
It is popular these days to talk about the future of software engineering. Among both people who are qualified and those who are not, everyone has their careful research, considered predictions, and the occasional ramrod opinion.
It is also popular these days to talk about the future of software engineers, but those conversations tend to be more blunt: Most people say that the career is barreling towards some sort of existential reinvention. As the robots get better at writing code, developers will do less of that particular task, and more of the tasks that orbit it: Architecting software, reviewing code, orchestrating agent workflows. Today, one of the defining skills of a software engineer is their ability to write pedantic technical incantations into code editors; tomorrow, the best software engineers will be those who are good at project management and robot resource allocation.
And, above all, the best software engineers will have taste. When you can manifest entire applications by simply asking for them, great products can be built by anyone—but not everyone can build a great product. So, over and over and over again, we’re told that taste will be the defining skill of the coming AI age. Because the machines can do, and the machines might be able to think, but the machines cannot feel. And so the engineer must. To be among the next generation of great builders, don’t be a technical genius; be a tastemaker. Don’t look like Ken Thompson; look like Rick Rubin.1
There is an obvious assumption embedded in this advice, though: That taste wins. That taste begats craft; that craft begats quality; that quality begats popularity. It’s a pleasingly wholesome loop—in a just world, isn’t this how things should work?
But in every competition, there are always cheaters.
Here’s a clever post by Ethan Ding about Cluely, that “cheat on everything” startup whose founder has bulldozed his way into the discourse a few months ago after being suspended from Columbia:2
[Cluely’s] AI coaching tool was vaporware, but they bet that conversational AI would eventually work. so they raised $5.3M and spent it on outrage—a manifesto calling cheating "leverage," a viral video of their founder lying on dates. 7.8 million views. "black mirror dystopian," everyone screamed.
perfect. because when real-time AI coaching actually works, everyone will remember them.
the genius is in the timing calculation:
the tech WILL get good
they need to survive until then
they need to own the mindshare when it arrives (hence the outrage)
…
these companies know their products are garbage. they're not delusional—they're patient. they're buying lottery tickets on GPT-6, on the next generation of video models, on the inevitable march of progress. their 90% churn isn't a failure metric; it's a holding pattern.
….
this is actually the correct strategy. in a world where foundation models improve 10x every 18 months, building infrastructure is like decorating a house you're about to demolish. the smart money isn't on building the best product - it's on being the default option when the underlying tech finally works.
the companies spending millions on R&D are playing checkers. the white-label ChatGPT wrappers are playing chess.
In other words: Taste and quality are dwarfed by marketing and mindshare. Though this may have always been partially true,3 it’s especially so today, because every new product sits on top of the same foundational models. Historically, early-stage startups prided themselves on how much they spent on engineering and product development, and how little they spent on marketing. Now, as Ethan says, that’s inverted—a dollar spent on building a brand is potentially more useful than a dollar spent building a product.4
$CHEAT
I’m not sure that story goes far enough though. Because, what exactly is Cluely?
On one hand, sure, it’s a startup that wants to be a successful business. Though that could mean different things—that Cluely hopes to make more money than it spends; that it hopes to get bought for more money than it raises; that it hopes to do neither but wants to make the world a better place—all of these answers are approximately the same. They say that Cluely exists to serve the interests of its shareholders, employees, and customers in roughly the same way that nearly every other entrepreneurial endeavor does.
On the other hand, that feels…insufficient? Cluely isn’t just a business; it’s also a vibe; a performance art project; an opportunistic meme coin. Despite ostensibly being a technology company, Cluely is more slogan than software. It’s not for the customer; it’s for the culture. And above all, it’s for the founder, as a vehicle for turning monetizing a viral fifteen minutes of fame.
Because here is another description of Cluely, but this time from Matt Levine:
A young man named Roy Lee was suspended from Columbia University after building a tool that let them cheat on job interviews. After the suspension, he wrote a funny, charming, somewhat irreverent Twitter thread explaining what happened. The next day the post went extremely viral and Lee became famous as the “cheat on everything” guy.
Lee was not, prior to this, part of the fame economy: He was a regular student, was not an influencer, and he shared the story from his own Twitter account. But once he went viral he was maximally interested in becoming part of the fame economy. How does one do that? How does one parlay one brief funny viral story into a lucrative career? There are well-understood approaches. You can probably get into reality television, perhaps followed by politics. You can get paid to promote products — in ads, or on your own social media — or to show up at nightclubs and other events. You can sell merch. If you can sing or act, you use your viral fame as a way into a singing or acting career. …
So! I mean! Here you are! You have become extremely famous for one thing. Parlaying that into a long and lucrative fame-based career probably will require doing other things: singing or acting or at least showing up at nightclubs. But you have a ton of meme value right now, even if you probably won’t in a year. So the thing to do is to capitalize that meme value into a meme company, give yourself most of the equity, and then sell shares to venture capital investors.
Ah, no, wait. That’s not what Matt Levine said about Roy Lee and Cluely; it’s a slightly adapted version of what Matt Levine said about the Hawk Tuah girl (née Haliey Welch) and her memecoin.5 But it’s kind of the same story, right? They were both at the center of a viral moment; that moment became singular to their identity; they immediately embraced that brand, and launched various ventures to capitalize on it. Though Lee’s version is more substantive than Welch’s—he says he tried to intentionally engineer his viral moment; he started a company while she launched a meme coin—there are clear parallels.
Which, makes sense! As Levine says, if you become an overnight celebrity on the crypto-obsessed Zynternet,6 “modern finance has created an incredible tool for directly monetizing your 15 minutes of fame,” and “this tool is the memecoin.” And if you are an engineer and become a niche celebrity on tech Twitter, Silicon Valley has created a very lucrative tool for monetizing that fame: A startup. Historically, that hasn’t been a terribly efficient tool, because you have to build a product and company, and that is, you know, hard. But today? You can incorporate a business in an afternoon. You can vibe-code an app over the weekend. Launching a new SaaS product isn’t that much harder than launching a memecoin.
And yes, fine, some people will say that startups aren’t like memecoins at all, because memecoins “have no business model, cash flow or fundamental value,” while startups are supposed to be rigorous commercial enterprises that make money and serve customers and all of that, but are they? Are you sure?
At the AI startup that promised to help people "cheat on everything," there are only two job titles: engineer or influencer.
"There are only two roles here. You're either building the product or you're making the product go viral," Chungin "Roy" Lee, the CEO and cofounder of Cluely, said in an episode of the "Sourcery" podcast published Saturday. "There's nobody who's not a great engineer who has less than 100,000 followers."
Though I sympathize with the people who criticize all of Cluely’s shenanigans, I think they somewhat miss the point. Cluely isn’t a technology company; it’s a shock jock podcast that manufactures outrage. It’s a brand that markets vibes. It’s a meme-company, created to sell a meme while the meme is still popular. The company is logo to put on the business card; the product is a mechanic for selling the meme. And measured against that metric—a job well done.7
Art App Content
Actually, we can probably go a step further than that. Because here’s a loose theory about technology, social media, and digital creation:
Before we all had computers and phones and Instagram, making art8 was hard. You had to have a fancy camera, or painting skills, or the ability to stitch together film strips into a video. Because art was expensive and somewhat scarce, we valued the art itself.
Then it became easy to make. You can create great art in seconds, sometimes without even meaning to. And as the cost of making it fell, the value and notoriety of each individual piece of art fell too.
So we started to care more about the creators than their specific creations. Like: Name that one great Kai Cenat stream. What’s your favorite Mr. Beast video? What’s Charli D'Amelio’s masterpiece? Some things might be more memorable than others, but there is no opus. Very little stands on its own. Popularity comes from a personality and an amorphous body of work.
Now, the cost of creating software is also going to zero, as they say. So would we not expect to see the same patterns here? While that doesn’t mean big software businesses will go away—there will always be workhorse products that do accounting and manage warehouses and fly airplanes, just as there are still big-budget Hollywood movies—could there not also be an ecosystem of influencers who make software that is popular because they made it?
I don't have a particularly precise idea how this might happen, and it seems unlikely that influencer engineers would be exactly analogous to more traditional social media influencers. But: Are Nikita Bier’s apps products or content? Is he an entrepreneur or an influencer? Is signull, an anonymous tech commentator, creating a product studio or a hype house? Is there even a difference?
This is another way to describe Cluely: It’s simply Lee’s first piece of viral content. No matter what happens to the business next,9 it’s already worked, because it’s made Lee a famous entrepreneur—and in Silicon Valley, there’s no brand more valuable than that.
Emphasis and lower-case letters are his.
For instance, there are a handful of data companies that burned enormous amounts of money on marketing before their products could live up to their ads’ promise. That was the explicit strategy though: Buy mindshare now, and hope the product can eventually catch up. A lot of AI companies appear to be doing the same thing, except, as Ethan says, they can outsource a lot of product improvements to foundational models. They don’t need to build new stuff to catch up; they just need GPT-5 to come out.
Marketing matters for AI products for another reason that is too often ignored, which is that most AI products are very subjective. There aren’t straightforward ways to compare different chatbots or coding agents against each other. Some people solve this with bizarre and fun tests; most of us solve it by buying based on brand and reputation.
Put differently, taste is memetic, as the menswear guy frequently reminds us:
I contend that beauty in personal appearance is subjective, not objective. In fact, its standards rest on the shifting tectonic plates of politics, economics, and technology.
…
The link between power and aesthetics becomes especially clear when you consider how certain styles became dominant in the first place. Either the ruling class had an innate eye for beauty, or we find things beautiful because they’ve long been associated with power and privilege.
Of course, Pierre Bourdieu noted this in his 1979 book Distinction, where he observed that our notions of "good taste" are nothing more than the habits and preferences of the ruling class. Georg Simmel said something similar in 1902 when he said fashion is a game of imitation.
Technology, and especially technology as subjective as generative AI, is no different. We like what we’re told to like by the people we like.
If you don’t know what this is, I’m definitely not explaining it here.
If you don’t know what this is, I’m definitely not explaining it here.
Of course, why did a16z buy the meme? One, because they believe Cluely can monetize the meme, and two, because VC is a vibes-based business too. If a generation of startups is going to be built on this model, what better way to prove that you’re the ideal partner to them than by investing in the strategy’s ultimate archetype?
Whatever’s next is apparently happening in four hours. I have no idea what they’ll say, but I assume it’ll make everything I said here catastrophically wrong.
everything you write is like the lid on a rabbit hole for me. and once i open for a peek, i'm always digging out gems––new ideas to explore, new people to follow. thanks for that.
I think Cluely is making a bet that controversial = attention = scaling. Meaning that in a world where every GPT wrapper and AI agent looks the same, the one that is most popular, gains user traction, and scales, wins. They grabbed people's attention better than any other GPT wrapper.
It's an interesting way to build a moat in a world of short attention spans. Algorithms promote the most controversial things, and therefore, they promote their startup. “We’re earlier than the latest YC batch of companies. Yet, we’re generating more views than every single one of them.”