Almost Good

On AI, authenticity, and the new economics of trust.

My book, The Girls Are Not Fine, is now available for preorders! Get your copy here.

Most things you consume online right now are almost good.

Almost good is not good. It’s not bad, either. Almost good almost clears the bar. Almost good almost gets it.

You know it. You recognise it. “It’s not that we have too much information, it’s that we’ve lost the ability to sit with uncertainty.” “It’s not burnout, it’s a crisis of meaning.” “It’s not about productivity, it’s about presence.” The structure that looks like an insight because it has the shape of one. The setup, the reversal, the landing, the punch, the aha. But when you press on it, there’s nothing underneath. No remainder. No specific mind that had to arrive at it the hard way. It could have been written by anyone. It was, increasingly, written by no one.

This is what almost good actually is: the shape of thinking without the thinking. The outline of a feeling without the feeling inside it.

I’ve been thinking about why, and I think it is this: it is too resolved. It has no remainder. A piece of real creation always has something in it that the maker didn’t fully understand when they made it. A question they were working out by writing it. A feeling they couldn’t name that ended up in the syntax. Real work has a shadow. AI work doesn’t. It is all surface, lit evenly from every angle, and that is what makes it feel like a vessel rather than a thing.

And yet we keep calling it good. Because we’ve been trained our whole lives to treat polish as a proxy for effort. A finished thing meant someone had done the work to get there. That’s not true anymore. AI is genuinely better at completion than it is at the work. It can take something half-formed and make it whole with a fluency that most human writers spend years trying to develop. Which means we’ve done something weird: we’ve made the hard thing easy and the easy thing hard. People today are more able to create a polished product than an unpolished one. We have inverted the difficulty curve.

For this essay, I define grotesque as not just unfinished but monstrous in its unfinishedness. We are not just talking about something in process; we are talking about something in process, and its incompleteness attacks the viewer in one way or another. This creates an aesthetic of provocation. The idea is not just incomplete; it is almost offensive in its incompleteness. It feels like a titillation, as if I shouldn’t be watching this at all. I shouldn’t be watching this uncompleteness at all, but I get to. It feeds into a larger voyeuristic want of a user who is fatigued by the polished and complete.

Now, if we keep this definition of the grotesque at the centre of this argument, we ask ourselves: can you manufacture the grotesque?

Theoretically, you can. AI can create rough drafts for you. AI can create the flubs for you. AI can create all the things that would imply the process. Imperfection can be generated as part of the process. I’d argue that as you generate things on AI, the first drafts are often quite horrible, but are they grotesque, as defined by this essay? I don’t think so. The problem with the grotesque and the titillation and the voyeurism that the grotesque creates is that it has a very ineffable quality. There is something deeply human about it, which means that you either recognise it or you don’t, and you either project it or you don’t. It’s a binary. There is no halfway house between the grotesque and the not-grotesque.

That creates a class of creators who are ineffably human in their quality, but it also creates an entire class of creators whose opacity makes us deeply suspicious of their outcome generation.

I like to look at Brandon Sanderson, live-streaming himself, figuring out a plot problem in real time, getting it wrong, reversing, getting it wrong again. The audience is watching him think rather than watching him perform having thought. Another artist who defines what the grotesque looks like in practice is Rick Rubin. Rick Rubin doesn’t know why something works until it works. He says so. He is so honest about his own lack of certainty that the absolute lack of artifice becomes the marker of credibility. These people understand, whether consciously or not, that the finished thing carries a lesser weight of proof. The proof has to live somewhere, so now the proof has moved upstream into the mess.

Compare them to creators who are almost hermetically opaque, who only post finished work and describe their output in mystified terms, like it came to them in a dream. A few years ago, that read as artistic, almost spiritual. Right now, it reads as deeply suspicious. Not because opacity is new. Artists have guarded their process fiercely for years. But because the stakes of that opacity have changed.

I now question opacity, not because of opacity unto itself. There are artists who are fundamentally opaque, and that opacity is the context of their art (take Banksy). I question the opacity of artists who have the capacity to show us the background of their work and choose not to, because then I wonder: what are they hiding? Is it just a process, as I would have assumed ten years ago, or is it more?

I started creating on the internet when I was 14 years old. I don’t recommend it, but it means that there is almost 16 years of what I think of as digital provenance.

There are people who have followed me for ten years, some for more. They have come in and out of my life. They have followed me, unfollowed me, followed me back again. That’s perfectly fine, because they have seen my process evolve. That is accidental trust. I never intended it to be a proof of existence, but it is now a proxy for faith in my process because they have seen me build it in real time around them. They have seen me be wrong. They have seen me be embarrassingly confident about things I later had to walk back. They have seen the bad writing and the better writing and the bad writing again. That history is a timestamp. Proof of prior existence. It functions now as provenance in a way I never planned for and could not have manufactured.

My provenance is embarrassing. It is a 16-year-old writing poetry on Ask.fm that she thought was profound. It is the Facebook statuses I cannot delete, and I would not delete them even if I could. It is the first essays that were trying too hard, the opinions held with too much confidence, the phases, notably, the activist phase, the startup phase, the I-have-figured-out-investing phase. All of it is somewhere on the internet, indexed, searchable, occasionally screenshot and sent to me with a laughing emoji. That archive is not curated. I did not build it as a trust strategy. It is just the residue of a person figuring herself out in public over a very long time. And that residue is now, weirdly, one of the most valuable things I have. Not because it is good. Because it is real, and because nobody would bother to fake sixteen years of being publicly mediocre on their way to being occasionally good.

Now I look at someone who starts creating today, with no 15 years of provenance behind them, and I genuinely don’t know what happens there. Can trust even be built in this era? Is it possible to create the same faith that somebody who has been creating for 30 years has earned? I honestly don’t know. Because trust takes time, and time is now a deeply complicated variable. AI has compressed the time to output. So the new creator is stuck in a genuinely strange question: Do I put out output fast enough to stay visible? Do I go slower and hope that slowness reads as intentionality? Does being slower elicit trust, or does it just mean fewer people find you?

Access was democratised. It is easier than ever to arrive at output. It is easier than ever to arrive at a finished product. But the act of building trust remains the mainstay of very few. It’s either people who can show their process compellingly enough to compress the timeline, or people who have simply been around so long that their provenance does the work.

The person this hurts most specifically is not a vague “new creator.” It is the 22-year-old from a small town who is genuinely talented, who has something real to say, who did not have the access or the safety or the platform to figure herself out publicly when she was younger. She arrives now with better tools than I had at 14, honestly. But no provenance. And she is trying to build trust in an environment that has seen too much almost good to extend it easily. The barrier to entry for creative credibility is higher now than it has ever been, precisely because the barrier to entry for creative output has never been lower. That gap, between how easy it is to make something and how hard it is to be believed, is where a whole generation of creators is currently stuck.

There is a concept I harp about called founder lifetime value. It doesn’t ‘exist’, per se, I just made it up.

How many times can you make a mistake, recover, and keep it from becoming the permanent story about you? The answer, almost always, is a function of runway. Financial runway, yes, but also social runway. The founder who went to IIT and whose father knows three partners at the fund gets more attempts. More iterations before one bad call defines him. The founder who scraped in from nowhere gets fewer. Sometimes just one.

I think the same is true for creators. The grotesque requires runway. It requires the security of knowing that one bad piece of work will not become the headline. That the audience will hold it as a data point in a larger story about you, not as the story itself. And that security is not evenly distributed.

A woman with an established platform posting a shitty first draft reads as vulnerable and humanising. The same shitty first draft from a creator without that platform or provenance reads as not ready. The grotesque is a luxury. Transparency is a class position.

Nothing made this clearer recently than what happened to Pujarini Pradhan, who creates online as lifeofpujaa. She is a woman from a small village in East Midnapore, West Bengal, who builds her following by discussing books, films, and feminism in English, from her home, in a cotton saree, no metropolitan apartment in the background. Nearly seven lakh followers. Brand deals with Netflix and Audible. And then, almost on cue, the accusations: industry plant. Fake. Too polished for someone like her.

Too polished. For someone like her.

The argument, such as it was, collapsed quickly. She had an agency for brand deals, which is completely standard, and she shot and edited her own content. But the accusation itself reveals the specific trap that a creator from a background with no history of being taken seriously faces. She was suspicious precisely because she succeeded. Her consistency read as evidence of a hidden hand. Her brand deals read as incongruous with her setting. Her fluent, accented English read as something that needed to be explained. When she said “they were fine with me until I started earning money” she is right. The mess was fine. The success was the problem.

This is the impossible position: she could not afford to be grotesque, because grotesque requires the audience to already believe in you enough to watch you recover. And the audience had decided before she started that someone from where she came from did not have that kind of credit with them. She had no runway. So every stumble was a verdict, and every success was suspicious.

This is who the democratisation paradox actually hurts. It is the creator who comes from a background where being taken seriously was never the default. Where the margin for error was always thinner. Where one bad take, one inconsistency gets read not as process but as proof of what the audience already suspected. You cannot show your mess when your mess is all the evidence some people need to dismiss you.

So when I say that visible process is the new credibility, I also mean that it is so much easier to say if you are someone who already had the credibility to begin with.

Let’s talk about the word authenticity. We accuse people of not being authentic. We say authenticity is what draws viewers in. But is authenticity not also a form of curation? It is curation done specifically with the intent of projecting an image of relatability. The authentic self you perform online is still a performed self. The difference is just in how well you’ve hidden the performance.

When I think about what authenticity actually looks like in practice, one of the most interesting cases to me is Becca Bloom. She is openly a billionaire. She does not pretend otherwise. And you would think that would be a problem, because your average billionaire is not particularly well-liked, and there is nothing organic about a billionaire. But people love Becca Bloom. They welcome her. They engage with her. Because she puts up a performance of authenticity so convincing that the billionaire part becomes almost incidental. No billionaire is going out of their way to do brand deals. She is. And we don’t question it too much, because she comes across as amiable, accessible, and easy to consume. That is authenticity as technique.

And this is where the real problem is. Because if authenticity is curation, and curation can be learned, then authenticity can be faked. The manufactured mess. The curated unfinishedness. The aesthetic of someone who is figuring it out. We are going to get very good at this, and audiences will sense it, and the cycle will tighten until even rawness requires verification.

There is also the trap inside the trap. The minute you know your rough draft is public, it isn’t really a rough draft anymore. The musician posting her flubbed take made an editorial decision about which flub to post. The writer sharing her cutting-room floor curated her cutting-room floor. I am writing this essay knowing it will be read, and that knowledge is shaping it in ways I cannot fully account for. The second you introduce an audience, you introduce a performing self. And what you’re showing is not your process but your process’s public-facing representative, which is a different thing.

I have been thinking about this personally. When I was younger, my parents had a rule: if it is not something you could discuss in the living room with your family, don’t post it online. I have more or less kept to that for 16 years. Does that make me inauthentic? I don’t know. I edit and curate what I put out in the world. As I’ve gotten older, I’ve become more suspicious of who consumes me online. I’m fairly convinced that among the people who do, there is a large, vocal part that consumes me out of jealousy or envy. It’s not the kind of audience I would choose if I could choose. But I can’t curate my audience. So I control the variable I can, which is me. Is that authenticity? Is it strategy? I genuinely don’t know where one ends and the other begins, and I’m not sure the distinction matters as much as we pretend it does.

So what is actually left? If the finished product is gameable, and the process is gameable, and authenticity itself is gameable, what isn’t?

I think it is irresolution. But a very specific kind of irresolution. Performed irresolution always ends with a lesson. It says “I don’t know” and then three sentences later it knows. It cashes out into a take because that is what the algorithm rewards and what the audience came for. Genuine irresolution ends without knowing. It doesn’t resolve because there isn’t a resolution yet. AI is structurally incapable of genuine irresolution because it always knows where it is going, even when it is pretending not to.

Genuine irresolution also has stakes. When you don’t know something and that not-knowing has actual consequences for your life, it reads differently than not-knowing as aesthetic. The person who is genuinely uncertain about whether to leave a marriage sounds different from the person performing uncertainty about their “journey.” You can feel the difference even if you can’t name it.

And genuine irresolution is particular. It is not “I don’t know what success means anymore” that is performed. It is “I don’t know whether the version of me that finishes this book is someone I am going to like” that is specific. The more universal the uncertainty, the more suspicious I am of it. Particularity is the tell. The detail that is too strange and specific to have been generated. The thing that could only have come from one mind in one life at one specific moment.

Which is maybe what we are actually looking for, underneath all of it. Not authenticity. Not process. Not the grotesque. Just evidence of a located mind. A particular mind with a history and stakes in what it is making. Not trying to be all minds or the most optimised mind. Just this one, here, now, working something out, and genuinely unsure how it ends.

I could not have written this book if I hadn’t spent 16 years creating a map to myself in public. I earned an audience, and now I am leveraging that audience to sell a book, and I am aware of exactly how that sentence sounds. I am also aware that there are people who would call me an influencer for it, and there are times I want to say, “But I’m not an influencer, I’ve been writing seriously for years, there’s a difference.” And then I stop, because does that distinction actually matter? Or is it just me doing the thing this whole essay is about? Curating which version of my credibility you’re allowed to see.

What I know is this: I don’t think I could have written this book at 22. Not because I wasn’t capable of the thinking, but because I hadn’t yet accumulated enough public becoming for anyone to trust that the thinking was mine. The book exists because the provenance existed first.

And that is what this essay is actually about. Not me. But the question of what it means to be a creator in a world where you are incentivized to bleed not just for the art but for the distribution. It is already so hard. There is no money in it. There is so much luck involved, ugly and arbitrary luck. What does it mean to be a creator in an era where even that luck has to be manufactured by a machine, and if you are not good enough at operating the machine, the machine will penalise you for it?

I do not have a lesson here. What I have instead is a harder question than the ones we usually ask about authenticity, process, or AI. It is not just what feels real anymore. It is who gets to look unfinished in public and survive it.

Because that privilege is not evenly distributed. Some people get to show the mess and have it read as depth. Some get to fail in public and have it folded into a larger story about talent. Some get to revise themselves in front of an audience that has already decided they are worth waiting for. Other people get one bad draft, one awkward video, one slightly off note, one stupid statement online, and it becomes proof that they were never serious to begin with.

This is not about whether AI can imitate the rough draft. It can. Not whether authenticity can be performed. It can. But whether, in a culture this suspicious and this saturated, a new creator can still build the kind of trust that used to emerge slowly and accidentally, over time. Whether a person without provenance can still become more than almost good.

Maybe that is what we are actually defending when we talk about art, or writing, or voice, or process. Not purity. Not even originality. Just the right to become in public without being destroyed by it.