Paul Graham

Eric Migicovsky gave 14 yo a Pebble watch. He loves it the way startups dream that early adopters will like their stuff. I asked Eric if I could invest. He just laughed.

R1: How can we distract people from the Epstein Files and the Iran War? R2: We could rename Lake Ontario. R1: Would they really buy that? R2: Think about our base. R1: R2: R1: Ok. Lake America it is.

I believe the most qualified person should get the job.

11% a quarter is almost 52% per year.

Someone asked how founders can make their companies huge without being overconfident. It was meant to be rhetorical question, but it has an answer, and the answer is an important one. Incrementally.

You put your head down and work. You focus on solving one problem after another. Gains accrue exponentially. And then you lift your head up a few years later and look around, and you're amazed at what you've built.

I spend a lot of time encouraging founders. Startup founders have a reputation for overconfidence, but this is just sample bias; the arrogant ones are also the loudest. Even now most good startups underrate themselves, so it really helps to encourage them.

Fortunately the fact that most good founders underrate themselves makes it easy to encourage them. You just tell them the truth about why they seem promising.

An email I sent to the summer 2007 YC batch to encourage them before Demo Day. It's all still true today.

In the essay I just wrote about how to prepare students to be founders, I talked about the importance of working on their own projects. In this essay from 2021, I explain in detail how and why to do that. A Project of One's Own: paulgraham.com/own.html

It's much harder to write an essay that has to cover a topic completely. In a normal essay you just do a core sample through some interesting topic and talk about what you notice en route. But if an essay has to be complete, you sweat over what you might miss.

"How to Do Great Work" (paulgraham.com/greatwork.html) was like this. It's supposed to be a complete description of the intersection of the principles for doing great work in different fields. And the intersection's big too. I doubt I'll ever have to work so hard on another essay.

How Universities Should Prepare Founders: paulgraham.com/prepare.html

At the high water mark of wokeness, the American Museum of Natural History in New York put labels on the glass of this diorama telling people the correct way to think about it. I took this picture in 2019. Are they still there?

When she was younger, Jessica sometimes felt she faded into the background.

LinkedIn users have been editing their work histories to add "AI" and remove "DEI".

Trump had to start a war with Iran to distract us from asking what it says about him in the Epstein files, but that went so badly that he now has to start a beef with Canada to distract us from the war with Iran. Expect two more years of these.

"I feel like I'm the squarest eccentric I know." — Jessica

Someone asked what I'd do if I were 17. I'd learn how to build LLMs from scratch, and then train ones as powerful as I could with whatever hardware I could get access to.

Notice that what I would not do is try to start a startup. Instead I'd build the foundation of knowledge to base a startup on later. Way better startup ideas would grow out of understanding LLMs deeply than trying to start companies based on what I knew at 17.

When a mechanical watch gets a full service, they take it completely apart, clean the pieces, and then put it back together again. This is (the components of) an Omega 552, one of the great automatic movements of the golden age.

When I posted this, it was considered right wing — a university had just claimed that saying it constituted a "microaggression." Now it reads as left wing, because it implicitly supports immigration.

It's generally a bad sign when people use the adjective "quick" to describe something they want you to do. Have a "quick call" for example. It means it will be unpleasant and thus not lasting long is a feature.

It's very common for founders not to know how well they're doing. In both directions. It's obviously common for founders to be making mistakes and not realize it. But it's also surprisingly common for them to be doing great and not realize it.

I think the reason is that running a startup is so unlike other kinds of work. By your third startup, you'd know if you were doing great. But with the first one you don't have anything to compare it to.

I did office hours today with a startup growing at 30% per month. They were worried this wasn't high enough. They stopped worrying when I pointed out that if they kept growing at this rate, they'd need an office 23x bigger every year.

If the Chinese government were sufficiently organized, they'd be funding opposition to data centers in the US. But the Chinese government is very organized. So they presumably already are.

I asked ChatGPT to make me a map of territories on the Italian mainland controlled by the Saracens. It is not very good at making maps yet. How strange to simultaneously worry about AI and get this from it.

Preventing companies from building data centers in America wouldn't slow down the rate of progress in AI. It would merely slow down the rate of progress in AI in America.

14 yo asked what DHL stood for. I said it was probably something German. Deutsche heilige Lager?

There may not be many people left in 10 years who can write, or want to read, things longer than a page or two. But there will be at least a few of us, and we'll be a powerful club.

The way to win at capitalism is to focus on what it actually does (make customers' lives better) rather than the socialist caricature of it (squeeze every penny out of them). The latter gets you 2x returns at best, whereas making new things for people can get you 10x.

There are plenty of people in business implementing the socialist caricature of capitalism. But they're mostly just opportunists and bean counters, not the big stars. The big stars make money by creating wealth.

I did office hours today with a startup that had three different paths to get huge. I told them to optimize for growth rate and let that decide. Maybe the winning path will be a fourth one we don't know about yet.

The dials of golden age watches are more vulnerable than modern dials. But I wouldn't trade. Modern dials seem too hard and perfect, in much the same way the digital versions of mid-century metal typefaces do.

14 yo: If Y Combinator had a mascot, what would it be? Me: I suppose it would have to be a cockroach. 14 yo: Not a very appealing mascot.

This essay (written during the 2008 financial crisis) explains where the cockroach thing came from. The Airbnbs were the quintessential cockroaches — which served them well not just in 2009 but during covid too. paulgraham.com/badeconomy.html

How should universities respond to the Arday scandal? Should they crack down on plagiarism and academic fraud? Ghent University's response, apparently, will be to fire the person who reported it.

@Itsjoeco @grok I'm going to delete this because Grok guessed wrong.

Teenagers rejecting all their parents' ideas has an advantage similar to the one that leaving fields fallow has in interrupting the lifecycles of pathogens and parasites. It's annoying, but it means stupid stuff isn't perpetuated forever.

The Tesla that Jessica just bought cost $48k and has a range of about 350 miles. The one she bought in 2015 cost $135k, adjusted for inflation, and had a range of 270 miles. People who think they dislike capitalism don't know how good they have it.

The trouble with having Zane Hengsperger visit your house is that instead of the usual random house gift, he brings you a 10 pound block of solid aluminum.

I was thinking about how I recognize AI writing, and one big tell is excessively colorful verbs. A handful of journalists might write that a proposal "drew" 100 votes, but any normal person will just say "got".

It's a bad sign when someone denying you something uses the phrase "at this time." That phrase means they've gone full bureaucrat.

If you're a good person, you should try to date people who are good judges of character, because they have fewer options. They're prevented from dating all the various kinds of bad people, because they can see how bad they are.

Obvious exception: when they want to date bad people, or at least don't care. But most people are over that by 25.

@ryangrim While the answer to the second question is no (changes in tax policy are not what made Larry and Sergey rich), the answer to the first is obviously yes (changes in tax policy could prevent anyone from becoming or remaining rich). No one doubts that.

There's a pervasive myth that "tax policy changes" are the reason economic inequality is growing. Actually it's because technology makes it increasingly easy to start companies, and makes them grow faster once you do. paulgraham.com/richnow.html

If you want to get into YC, don't focus on getting into YC. Focus on building stuff and understanding your users. That's what gets you into YC.

Interesting example of the fixed pie fallacy in the wild. If someone else is getting richer than you, "their wealth is growing at your expense."

Sunday morning, Cafe Borrone. I see a man in profile. Are those flip-up sunglasses, or eyebrows? They're eyebrows. There's only one man it could be. It is.

7am, by Craig Simpson.

Using immigrants as scapegoats worked for Trump. Maybe scapegoats will work for us too. But who should we attack? I suppose rich people are the default on the left. Ok, rich people.

I never consciously realized this before, but the fact that something has a cult following doesn't mean that it's cultish. Wodehouse for example. His fans are fanatics, but no one could have been more matter of fact.

Some founders I talked to today were surprised that I spent 80% of the time reassuring them. They asked if all I did in office hours was tell people not to worry. I told them it was only because they were doing well. Otherwise I'd be telling them what to worry about.

It's obvious that one needs to tell startups when they're making mistakes. But, strange as it may sound, it's also important to tell them when they're not. Investors will judge them partly by their confidence. So if they have reason to be confident, they should be.

Someone I've known for a long time asked if I really wrote essays in vim. I was surprised that he was surprised. So it must be a strange thing to do. But I don't see why it should be. Essays are just text. What else do you need?

Since several people asked why I don't use emacs, here's why:

One of the most conspicuous trends in recent YC batches is that there are more single founders. Single founder cos went from 9% last summer to 18% this summer.

The main reason is that there are more single founders applying. Presumably because AI enables one person to get more done. I still don't think it's a good idea though. The reason to have a cofounder is not just to get more done. It's to have someone to help bear the stress.

One startup today was ready to give up because their idea wasn't working. I suggested they just start over from scratch with a new idea and defer to the next demo day. I sent them to have lunch and mull over potential new ideas and come back at the end of office hours.

When they returned, the first new idea they mentioned was great. Even better than they realized. We only had to talk about it for 10 minutes. In 3 minutes more we found a good domain for less than $5000. So 15 minutes total and they walk out with a good idea and a good name.

The cheapest way to get a watch with one of the great movements of the golden age is to buy a steel "cushion-cased" reference from around 1970. They're unfashionable now and can be got for under $1000. This has a cal 854, which has a reasonable claim to be the peak movement.

And peak golden age watch movement = the culmination of 700 years of work on mechanical clocks. Mechanical clocks first appeared in about 1270 and since then many of the great minds of their times have worked on them. That's why I'm interested in watches.

Making your model faster is a very smart move when your customers are paying by the token.

A couple days ago a founder sat thinking silently for a minute and fifteen seconds before answering a question I asked. (This is an estimate, because I only started timing him after about thirty seconds.) One of the smartest founders I've met too.

14 yo is fascinated by the idea that there are adults who not only listen to my advice, but actually act upon it.

I often write "I'm not sure why" about something in drafts of essays, but these rarely make it to publication without being followed by some guess about why. I don't have deadlines; I should at least try to answer.

A year ago lots of startups were tuning open weight models. Then the consensus became that this was a waste of time, because the model companies' next versions would surpass them a couple months later. Now tuning open weight models seems to be back.

I wouldn't be surprised if there are several more back and forths. I have no idea who ends on top. If AI tops out somehow, open weights win, and there are a thousand model companies whose moats are proprietary training data. Or four after a new JP Morgan combines them.

YC doesn't have a word for this, but this is what it looks like when I talk to a whole bunch of startups at once. I only talk to each co for about 5 minutes, but it's like office hours: we go right into what they're working on, and often come up with new ideas for what to do.

The one thing for sure that gets fixed in these conversations is how each company describes itself. At this point, in the middle of the batch, these explanations are still often terribly confusing. But after I decrypt them I repeat the decrypted version back to the founders.

Just when you thought things couldn't get any dumber.

This gap before office hours was just big enough to start a new essay. I wrote 814 words of the first draft. (We'll see how many make it into the last one though.)

Since everyone will ask, it's about some of the techniques I use in office hours for figuring out how to make ideas bigger.

One of my office hours today was with a founder from 2011. Company's still going, doing pretty well. I asked him to give some advice to the next group, from the current batch. He said the biggest test is your own psychology. Can you keep going despite all the obstacles you face?

I think at that moment the new founders started to grasp how hard startups are. Not just from what he said, but his face as he said it.

One of the main topics I talk about with the founders of the big, established YC companies is how to combine work with family. It's a hard problem. But worth solving; arguably no problem is more worth solving.

A lot of founders I talk to are worried about what will happen with AI. So am I frankly. But they're actually better off than anyone else. If the world is going to get turned upside down, the safest place to be is in a small, fast-moving company that can easily change direction.

This is the 47th YC batch. I missed talking to a couple during covid, but except for that I think I've talked to all of them. Strangely enough I actually write new talk notes each time, though I bet there's a huge amount of overlap with what I've said in past years.

One common question I get now is how much of the standard startup advice is still true in the AI era. So far almost all of it. There's more difference between the advice I give different startups in one batch than between what I've given pre-AI and post-AI startups.

The reason they're called vertebrae is that you can't expect Demo Day investors to remember more than about 5 things about you. So you want to decide in advance what the most important 5 things are and make them the vertebrae of your pitch.

As often happens, this exercise was also useful for the startups themselves. For example, trying to cook up super ambitious future plans to present to investors would often result in genuinely good new ideas.

The reason we're standing next to the wall, incidentally, is to take advantage of the acoustic foam. That's what the spiky orange stuff is. Our architect Cate Courteau decided to own it instead of hiding it, and now it has become part of YC's visual identity.

Talked to startups at YC for 7 hours straight. I think that's about the limit. I didn't quite saturate demand, but I got within about 20 founders of it.

@lreyzin So while AI could indeed give these fake scholars an advantage in publishing new work, it will simultaneously make their previous work vulnerable in a way they never expected. And it's all out there, published forever, for any future technology to analyze.

Rob Miles is worth following. He delivers the rarest thing of all: insights that are very general, but also novel.

(Of course novelty is relative. I'm sure someone in the history of mankind has said something similar. But it's certainly not a widely appreciated principle.)

Is the term "sales" being replaced by "GTM"? Seems kind of bogus frankly.

You can tell no one knows that New York is the Empire State from the way everyone emphasizes the second word of Empire State Building rather than the first.

You can be pro-meritocracy or anti-immigration, but you can't be both.

Are the people who were enraged by this tweet going to be even more enraged when I tell them I met a founder today who sometimes writes 50,000 lines of code in a day? Or is everyone now used to this? x.com/paulg/status/1…

Once again, limit case: super programmer building a new project from scratch that duplicates a lot of the functionality of an existing service.