Fear and self-loathing in Silicon Valley

I somehow ended up at YC Demo Day.

Where are we taking everyone?

Hundreds of us crowded by the door, as the security guard slowly let us in ten at a time. We looked like the boarding line of flight to Seattle, if the plane was nothing but business class: A disorganized traffic jam of middle-aged men—and we were mostly men; I counted the first fifty people I saw as I scanned the crowd; 43 were men—wearing our nicest jeans, our tastefully-branded fleeces, and carrying tactical backpacks that betrayed our best efforts to be corporate chic. We talked loudly about AI, about money, and about how Silicon Valley’s leading startup incubator should have a more streamlined registration system.

We were there—piled in front of a refurbished building in Dogpatch, which is a refurbished neighborhood in San Francisco—to attend Y Combinator’s Demo Day. Y Combinator, or YC as it's more widely known, is an elite startup accelerator that now incubates about 500 companies a year, spread across four seasonal batches. Demo Day is the program’s capstone: An all-day marathon of pitches, one matriculating YC startup after the other, presenting to an eager audience of hundreds of investors.

We were the investors. Or, they were the investors; I don't know why I was there. I had been inexplicably sent an invite,¹ happened to be in San Francisco last week, and was curious to see the spectacle in person. Demo Day has always been a cultural lodestar in tech: Dozens of famous startups and founders, from Airbnb, DoorDash, Coinbase, Instacart, and Dropbox to OpenAI’s Sam Altman and Scale AI’s Alexandr Wang, debuted on its stage. It is a stylistic measure of the Valley’s mood; of the industry’s bleeding edge ambitions; of what is rising and falling; of what’s hip with the kids these days. There is also a free lunch.

It’s an imperfect scale, of course, in no small part because YC’s own thumbs are all over it. The accelerator has a “tradition of sharing ideas [they’d] like to see founders tackle.” I heard from several more seasoned attendees that YC has, in recent years, preferred investing in enterprise products over consumer applications; in this particular batch, there were also requests for an even greater emphasis on “boring” industries like accounting and compliance.

An army of founders had heeded the call. According to our host’s opening remarks, the 140 or so companies that were about to hear from had been selected from a pool of more than 15,000 applicants. Some of these companies will be the next decacorns and hectacorns, the emcee told us.

“Happy hunting,” he said.

—

You could follow along with the pitches at demoday.ycombinator.com, a custom site that YC set up to help investors find their favorite startups and contact their founders. When I first typed in the URL, I accidentally left a letter out of “combinator.” I landed on a phishing site, the sort that immediately assaults you with popups and pharmaceutical scams, offering me magic remedies for a variety of ills.

Something about it felt appropriate. Our host was probably right: There are the seeds of unfathomable wealth in this room, if you pick well. But if you’re off by one pitch—one erroneous choice, one wrong keystroke—you could instead give your money to something that first promises you the world, and then blows you up.

—

In its early years, when YC accepted fewer companies, startups gave product demos on Demo Day. Now, it’s a misnomer—the demos are a single static slide, and a brief one-minute pitch from the company’s founder.

All the presentations had the same cadence; the same sing-song rhythm. Maybe that’s because they were all coached by the same playbook; maybe that’s because you can only say so much in a tightly-governed 60 seconds; maybe that’s because, after 100 pitches, they all begin to bleed together into one miasmic blur of optimism, chesty credentialing, and inventive accounting.

Two founders walk on stage; usually young; also usually men. The speaker is stage left; their silent partner—they has nothing to say; they weren’t allowed to have anything to say; only one person was given a microphone—stands to their right, trying to figure out what to do with their hands.

They tell us about their vision: “We are Stanley,” the founder begins,² “and we’re building the future of construction workforce automation.” A pause, to let us take in the gravity of what they’re saying. “Stanley is the first-ever AI agent purpose-built to help general contractors optimize the output of their day laborers. Construction sites lose $150 billion a year because people aren’t working as efficiently as they could,” says a 23-year old with a computer science degree from Carnegie Mellon, “and we’re the team that’s going to fix it.”

They tell us about their résumés: MIT; Stanford; Harvard, read even more proudly if they dropped out. They worked at Stripe; they worked at Google DeepMind; they worked at JPL; they worked at McKinsey, and now know better. I don’t remember anyone who worked at OpenAI, because former OpenAI employees can make plenty of money without needing YC’s help.

They tell us about their traction: Some companies used real financial metrics; others made stuff up. We’re at $600,000 in ARR, and profitable. Our cARR is $120,000. We have LOIs for $450,000. We’re in negotiation for over a million dollars of new contracts. MRR is up 2,000 percent week over week, to $8,000. Our Series A ACVs are $100,000.³ We have users from half the Fortune 500. We have signups from half of the Fortune 500 on our waitlist. We have reached out to half of the Fortune 500 on LinkedIn. Something is up and to the right; something is exponential; something is going vertical.

They close with the same light punch. “We are Affirmative Health, and we are reinventing how clinics can for their patients,” said a software engineer. “We are Scurry, and we are solving logistics with AI,” said a founder wearing a QR code on his t-shirt. “We are Dev Defender,” said the one person pitching Dev Defender.

—

After the last pitch, the founders and investors all milled about during a closing happy hour. A robot mixed cocktails. The founders repeated their pitches. The investors did their best to also impress the founders, but you can only be so impressive with a backpack on.

—

As we were talking, some 400 miles south of us, Los Angeles was going to war. Marines were being deployed into the city; protesters were taking rounds; marches were being organized around the country.

For nearly a decade now, the world has felt like it’s been combusting, and last week was particularly acute—deportations, the crisis in Gaza, the threat of nuclear armageddon in Iran. Our politics volcanic; our society strained; the world a feverish train barreling down the tracks, faster and out of control, shuddering louder and louder now.

At Demo Day, we were unphased by the external chaos; cloistered from the outside world. The invite to Demo Day—held two blocks from a subway stop—said “NO parking available on-site. Please rideshare.” On the floor below the main stage, TBPN, a trendy daily talk show about technology—the All-In podcast on Twitch, with the manic visual aesthetic of Squawk Box—was broadcasting live. The show’s on-air hosts interviewed YC founders; the show’s representatives walked around the YC complex wearing highlighter-yellow hats, with the confident swagger of celebrity.

That was the energy of the day: Ascendent, triumphant, “festive,” as a YC executive put it. People took selfies with founders and in front of neon signs. “These things are always exciting," said the Atlanta-based investor sitting next to me. Technology had won and was exiting its villain era. Or was it entering its villain era? The atmosphere—of Demo Day, and of San Francisco more widely—wasn’t the nerdy computer lab of the 2000s or the delinquent frat house of the early 2010s; it was the country club that the frat boys grow up into. Rich and unashamed of it; unrepentant and unthreatened; free from the yolk of broader responsibility. That freedom didn’t come from reform though, or as a reward for successfully “making the world a better place;” no, today’s Silicon Valley is free because it no longer cares. It no longer needs to care; that branding is no longer useful.

I went to another event after Demo Day, and an investor at a blue chip firm commented that many founders these days, including those who are still in college, are as well-versed on the financial mechanics of startup fundraising as he is. He said he thought it was because they’ve grown up Very Online, raised by Content, mainlining tech Twitter and shows like TBPN before they can drive. Maybe, though my sense is their sophistication is from something else: Startups are just businesses now. They aren’t the passion projects of technical enthusiasts who want to build something for the sake of creation and curiosity; they are an opportunistic hustle. They are unapologetically about making a lot of money. Esoteric clauses on fundraising term sheets do not matter to the technologist, but they do matter to the capitalist.

The inverse is true for politics and social concerns: They are below our line. Which isn’t to say that today’s generation of startups are unprincipled; it is to say that they are unburdened. They appear to be, as much as they can, clinging to the coconut tree, avoiding the social context in which they live. There is the company; there is its market, its customers, and its competitors; everything else is a yawning void of distraction.

—

No, wait, that's not right. The discourse did come up once. “We’re starting with sheet metal,” said the founder of a startup hoping to build autonomous factories for manufacturing sheet metal, “and we’re staying, because the future of western democracy depends on it.”

—

Maybe that’s the point, though. You join Y Combinator—or simply move to San Francisco—to put your ideas through a pressure cooker: Seal them in, kindle them faster. Does the founder of a company really have any obligation other than that? Zuck is not unloading a titanic cannon of cash to develop a socially-responsible AI; he’s doing it to defend his empire. His competitors have to operate with the same singular obsession. Startups cannot opt out of the communities that cultivate them, but they can’t opt out of the ruthless maw of the markets either.⁴ Worrying about anything other than survival is, in some sense, a privilege. Relative to their recent predecessors, are today’s founders unburdened, or are they simply focused?

There used to be an undercurrent of self-loathing in tech, or at least a twinge of shame. We sat behind desks and typed secret codes into a computer, and then became the richest professional enclave in history because of it. Everyday employees were paid embarrassingly well for doing so little; tycoons were paid embarrassingly well for breaking so much. It was gauche to talk about it, so we pretended it wasn’t true.

That shame seems gone. Perhaps that’s healthy? Perhaps that’s self-care? Perhaps that’s just a more convenient belief to hold—it is easy to get a man to believe something, when his salary feels more deserved when he does. Or perhaps this is just an industry maturing, from a quirky upstart, through its ideological adolescent, and into a more pragmatic middle age.

Still, there is something striking about the audacity of Silicon Valley’s current ambitions, and the narrowness of its field of view. We saw startups that want to replace accountants, sales people, and paralegals. They want to automate manufacturing and build lights-out factories; they want to streamline health services, pepper the country with robot car detailing services, and fully replace grocery store supervisors with AI agents that monitor checkout clerks’ performance. They want to change how we develop new drugs, how we get mortgages, and how teachers teach math.

These may well be good things; they may be progress; they may be the application layer of the gentle singularity. Or they might also be the first dominos in a long cascade of unintended consequences. What happens when stores are run by inscrutable electronic managers? When we all have personal assistants doing our chores? When we can conjure worlds on a whim? When we forget how to think? What happens when we request that thousands of 20-year olds reinvent entire industries, for the sake of their own entrepreneurial aspirations first, for our returns second, and for the consequences third?

What happens when we split this many atoms this quickly; when we throw off this many rogue neutrons? Do we have any idea what they might collide with next? Are we building worlds, or destroying them?⁵

Truly, I don’t know. Nobody knows. But that is tech now: The car is unstoppable—ambition is a hell of a drug, and startups are going to build big things, regardless of whether or not YC asks them to do it—and I have no idea if I should be excited or afraid.

—

Early in the day, a startup pitched a product that makes it easier for health care providers to fight denied insurance claims. More than $260 billion in claims are denied every year; with denials on the rise, the system is collapsing under its own weight, they said. Our agents will help hospitals rightfully recover the money they’re owed. Later, another startup pitched their insurance claims product—a platform for insurance adjusters to automate how they collect and initially evaluat claims.

An AI agent for both sides. An arms race; a relentless fight between competing intelligences, with Silicon Valley’s investors and compute providers supplying both sides. Will this fix our miserable health care system? Will it make an insurance industry that literally drives men to murder even more ruthless? Or will the chatbots slaughter themselves into a stalemate?

—

The day I left San Francisco, I spent the afternoon sitting in a coffee shop in Chinatown.⁶ Two men in their twenties were sitting next to me. One appeared to be pitching the other, though I couldn’t tell who was who—most of the conversation was one-upmanship of predictions and bold takes: AI will replace all user interfaces; no, of course, React doesn’t make any sense anymore; well, obviously, in the future, we won’t need keyboards.

At one point, one of them asked the other when he last left the Bay area.

“I don’t remember,” he said, “it’s been a while.”

“You really need to get out. I’m from Florida; people are different there. It’s good to talk to them. The way they think about software and AI…they’re a very different type of buyer.”

That’s a start, I suppose. If we’re going to rebuild the world, best to visit it—even if only to sell to it—first.

1. Perhaps it was a mistake. Perhaps someone mixed me up with that other two-n Benn on Twitter, who appears to be a real investor. Perhaps an AI bot found benn.ventures and thought it was real.

2. This isn’t a real company, nor are any of the named ones listed below. My point here isn’t to call out any specific company (which are all listed in a public directory), but to describe the character of the event. The other quotes are as accurate as I can remember them, but my memory is blurry and my notes were loose, so consider the details here more atmospheric than journalistic.

3. More completely: “We are making $150,000 in ARR, and our series A ACVs are $100,000.” Which I think means they have one customer that has raised at least a series A, and that one customer pays them $100,000?

4. Similarly, just as startups can’t opt out of politics, politics can’t opt out of technological progress. With or without YC, this stuff is coming. Nobody can stop that reality; we can only hope to guide it.

5. Of course, one possibility is that nothing happens. Maybe self-interested VCs are exaggerating the potential of AI; maybe it’s all just a bunch of hocus pocus (source: A mural, in SF). But if it is, man, this is going to leave behind one unholy crater of technological and financial wreckage.

6. Pour overs and industrial minimalism have nothing on Cookie Crisp lattes (which have nothing on Cinnamon Toast Crunch lattes).