely to get wings

if you like let it rip

and token max.

This is sort of like the classic problem.

It's like if you have taste and you understand technology,

you are particularly the people who should would benefit the most from getting this.

All someone has to do is you know, believe.

>> Right.

>> Yeah.

So, stop fighting.

And just open Claude Code and try it, you know.

>> I think another thing that's potentially going on is just like the experience

is very dramatically depending on like the the models

and the harnesses.

Um I think something I've noticed is any sort of like semi-complicated programming task

I try and do through my open claw agent just like kind of fails.

Um like it's exactly the same model

and sort of like Opus 4.7

as Claude Code,

but it just like like anything above like a simple script,

I just find like it's not like that great at.

So, I'll go back into like Claude Code

and then it was sort of a moment for me where I realized,

oh, like this is how it used to feel.

Like this is how like even 6 months ago it used to feel like,

oh, like you try and like these things Yeah,

these things aren't quite there yet.

And then Claude Code with like Opus 4.5 was like, oh,

like it's actually like here.

>> It's about to recur.

Like right now people sort of are feeling like open claw

or Hermes is like not quite there

or it's like a lot of work.

And then I guarantee you like this time next year like everyone's going to

be saying what you heard here first,

which is like every single person on the planet will have their own personal

AI.

We could either live in a world where we have our own AI,

where we have our own data, our own integrations, like we see what's happening,

we write our own prompts, and we have control over what we see,

uh or it's corporate controlled.

It's something, you know, you go to a host,

it's kind of like your Facebook feed, and like you don't know what that,

you know, who wrote

that algorithm and who does it benefit

and like what model is behind it,

like nobody knows.

The most powerful idea that like was a gift was the personal computer revolution,

and we're about to go through exactly that same shift with personal AI.

And it's going to be a choice, like, you know,

people are going to have to figure out,

am I willing to write my own prompts?

And, you know, I think I wish Pete Koomen were here,

like that's one of the things we learned from him, too.

It's like, unless you have your own prompts

and you can write it for yourself,

like you are, you know,

below the API line for some PM or developer that is not you,

who like will not understand you, will not understand your needs,

will not understand what you uniquely care about.

And I think that's like the defining question,

like will you have control over your own tools

or will your tool your tools have control over you?

>> And I think this is the one of the disconnects that the public has,

I think, is a lot of these capabilities you have to be on the

latest and greatest models,

and it's actually quite expensive to use them

and burn all the tokens for now.

It's coming down, but I think maybe people are just trying like Sonnet

or the free model

or having the basic Claude Pro subscription only.

>> Yeah.

>> And part of it is maybe we have to address

that this new way of really getting all this almost ASI AGI moment for

for building is you have to be burning lots of tokens the whole token

maxing paradigm.

>> It actually reminds me of rent, San Francisco rent.

Like one of the things

that I feel like we always have to do um with YC founders is

that it's like a general thing.

It's like, "Oh, like I don't want to move to San Francisco

because it's like so expensive to live there."

But it's like it's so expensive to not live there.

>> Yeah, yeah, yeah, exactly.

That's the whole point, right?

Like early on in a YC batch,

like I'm just used to like a founder being like,

"Like this like this apartment is like thousands of dollars a month in rent.

Like seems ridiculous.

Like should I like pay it or not?"

And it's like, "No, you should absolutely pay it.

And if anything, you should pay more to not just be in San Francisco,

but be in like the Dogpatch

and just like be in like neighborhoods where you create this serendipity."

Like token maxing is going to be one of those things for founders

that we sort of have to teach them where it's not immediately obvious

that you shouldn't This is actually like rent.

Like this is one of the things where you should like spend

as much as you can to like get the like most utility out of

it versus treating it like the the office desk

or something.

Like sure, you can economize on

that or you don't need like a super expensive like couch,

but like when it comes to like actually using the models

and your token spend,

you should probably be like pushing pretty hard on that.

>> Yeah, one of the key maxims for YC is, you know,

how do you find good startup ideas?

Live in the future and build what's missing, right?

And so this is a profound version of

that where all you have to do is commit your brain to look at,

you know, spending $500 in a single day on tokens and say actually like,

you know, as long

as I'm building something that's actually of great value to me,

you know, and I'm building the right thing, uh I'm going to do that.

>> Hey Garry, I have a weird question.

Do you think that in some ways the fact

that you tried to build all of this

while also being the CEO of Y Combinator actually helped you?

Because like your time is so scarce,

you have to like try to figure out how to write hundreds of thousands

of lines of code with just like spare minutes in between meetings.

Unlike a a full-time software engineer that could, you know,

just take the time to like open the website

and like click around

and like just test it.

Like those minutes were like insanely scarce for you

and so you were constantly pushing yourself to figure out how to like automate

everything.

>> Yeah, I I envy time billionaires.

You know, sometimes look at I mean,

I'm look at my kids

and it's like these kids are time billionaires right now

and I'm like,

you know, you can just like do thing, you know, you know,

run across people at Startup School all the time and it's like,

you're a time billionaire right now.

Like this is incredible.

Like you could just do anything you like learn about anything.

This is so great.

So yeah, I'm you know,

personally like I think my philosophy is I am in a crazy rush.

In my brain, I'm like probably lived 10 billion lifetimes living in this body

right now and I need every single moment to count.

Uh and then if you can token max, it's like I mean,

you could buy millions of years of consciousness of machine consciousness.

Now, I can be a time billionaire.

It's not, you know, my own time.

It's the time of a machine.

Like doing work for me

and like the human entities

that I care about working on the causes

that I care about,

right?

I care about YC.

I care about builders being able to build.

Even in a lot of our internal meetings last year, remember in our offsites,

we would talk about like how do we teach the next generation how to

use these tools?

And so, you know,

I'd like to I wish

that I could say like

that was all a part of the grand plan

and that's how it started.

It's not like but you know, subconsciously I actually think it was.

Like I think subconsciously from doing like cone and like talking about this stuff.

Like sitting side by side with Boris Cherny right here was a very powerful

moment for me because I realized like he's he started saying things

that like I could do myself.

It's like he said, our team doesn't write a single line of code.

I'm like, oh, actually like I can do that.

And like the people who are watching right now,

it's like you and I are not different, right?

We're the same.

Like we started in the same place.

I don't think of myself as like, you know, in the sky yet,

even though people seem to talk like I am, you know,

like I'm just a person trying to do a thing.

And if I sit next to Boris, I'm like, you know,

this guy is one of the best engineers I've ever met.

But also like, if I just open a prompt, we have the same prompt.

We have the same MacBook Pro.

And, you know, there's nothing

that stands between like me

or you or any of us from like drawing on millions of years,

potentially, of like tokens to like serve humanity.

>> Well, Gary, I think that was a beautiful quote that should be retweetable.

>> It shows >> Got to get it on the X right away.

>> You could have infinite time by borrowing the time from the machines.

>> Yeah, what a time to be alive.

>> That's a beautiful thought to end on.

Thanks, Gary, for showing us the future.

>> Thanks, guys.

>> All right, thanks for watching

and I'll see you on the next episode of Light Code.