We are insufferable.

And I say this as someone who is very much one of us.

I'm a software engineer through and through. I think about distributed systems problems in the shower. I talk about work issues on dates. I have opinions about infrastructure. I'm not not a part of this.

And that's precisely why I feel confident speaking up about the culture that we have created for ourselves.

Last week, a founder launched "ABG CMO," an AI tool that generates a fake Asian female influencer for users to post content under. The product's pitch: "Design a hyper-realistic persona in seconds. Pick a look, a niche, a vibe."

On X, the replies came quickly. "So sick!" someone wrote. "I'm going to get 10 CMOs." The founder responded: "who's your ABG CMO boss man."

The whole thing left me gravely disappointed.

ABG is already a reductive abbreviation in its own right. It stands for "Asian Baby Girl." It's a term with enough internet history that ignorance is not a credible defense. And this product doesn't just invoke the ABG archetype—it makes her puppetable. You pick her look, her niche, her vibe, and then you profit off her image without her existing as a person at all. The intended customer, per the site's own testimonials, is overwhelmingly male founders and growth guys. These men are operating in an industry that is, by its own numbers, predominantly male. The same rooms that have no space for real Asian women are buying a tool to manufacture fake ones for profit.

I keep asking myself what emboldens this. What makes a person so thoroughly unconcerned with the weight of their own actions?

The answer comes from the top.

When the figureheads of an industry begin eroding the standard of how we conduct ourselves—that is, diminishing the importance of professionalism or the responsibility of upholding a position of visibility—it doesn't stay contained to them. It trickles down to the little guys. It always does.

And what we have arrived at now is not the meritocracy that was promised. Instead, we are building a wildly regressive and destructive framework in the name of growth and innovation.

---

Paul Graham, co-founder of Y Combinator and one of the most influential figures in startup culture, wrote an essay called "Good and Bad Procrastination." It is, on its face, about productivity. But embedded in it is a philosophy that has been absorbed—and badly misapplied—by a generation of founders and builders.

Graham's argument is elegant: the most impressive people he knows are all procrastinators of a specific kind. They put off small things to work on big things. They blow off errands to pursue what he calls "real work." And he recognizes that this comes with a social cost: "The people who want you to do the errands won't think it's good. But you probably have to annoy them if you want to get anything done."

Deep work requires protection. Sustained attention is a scarce resource, and guarding it is not selfishness—it is a prerequisite for producing anything of consequence.

But what's missing from this framework is what we owe each other. We take from our communities constantly—the infrastructure, the institutions, the people who show up and extend good faith to strangers. And I believe that the social contract we agreed to implicitly by dwelling in these communities deems that we must give back in equal measure. We hold doors. We extend empathy. These are not niceties. They are the load-bearing walls of any community worth belonging to.

Graham is not wrong that some errands can wait. But a generation of builders heard, in this ethos, that they were all working on something important and therefore exempt—that the social contract itself was the kind of errand Graham was telling them to skip. Why return the favor? Why participate? Why move through the world as though other people exist and matter? They were building, so the externalities were someone else's problem.

I do not think that most technical endeavors in this day and age are so great and so urgent that they preclude a person from doing the things we are all but contractually obligated to do. I do not think that building an LLM wrapper, or founding a startup that will pivot twice before being acqui-fired into obsolescence, exempts you from basic human decency. I do not think that ambition is a license for abrasiveness.

But we have made it one. We have made it a signal. The abrasive founder is not a cautionary tale. He is, in the current cultural register, something to aspire to.

---

I'll say it plainly: this pattern of behavior makes you a bad person.

You are not a misunderstood visionary. You are not someone too focused on the work to bother with niceties. You are just a bad person. Or at least, someone who has decided to act in ways that are bad, which amounts to the same thing over time.

And the consequences of this extend well beyond the personal. The same logic that licenses the abrasive founder—that his work is important enough to exempt him from ordinary consideration—is the logic that builds exclusionary rooms. If other people's feelings are errands, then other people's presence is optional. And when presence becomes optional, the room fills with whoever was already comfortable in it, and it stays that way.

This matters not just as a moral observation but as a practical one, because there are real and measurable costs to making a community hostile to the people who aren't already inside it.

In 2020, Timnit Gebru—then the technical co-lead of Google's Ethical AI team and one of the most prominent Black women in AI research—was fired. The circumstances were contested; Google maintained she had resigned. Gebru maintained she was pushed out after submitting a paper on the risks of large language models that the company found inconvenient. The letter condemning her departure was eventually signed by thousands of Google employees and academics.

What was lost was not just one researcher. Rather, we lost the institutional capacity to scrutinize ourselves—to accept the presence of someone who is asking, with rigor and credibility, whether the thing being built is good.

Gebru's work on algorithmic bias had shown that a particular facial recognition implementation was 35% less likely to recognize Black women than white men. This is the kind of finding that emerges from a team that looks different and has different stakes in the outcome. When you remove that person—when you make the room inhospitable enough that she leaves, or is pushed out, or never even arrives in the first place—you don't just lose her, you lose the question she was asking. And the people left in the room don't always notice, because they were never the ones the question was about.

What you get instead is a room that has never had to justify itself. The assumptions go unexamined because everyone shares them, and the blind spots go unnoticed because no one is positioned to see them.

And then we formalized it. The coordinated dismantling of DEI programs—the contempt expressed, by powerful men in powerful rooms, for the idea that the composition of a team might matter—was a ratification. The room had already decided who belonged in it, and the policy made it official.

Meanwhile, a billionaire venture capitalist had gone on his podcast and said that a genocide is "below his line". That other people's suffering was, on his personal priority list, not worth his attention. He called it a "hard ugly truth." His co-hosts pushed back a little, and then the episode moved on, and the world kept turning.

This is what the room sounds like when there has been no one in it, for long enough, who might have made him think twice.

The exclusionary club harms the people it excludes, the work they produce, And eventually, the character of everyone left inside it.

I want to speak directly to the person who finds this unremarkable—who hears "a genocide is below my line" and thinks "at least he's being honest."

Moral consideration is not a distraction from the work. The insularity that makes caring about a genocide feel optional is the same insularity that produced a facial recognition system 35% less likely to recognize Black women. The monoculture is a liability.

In 2014, Apple launched a health tracking app comprehensive enough to log blood copper intake and blood alcohol levels but could not track menstrual cycles. That is half the population completely unaccounted for. They shipped it, were publicly humiliated, and spent the next year fixing what a more diverse room would have caught before it ever went out the door. The moral argument quickly became a product failure with a price tag.

You do not have to care about people to understand that building without accountability to anyone unlike you is a bad engineering decision. But I would also like to submit, with full awareness that the people who need to hear it most probably won't: you should also just care about people. It is the bare minimum.

These are the large and visible failures. They are the end state of something that begins much smaller and much closer to home—in the everyday decision to take from a community without giving back to it. To let someone else hold the door. To skip the favor, ignore the ask, move through shared spaces as though they exist for your benefit alone. None of these things feel consequential in the moment. But they compound. Each small abdication makes the next one easier, and the one after that, and eventually you have a culture that has practiced not caring about other people for so long that a genocide barely registers as worth mentioning.

The chores matter. And the person who won't do them is practicing the belief that other people's needs are not their problem. That is a belief with consequences that do not stay small.

---

I believe in this industry. I believe in what's being built. I work in infrastructure, and I've spent the last two years thinking very hard about technical problems that I find genuinely captivating. I'm not immune to the pull of it—the seductive logic of 996 culture, the idea that total immersion is the only honest way to build something that matters. I understand the appeal. I feel it.

But I will still clean up after myself. I will still be kind to the people around me. I will still uphold my end of the social contract—the small, unglamorous, non-negotiable part of it—because exemption from the social contract is not something you earn by being busy. Graham's absent-minded professor forgot to shave. That is endearing. What's less endearing is the person who remembers perfectly well that the sink is full and has simply decided it isn't their problem.

I believe that technical excellence and basic human decency are not in competition. I believe this not as a platitude but as a conviction I have arrived at by watching what happens when people treat them as mutually exclusive. The work suffers, the room suffers, and the people the room was never built for suffer most of all.

We have glorified founding. And more than that, we have glorified the character that is now associated with it—the ruthlessness dressed up as focus, the contempt dressed up as sincerity, the bad behavior dressed up as a side effect of genius. We have made the social contract optional and called it a feature.

This is not the industry we should be creating for ourselves. We cannot build by making ourselves a burden to the communities that hold us. We can do exceptional work and operate with empathy. The two have always been compatible.

That is not a high bar. We have to clear it.

We have to bring back decorum.