by Will Schreiber

"It's a terrible thing that we made"

Richard Feynman's wife died on June 16, 1945.

He got the call saying her condition was deteriorating, borrowed a car to drive the 100 miles from Los Alamos back to Albuquerque, and made it to her bedside. Barely.

After she passed, he went for a walk near the hospital. He was surprised that he wasn't more upset. Her seven-year battle with tuberculosis had probably softened the blow.

Exactly one month later, on July 16, 1945, Feynman arrived back in Los Alamos for Trinity, the first test of the atomic bomb. Feynman claims he was the only one to actually see the explosion since everybody was wearing extremely dark - too dark - sunglasses. He knew glass blocked ultraviolet light, which is what would actually damage the eye, so he took off the sunglasses, opened a car door, and watched through the car's glass pane from twenty miles away.


It was so bright that he fell to the ground. At first he saw white splatches of light as he flashed his eyelids open and closed. Then he stood up to see the rising fireball plume.

A minute and a half later the sound wave came roaring by. Then came the heat. Then came the celebrations.

Everybody was thrilled. They'd done it. They had engineered an atomic-level chain reaction. There were a lot of parties amid a flurry of excitement. For his part, he played drums in the back of a Jeep.

One man, Bob Wilson, wasn't celebrating. It's a terrible thing that we made, Bob said.

In another month, victory over Japan was declared. In two more months, Feynman left the desert to start teaching at Cornell.

On his way to Ithaca, he stopped in New York City and went for another walk, this time up and down the city streets. First 10 blocks, then 20, then 30, then finally stopping for dinner.

As he thought about how far he was from 34th Street, and how everything in between him and 34th would be reduced to ash by an atomic bomb, he had a terrible feeling of hopelessness.

Why was anybody doing anything? Why are people building bridges? Why are they building buildings? Why are people paving streets and piling into the Subway and riding up and down 5th Avenue? Don't they understand? Don't they get it?

Don't they see there is no point?

I'm not sure where Feynman walked exactly. But I made a similar walk last winter. I'd just spent a few hours asking GPT-3 for help with some complicated multi-JOIN SQL queries, which it studiously and earnestly and instantly gave back to me. And then OpenAI's API went down. I felt stranded. Agitated. How am I going to finish this project without help?!

So I put on my coat and left the office. I made my way across Union Square, down 15th, past all the tourists at Chelsea Market and into the warm and inviting doors of the Bus Stop Cafe.

I sat alone in a booth and ordered mint tea.

There were a lot of regulars around me, everybody laughing with the staff. Couples kept streaming down Hudson past my window.

Don't they get it? I thought. What's the point? Don't they understand?

My job writing software is like making kerosene lamps in the same year Edison announced his lightbulb. It's too late, I think. Human achievement is over.

Feynman wrote his autobiography 40 years after his lonely dinner in New York. "Fortunately, it's been useless for almost forty years now, hasn't it?" He remarked. "So I've been wrong about it being useless making bridges and I'm glad those other people had the sense to go ahead."

And go ahead we must - ahead with building bridges. And kerosene lamps. And silly little blog posts.

Feynman tells this story in basically the same order - but with much better detail and many other funny tidbits - in his book Surely You're Joking, Mr. Feynman.

by Will Schreiber

It’s good when things are bad

I mentioned the “We’ll See…” Fable1 to a friend at Friday Coffee this morning. He immediately raised his mug for a cheers. “Brother I love that fable.”

The status quo never stays. Bad things don’t last forever. Good things don’t last forever.

This has been so universally true in my life that, by the Law of We’ll See, it’s good news when things are really bad.

When there are layoff rumors, when the stock market has cratered, when inflation is rampant, when banks collapse, when sales efficiency slows and when weather turns for the worse and when AI threatens humanity’s existence… get ready.

  1. I first heard the fable from Derek Sivers. But it’s all over the place now.↩︎
by Will Schreiber


Yesterday, I woke up, hopped on a Citi bike, and rode to the gym. It took 5 minutes.

The gym was so crowded I left and walked to Joe & The Juice back by my apartment. I don’t like Joe & The Juice. But I do like their laptop policy.

After pushing code all morning, I went home to mix some dough before riding another Citi bike to Soho for my Apple Store appointment. Sandra unclogged my phone’s Lightning port.

From there, it was a 19 minute walk straight down Greene to my dentist for follow-up x-rays.

After getting scanned, I went to the gym across the street in Tribeca. This time it was empty. 3pm is the sweet spot.

After working out, I rode the R train to Union Square. 10 minutes. 16 minutes with walking.

I worked for two hours at Industrious, then decided to meet Elizabeth for a walk before the sun set. So I got back on a Citi bike, rode down 13th street, then over to the West Side Highway.

From there we walked home.

I am in love with NYC.

by Will Schreiber

AI vs. AGI vs. Consciousness vs. Super-intelligence vs. Agency

GPT-4 surpasses all sane definitions of “Artificial General Intelligence.”

AI (without the “G”) is a fancy way of saying machine learning - finding patterns within giant datasets in order to solve a single problem. E.g. analyzing billions of interstate driving miles to build Autopilot, billions of pictures to classify faces, or billions of minutes of audio to transcribe speech into text.

None of these AI (without the “G”) tools show human-like intelligence capable of accomplishing a bunch of different tasks. They are programmed to do one thing, and they can only do that one thing.

But GPT-4 can do lots of things.

It can carry a conversation, even if the chat format is “just” a parlor trick. It can write poems, brainstorm names, edit paragraphs, outline essays, and make jokes. It’s a Tower of Babel, capable of translating and answering questions across languages. It can solve math problems while showing its work. It can program Python apps, fix its own coding mistakes, suggest medical diagnoses, and help doctors communicate devastating news to patients. It can even learn how to do stuff by extrapolating from something as small as a single example.

GPT-4 has nearly aced both the LSAT and the MCAT. It’s a coding companion, an emotional companion, and to many, a friend. Yet it wasn’t programmed to be a test taker or a copywriter or a programmer. It was just programmed to be a stochastic parrot.

This is general intelligence.

When people say GPT-4 isn’t AGI, I think what they really mean is AGI is not conscious, GPT-4 is not super-intelligent, and GPT-4 does not have agency.

I agree.

So we are here on the progression of artificial intelligence:

✅ AI (single-problem machine learning)
✅ AGI (seemingly anything you throw at it)
🚧 Conscious AGI
🚧 Super-intelligent AGI
🚧 AGI with agency

Conscious AGI

As much as the goalposts have moved for AGI, the goalposts will probably move even more for “Conscious AGI,” mainly because we don’t have a clear definition of consciousness.

Richard Dawkins has the best explanation of consciousness I’ve heard. He explains how brains developed the ability to make predictions in order to help their bodies survive. Is that rock going to hit me? If I move, will the prey see me? If I steal that food, will I be chased? Our brains are incredibly efficient prediction machines, constantly modeling the world around us and stack-ranking probabilities in realtime.

Once the model turns inward and begins modeling itself by predicting its own reactions, then consciousness arises. If I steal that food, how will it make me feel?

In The Selfish Gene, he makes the somewhat obvious observation that consciousness isn’t all-or-nothing. Some animals have more levels of consciousness than others. I think we intuitively know this, which is why we don’t grant human rights to dogs, even though we know dogs are conscious.

Considering LLM’s are really good at modeling and predicting (they’re complex language prediction machines after all), they’re standing right at the ledge of consciousness, one step away from falling down the hole of modeling themselves. How big of a leap is it from asking Will this response satisfy the request to Will this response satisfy myself?

Add the fact that consciousness is a spectrum1 — and that species with far less developed brains seem to have some degree of consciousness — I think it’s likely LLM’s develop a level of consciousness before they become super-intelligent.

What we don’t know is how we’re going to figure out if the LLM is conscious or not.2 I haven’t heard any good ideas, mainly because I haven’t heard any consensus on what consciousness even is.

Super-intelligent AGI

Super-intelligent AGI is AGI that far surpasses even the best humans in a field.

Even if GPT-4 aces every test ever given to humans, that’s merely matching the very best of humans. To give ourselves credit, we’re a fairly smart bunch.

To be considered super-intelligent, AGI needs to contribute meaningfully to human knowledge. It needs to create a new important programming language, discover a new drug, generate new ideas, and write new stories and screenplays.

Diagnosing medical conditions the way GPT-4 has done is impressive, but a doctor who isn’t distracted by an onslaught of patients can also properly diagnose diseases.

A super-intelligent AGI would recognize new diseases, categorize symptoms in a new way, invent new words and theorems and finally explain subatomic entanglement to us.

That’s super-intelligent.

AGI with agency

Agency is actually the final unknown. Almost all AGI doomsayers assume AGI will have agency. They have this vision of the machine deciding it’s time to end civilization.

They might be right.

But just because something is super-intelligent or is conscious doesn’t mean it has agency. There are lots of hyper-conscious, hyper-intelligent humans stuck inside PhD programs unable to change course. (Kidding of course.)

Some humans — despite being fairly intelligent and fairly conscious — display tiny amounts of agency, barely able to alter their goals, living conditions, or diets. Others decide to drive motorbikes off ramps and over the Grand Canyon.

How does agency arise, and is it something wholly separate from consciousness?

I have no idea.

But I think we’ll see a conscious, super-intelligent AGI before we’ll see one with determined agency.

  1. Of course this means AGI might eventually become far more conscious than all humans other than Buddha have ever been.↩︎
  2. I like what Sam Altman said on the Lex Fridman podcast, that we can tell if a model is conscious by being very careful to not feed the model any mention of or description of consciousness during its training, and then seeing if the model is able to describe or identify with the idea of consciousness. HN Comment↩︎
by Will Schreiber

Fragile passengers

I tolerate longer lines when buying coffee than I do when going through the airport. Why?

I got a push notification at 5:15 this morning on the dot: “Your Lyft has arrived. Gabe will wait for 5 minutes.” Right on time.

We crossed town, and then crossed the Williamsburg Bridge. As we rode up 278, we passed a car flipped upside-down going the other way. Traffic back into the city was miles long. Not for us, though. We got to LGA in 26 minutes.

A screen posted in front of the CLEAR PreCheck line said it was a 10 minute wait. 15 minutes for normal PreCheck. 5 minutes for no PreCheck.

I decided 5 minutes wasn’t worth taking my shoes off, so I stuck with the CLEAR line. The guy behind me mouthed off to a CLEAR employee, “This line’s so long, what’s the point of paying all this money?!” He was told to go through the normal security line. And let’s be honest, Amex paid his membership fee.

Soon it was my turn to scan my eyes. “Random ID Check” flashed on the screen. I started getting annoyed as I dug into my bag for my wallet.

After the TSA agent scanned my ID, I loaded my bags into the X-Ray and walked right through the metal detector. “BEEP BEEP BEEP.” Another random check.

Am I on a security watchlist?!?!

As I stood waiting for three elderly women to get their hip implants manually scanned, I started bouncing up and down, watching my bag as it sat at the far end of the conveyor belt.

I started doing my best dad-looking-for-the-nowhere-to-be-found-waiter impression, craning my neck as I dramatically looked around for more TSA agents.

Nobody came. So I waited about three minutes, got a pat-down, then was reunited with my bag. A couple minutes after that, I had a steaming mug of Americano and a view of planes taking off inside the brand new Terminal C SkyClub.

Travel is so seamless now.

Instant chauffeur pickup via an app. Iris scanners to speed up security. A ticket that loads into Apple Wallet. Push notifications for delays. Lounges filled with free food and coffee while I wait to board a $100 million piece of equipment, maintained by an army of mechanics, piloted by an ever-rotating crew of employees, which will fly me hundreds and thousands of miles away. Usually on-time. Often early.

And yet there’s this pervasive frustration passengers have the moment they enter a terminal.

I somehow expect shorter lines and fewer delays from Delta Airlines than I do from the local Italian spot. A restaurant can bump my reservation, tell me they’re running an hour late, charge me $100/pp, and I’ll still recommend the experience to a friend.

But a flight that costs half of a dinner with wine in NYC? If they don’t get me there 10 minutes early, you’ll never hear the end of it.

by Will Schreiber

Will training data matter anymore for self-driving cars?

The coolest thing about these new LLM’s is their ability to handle few-shot learning. Give it a few examples, and GPT-3 will extrapolate that out to whatever else you throw its way. There’s no need for hundreds of thousands of pieces of training data just to classify a paragraph’s sentiment as “positive” or “negative”.

It makes me think a generally-intelligent system capable of few-shot learning will replace almost all neural nets trained on insanely big data sets.

Take driving. All humans in the world can learn how to drive in under an hour with almost 0 training miles.1 Put any teenager behind the wheel, show them the gas pedal and then how the brakes work. Tell them which side of a road to stay on. Tell them not to hit other cars. Stop at red lights. Otherwise, go.

And off they go, with shockingly few issues.

Tesla Autopilot has now been trained on what, a trillion miles driven? And they’re still having issues with roundabouts? I’ve watched tons of those Cruise videos too. While they’re insanely impressive, they also seem to suffer from brittle edge cases.

Self-driving neural nets seem to need training data for every single possible driving scenario in order to properly inch into traffic, take turns, and stop for pedestrians. Sort of like how IBM Blue needed to ingest every chess game ever played in order to take down Garry Kasparov, Tesla autopilot is insatiable in its thirst for training data. And even though it’s drowning in data, Autopilot gets nervous and stuck all the time.

Will a trillion more training miles really help much at this point?

Prediction: the first real self-driving system will be trained on less than 100 miles of driving.

Instead of being fed a billion or a trillion miles, we’ll simply show the system the rules of the road and off it will go - just like how GPT-3 only requires 1-2 examples in order to accurately perform classification, completion, and data extraction.

Billions and trillions of miles trained not required.

  1. I’ve previously written this about our narrow band of intelligence:

    We’ve spent over three decades, millions of man-hours, and tens of billions of dollars trying to teach computers how to intelligently stay between the lines.

    Yet, a few years before Alan Turing built the first computer, my grandfather was on a tomato field in rural Virginia. In two weeks, out of necessity, he figured out how to slip the red stickshift tractor-trailer into first gear, and then back to neutral. Into first gear again, and then back to neutral. Then all the way up to third gear and into town, to haul the tomatoes off. He was 11 years old.

    Nearly any human who has tried to learn how to drive has been able to do so in a short amount of time. Over 70 years since Turing’s first machine, we still don’t have self-driving cars.


by Will Schreiber

Two things have changed since 1990

I was born in 1990. Two things have changed since then:

  1. Water fountains
  2. Lightbulbs

A couple years ago, I landed in Albuquerque en route to Taos. The jet bridge from the Southwest plane to the terminal was a time machine to the 1990’s. Every store logo had that brutalist Seinfeld font aesthetic. The wallpaper was a blue-splash pattern like those old coffee cups, the windows were small, the brick was multicolored, and there wasn’t a water fountain in sight.

It took me back to elementary school, standing in line after gym, waiting for my turn at the dinky metal drink fountain. I’d push my entire body weight against the panel hoping for just a dribble of water.

Now all of those dinky fountains have add-ons sitting on top, which fill bottles at a torrential pace. And everybody seems to carry a bottle everywhere they go. When did this bottle craze begin?

The only change bigger than the water bottle has been the lightbulb. I dropped a plastic IKEA LED bulb yesterday. It hit the floor. Nothing happened.

I still remember my dad slicing his hand as he tried to catch a falling glass bulb. He probably juggled it because it was insanely hot after taking it out of its socket. The radiant heat from those bulbs made reading in the summer a tortuous event.

I do miss the startling “pop” you’d hear every once in a while, when the little wire inside the bulb would burn to a crisp and leave a black burn mark on the glass.

There are lots of things I don’t remember. I don’t remember NFL games being grainy. I don’t remember how all the movie trailers had that corny deep voice explaining “In A World…”. I don’t remember packages taking a long time to ship. I don’t remember wanting to listen to a song, but not being able to.

Everything seems the same in retrospect. Everything except the fact that humans were camels, and all homes were lit with a warm yellow glow.

by Will Schreiber

What’s in and out for 2023

I made this list on New Year’s Eve. So far, so good.


  • 1pm-9pm eating window
  • Bill’s Pizza Night every week
  • Daily writing
  • AI
  • Sunday tea
  • Fewer todos
  • Short sprints instead of long jogs


  • Scrolling in bed
  • Scrolling in general
  • Fake deadlines
  • Deli meat
by Will Schreiber

My experience Wednesday in SF

It’s pouring rain.

I take the Bart to Fidi. Everybody on the train is wearing a mask. As we roll to a stop at Embarcadero, the train loses power. We wait “for the generator to come on so we can open the doors.”

Finally outside, I pass 10 people on the three-block walk to 345 California. Half are “our unhoused friends.”

The Cafe-X coffee robot tent at the corner of Pine is no longer there. Somebody is bundled up, sleeping in its place.

I roll into Industrious at 9:15am. All of the bagels from the breakfast spread are gone. This isn’t San Francisco’s fault! Although since most private offices are empty, I’m not sure how I missed the rush.

I get some third wave drip coffee and begin my morning scroll.

I learn Facebook is laying off 13% of their employees. Just down Market Street, Twitter is embroiled in an internal war - payroll, servers, and debt payments all competing over declining ad revenue.

Worst of all, it turns out the guy with the curly hair plastered on every billboard around town is a total scam who rug pulled every FTX customer, creating an $8 billion hole in his balance sheet and sending crypto into a free fall. Bitcoin is down to $14k, not far off its price when I left San Francisco in August of 2020.

I take a Lyft to dinner. We get stuck behind two ambulances on Polk. As we idle, watching a woman get loaded onto a stretcher, I realize nine years ago I was picked up for my first-ever Lyft ride just a block away. I tell my driver about this personal history, about how magical it felt to order a car at the push of a button, how a Honda Accord adorned with a pink mustache pulled up to the curb, how I donated the suggested $8 to be driven to a bar near AT&T Oracle Park.

He asks me how much Lyft is charging me for this ride. $22.50.

He grunts. They quoted me $9.

I get to the bar where I’m meeting two of our investors. One has just come from a CEO roundtable. He decides not to drink because he has to “wake up early now.” Everybody is planning layoffs, he explains. People are burning millions a month. Capital markets don’t exist. I don’t think half of these people are going to have businesses in a year.

A couple hours later, I schlep back to Oakland. Lyft quotes me $0.50 more than Uber, so I flip a mental coin and go for the Uber. Boy does my gamble pay off. I’m matched with a Model 3.

In bed by 12:30, awake four hours later, and hungover from my three glasses of wine, it’s time to catch my Delta Main Cabin 5.5 hour flight back to NYC.

by Will Schreiber

I forgot my laptop

I got home and unzipped my backpack. My laptop wasn’t in its padded pocket.

I fired up Find My… and there it was, pulsing on the map, at Park and 17th about 1.5 miles away.

My first reaction was to panic. How am I going to provision these new accounts?

And then my second reaction was calmer. Maybe I don’t need to do anything tonight.

This is the first night I’ve slept further than 10 meters away from my laptop since our 2019 trip into the Kyrgyzstan backcountry.

So here I am, writing.

It’s nice knowing I can’t do the things I have the urge to do tonight. I was going to manually provision accounts for new Bottle 2.0 merchants — but I shouldn’t still be doing those anyway. Instead, we need to build an internal New Merchant form.

My dad jokes that in college his job was to tap the keg. For four years he tapped every keg. When I graduate, surely they’ll be living in a beer desert, he thought.

August rolled around and nobody called him. The keg got tapped. People partied.

My forgotten laptop is an act of forced delegation. I don’t have access to my console. All I can do is write a spec for an internal form, and then wait for it to get built.

No git checkout -b billy/fixes for me tonight.

… I do still have my iPad though :)

by Will Schreiber

My Kindle is dead

I ordered a paperback book from Amazon last night. The last book I ordered was a paperback, too. So was the one before that.

In every dimension, the Kindle is better. It’s slimmer, lighter, holds thousands of books, syncs to the computer, has a built-in dictionary, simplifies highlighting, and I can read it on red-eyes without turning on that glaring overhead light.

Yet there’s something about seeing The Wind-Up Bird Chronicle on the bedside table that nudges me to pick it up. There’s something different about turning the page, about feeling how much more’s left to read. Real books hit different.

I keep seeing surveys where workers say Work From Home is better than being in the office in every way. The commute is shorter, there’s time for deep work, it’s cheaper, it means more time spent with family and healthier lunches and fewer wasted meetings and less time spent getting dressed.

Last month, Andy went back into Industrious for the first time in a year. He called me as he walked home. “It’s amazing what we humans justify. I thought WFH has been fine, but after a day in the office, I feel like we’ve lost a year of our lives.”1

Yes, going in to work sucks. Everybody’s cramming on to the train at the same time. The bougie lunch stall is hawking $15 sandwiches. It’s a drag to shave and put on pants and talk about how “yeah this rain is crazy” 30 times. Being in-person is worse in every single dimension.

But dammit, I love running to catch that train. That’s my favorite lunch spot. And sometimes you’ve got to keep shaving to stay alive.

I keep getting introduced to people in NYC. Not a single person has suggested we do a Zoom call.

“Let’s grab coffee this week.”

  1. I am long remote work and flexible work. WFH != remote work. Bottle is a remote company. I’ve been bouncing around for the past three years. But the death of the office is vastly exaggerated.↩︎
by Will Schreiber

Decentralization is a narrative mirage.

Technology is a concentrating force. It always has been.

Everybody sewed their own clothes until we built textile factories.

It used to take one farmer to feed four people. Now each farmer feeds 130 people.

Home Depot killed thousands of local hardware stores. Opendoor is replacing legions of Keller-Williams agents. Salman Khan teaches algebra to more kids than the next 10,000 math teachers combined.

Gold, silver, USD, CNY, and EUR are all converging on the blockchain.

Human history is a story of increasing centralization. From roaming the plains of Africa, to settling down and building homes, to buying food in central markets, to instituting courts of law. Progression is compression. How can I make it so everybody isn’t making their own shirts? Deciding their own justice? Tabulating their own spreadsheets?

The internet, contrary to popular opinion, is accelerating the pace of centralization.

Software is centralizing in the App Store. Compute is centralizing in AWS-US-EAST-1. Culture is centralizing on TikTok. Stock trades are centralizing in Citadel’s dark pool. Politics are centralizing around personality. Work and cities are centralizing online.

Decentralization is a narrative mirage because three trends are frequently misinterpreted: the downfall of institutions, the death of the gatekeeper, and remote work.

These aren’t representative of decentralization, but instead are examples of increasing concentration.

1. “Institutions are failing.”

Personality is the force majeure of the internet. In a noisy world, it’s much easier to diligence an individual than it is to understand institutional motives. Who supports NPR? Who’s the NYT’s customer? Advertisers? PR people? Me? What are we being sold?

Institutions aren’t crumbling because of decentralization, but rather the opposite: institutional power is being compressed into personalities.

Trump became the GOP. AOC is becoming the DNC. Dave Portnoy is becoming ESPN. Sam Harris is now NPR. Elon is Tesla.

Organizational power is inherently decentralized. Shareholders vote for board members. Boards approve plans. Plans coordinate across loosely-coupled internal orgs.

But individual power, like autocratic government, is singular.

Directives no longer need to flow through layers of hierarchy. Tesla no longer needs a network of dealers to sell cars on, nor do they need an army of salespeople. Why listen to Bob at the local showroom when you can follow Elon on Twitter?

It’s the same in politics. We used to elect state representatives who would elect electors in the legislature who would elect a President. Then communication and transportation improved. We got railroads and national radio, so we put the Presidential candidates directly on the ballot.

Nowadays the President doesn’t even need TV stations: they can instantly drop messages into the pockets of 80M followers.

Betting on the internet means betting against institutions,1 not because institutions are exploding supernovas but because institutions are imploding black holes.

2. “The gatekeepers are dead.”

It’s true, gatekeepers are dead. It’s a huge improvement over the pre-Internet era.

I never learned about Richard Dawkins growing up. Then I went down a Carl Sagan Pale Blue Dot rabbit hole. I discovered Christopher Hitchens talking about evolution, consumed all of his religious debates, and then came across this Oxford evolutionary biologist who wore funny ties. The Selfish Gene fundamentally changed the way I view the world.

YouTube extends far beyond space videos. There are debates about quantum theory, DIY construction how-to’s, makeup tutorials, and an unbelievable amount of unboxing videos.

The best part? There’s no school board deciding what you can and can’t watch and there’s no MTV exec telling David Dobrik what he can and can’t upload.

It’s easy to think the unlimited nature of content is a decentralizing force. Yet the reality is that winning content in an unlimited world is watched over and over and over again.

Taylor Swift can reach billions of listeners, Salman Khan can teach algebra to millions of students, the best Stack Overflow answers get upvoted thousands of times.

Imagine if San Francisco had no zoning rules. There’d be cranes and giant skyscrapers everywhere. The city would be reaching Tokyo-level density, concentrating people in the Bay Area from across the country.

The internet is a supercharged no-zoning Bay Area. YouTube, TikTok, Amazon, Shopify are all limitless. Their success is exponential, like the weight of gravity, pulling in more and more eyeballs.

The gap between first place and second place is far more pronounced online than it ever was offline. 30-45% of all e-commerce goes through Amazon. Walmart managed to handle only 9% of physical retail at their peak.2

Power is shifting to the dancer who can get featured by TikTok’s algorithm, to the business that can place first in Google’s results, to the reseller that pays the most to top Amazon’s catalog, to the singer that can get featured in Spotify’s Discover Weeklies.

The long tail of content is a distraction from the reality: online power is concentrating in an extreme power law. Yes, anybody can build a following. But the winners dwarf the long tail.

After all, audiences are so concentrated that it only took three companies to silence the sitting POTUS.

3. “But what about remote work?”

I miss going into Industrious. I’d wake up, walk to FiDi, pour myself a cup of coffee in a grey mug, sit in a mid-century modern green chair, and get to work.

When I flew to Chicago, I’d take a train to the Fulton Market Industrious, pour myself coffee in a grey mug, sit on a mid-century chair, and get to work.

Fly to Atlanta? Ride to Ponce. Coffee, grey mug, green chair, work.

New York? Union Square. Coffee. Chair. Work.

As people move their work lives online, the experience everywhere is going to be the same. Morning ride on the Peloton, get to WeWork, drink some Blue Bottle, hop on a video call, order groceries from Instacart, rinse and repeat. The only difference will be whether we see palm trees or wildfire smoke out the window.

The “everyone is leaving SF” narrative isn’t about decentralization, it’s about centralizing online. Everybody is on Zoom now. “Remote work” doesn’t mean you can now live some exotic life in a faraway place, it means that people in Birmingham can live the same as those in Cow Hollow. And work for the same companies, too.

Not only are we all working on Zoom, but we’re all shopping on the same “hip” street, the one with Warby Parker and Bonobos and Lululemon and La Colombe all lined up side by side. Sub-in Blue Bottle, and I’ll tell you if you’re east or west of the Mississippi.3

Part of it is that capital is cheap and management can be centralized, so chains are able to grow quickly nationwide without leaning on franchise models.4

But the other part is that our tastes as Americans are compressing. Every town has incredible breweries, third wave coffee, Trader Joe’s, and a Costco out in the suburbs. We’re all watching Netflix and Charli D’Amelio. We’re all peddling to the same bike instructors as Biden. How much longer until we all lose our regional accents, too?


Everything is shifting online. From how we find office space, to get rides, to feed ourselves, to exercise, to work and to live.

The meme that this constitutes “decentralization” is like the ocean on the edge of the Saharan horizon. It is an illusion.

The spectrum of opinions, of tastes, of behaviors is compressing. The platforms and personalities with power will continue to grow in size. The power law of the internet remains irrefutable.

Soon one farmer will feed 1,300 people, robots will sew our Nikes, there will be one nationwide math teacher and a single wealth ledger.

And our lives will be fully uploaded.

  1. If you want to bet with the Internet, bet against institutions.

    — Naval (@naval) February 1, 2021



  3. I’m going to get flack for “being in a bubble.” I’m just using upper-middle-class-young-professional as the example. But shared tastes are true up and down the socioeconomic spectrum. Whether you’re shopping at Aldi or Whole Foods, or eating at Chili’s or Soho House, your counterparts across the country are doing the exact same thing.↩︎

  4. Franchises are dead

    OEM to dealerships -> Tesla
    KW/Coldwell/Compass -> Opendoor
    McDonald's -> Starbucks/Chipotle

    Franchise model used to unlock capital and management. Now you've got 0% interest rates and computers.

    — Will Schreiber (@breakfastbybill) January 27, 2021


by Will Schreiber

Dark and dirty dorm rooms

It’s endearing how college kids abbreviate everything.

“How’s that prof?”
“Meet you in the caf.”
“Skirt skirt.”

I lived on “Lup 7” in college. There were 16 of us crammed like sardines in cinderblock cans at the end of the hall. Two to a room.

One Thursday, I left Tin Roof early. I had an early 10am class the next day! I got back to our floor, microwaved some Bagel Bites, put headphones on, and climbed into my lofted bed.

As I closed my eyes, I heard laughter coming from down the hall. The noise got louder, then louder, and then “click.” The door opened. My roommate had brought somebody home.

I tried not to be annoyed. My monitor’s glow usually keeps him up, I thought. Karma.

As I pretended to be passed out, I imagined being an old man, waking up one day in a big house, a big house with vaulted ceilings and floor-to-ceiling glass, sitting on a Restoration Hardware couch looking out at snow-covered pine trees, sipping espresso all alone, wishing I was back here, right now, surrounded by friends, dark and dirty dorm room be damned.

Then I fell asleep.

by Will Schreiber

Four years ago

Four years ago, I was in Nashville at a friend’s house, eating shrimp, drinking Bud Light. The election was called not long after we finished dinner. I remember thinking even Trump looked surprised.

Sam and I walked back to our house on the other side of 12th South. It was cold.

Two weeks later, I loaded my bed and Apple Cinema Display into a Penske truck and drove up I-65.

I first lived in Gold Coast, walking distance to The Loop in downtown Chicago. I went to Carmax to sell my car. I bought my first down coat. I started reading Vonnegut novels. I fell in love with riding the L.

Then I moved up to Lakeview. I got a tortilla press. I started a sourdough culture. We paid ourselves with money from Bottle for the first time.

Our apartment overlooked the Paulina Brown Line stop. I worked from my bedroom, but enjoyed watching people run from Starbucks with their coffees in hand, trying to catch the morning train.

Then I sold my Ikea bed and Ikea desk and Ikea chairs. I packed a 55L Osprey backpack. We left Chicago.

I stocked up on Rx Bars from the Midtown Atlanta Industrious. I didn’t know if I could trust China Southern’s inflight dinner.

I flew to Phnom Penh. We looked at all those temples near Siem Reap. We crossed the border into Laos, went backpacking in the jungle, ate noodles in Pai, crashed a motorbike, lived a yuppie month in Bali, slept out of a minivan at the bottom of Mt. Cook, spent the winter skiing in Sun Valley, and drove a Subaru from San Francisco to Bar Harbor, Maine.

Then we settled down again. We moved to San Francisco. We landed in SOMA before moving to Alama Square.

I went to Napa for the first time, got a coworking membership, and even bought an Arcteryx bag for run-commuting.

Then we quarantined. Then we bought a new Subaru. Then we left.

It’s a new election night. I can’t believe it’s been four years.

by Will Schreiber

Inside every brown take-out bag…

“I wanted Uber Eats because it was raining. But didn’t end up ordering because it was raining.” - Recent text from a friend

Paying for somebody to deliver burgers and fries usually feels fine. They could decline the gig if the price isn’t fair.

But when you’re sitting on a bench across from a mid-forties man rubbing his knee and popping Advil, or when a woman is standing outside the door, dripping wet, brown bag in hands, glancing at the gray Scandinavian couch and 65″ Sony on the wall, guilt creeps in.

Maybe Doordash and Instacart aren’t stealing the tips. Maybe they are. I don’t know. But I do know something feels weird about paying somebody to mask up and pluck items off the shelf while I sit on the couch. For only $8!

I’d order more often if I didn’t feel guilty.

But the gig economy is stuck in a prisoner’s dilemma. Doordash can’t raise prices to pay Dashers™ more. If they did, everyone would use Postmates or Uber or Amazon instead.

Game theory means it’s cutthroat prices and cutthroat wages. Game theory means inside every bag of hot food left on the doorstep, there’s a little feeling of shame. They could decline the gig if the price isn’t fair.

When price for delivery goes up, people order less often. Same in the other direction. But I wonder if the gig economy prisoner’s dilemma is suppressing overall demand for food and grocery delivery.

If seeing Palm Oil on an ingredients list didn’t make customers think about deforestation and global warming, then more products would have palm oil.

The same goes for delivery and labor practices.

by Will Schreiber

Acting like a Github recruiter

I interrupt the highly irregular but continually aspirational drip of daily Second Breakfast prose to briefly act like one of those recruiters who scrapes your email off Github.

Two things.

  1. We’re looking for a Rails developer who can contribute about 60 hours/month to help us ship an amazing Bottle 2.0.
  2. My friend Jane and her cofounder Sabrina are looking for a founding engineer for their women’s health company, Pollie.

First, Bottle.

Bottle combines memberships + texting + order-ahead to fix retention for local businesses. Bottle merchants frictionlessly text with their customers, build followings, and remind members to order without forcing subscriptions. Drives higher weekly revenue.

We started Bottle in 2016. We’ve grown using the money our customers pay us every month. And now we’re building Bottle 2.01 with Rails and Vue.

We need a Rails developer to help us with the backend. This developer will work with me to plan the app’s logic and build necessary API endpoints. We use Heroku, Postgres, JSON:API, Pusher, Algolia, and coffee.

Perks: you get to laugh at all the mistakes we’ve made.
Cons: you have to cry at all the mistakes we’ve made.

Second, Pollie.

My friend Jane and her cofounder Sabrina started a women’s health company called Pollie to improve the patient journey for PCOS, endometriosis, and other hormone imbalances. I love what they’re doing.

They’re hiring for a founding engineer role.

Hit me up or email will @ sendbottles if you’re interested in chatting about either.

  1. Here’s the Bottle 2.0 checkout experience we’re building. Now we need to bring it to life: