‘Letterboxd for lattes and sandwiches’: the new social media platform putting human curation before the algorithm

Perfectly Imperfect is a new app designed to fight against AI-slop, attention traps and echo chambers.

It feels like 2025 has been quite a defining year for social media, so far. With our apps becoming increasingly infested by AI-slop, sinister targeted advertisements and short-form content, more and more people have become disillusioned with the platforms they once sought community and entertainment from. And yet, it is these very features that make it so hard for us to detach from them, each one psychologically designed to entrap us.

But the internet hasn’t always been like this. In their early days, platforms such as Instagram, Tumblr and Facebook were genuinely places for discovery, communication and taste-making. They felt like they were truly bringing humans together, as opposed to driving us apart. So what has changed?

Social commentator Corey Doctorow calls it the ‘enshitifcation’ of the internet; a process where online platforms initially attract customers with an authentically user-centric and beneficial service, lock those customers in by essentially getting rid of all competition, and then finally degrade their service by prioritising profit and shareholders over user experience. Well, it certainly feels like we have reached phase three.

But it looks like there might be an alternative. Launched last year, Perfectly Imperfect (PI) is fast gaining a name for itself as the new social media platform fighting against brain rot, attention traps and echo chambers. Instead, it harks back to an era when corporate algorithms didn’t control cultural discovery and content was genuinely created for and by the user base. Or, in the app’s own words, PI is “the anti-slop taste network”, where “human curation reigns supreme”.

PI started its journey as a newsletter during the 2020 lockdowns. The premise? Asking people one simple question: “What are you into right now?” Each contributor provides recommendations that are collated into the publication, a formula that has now been transferred onto the platform, whereby each user curates their own page (deliberately clunky and old-school in design) with pointers, tips and opinions.

Having made an account myself, I was struck by the simplicity of its design. Navigating PI feels like being on an early version of the internet – at least what I imagine that to be like. It’s basic, not particularly aesthetic and 2d – there is no moving image whatsoever.

At the bottom of my screen, I am confronted with three options: I can post a ‘REC’ (something I am reading, eating, listening to, etc.), an ‘ASK’ (if I would like to know what other people are doing, thinking, consuming, etc.), or an ‘EVENT’ (clubs nights, tours, poetry readings, exhibitions, etc.) I can also respond to other people’s ‘ASKS’ (two I got today included: ‘What’s a DJ set, mix or playlist you’d recommend?’ and ‘What’s your favourite game from childhood?’), read full-length editorial pieces (like Substack), and strike up conversations with other users (like Reddit). It is worth noting that both Substack and Reddit have gained popularity in 2025, particularly amongst young people searching for something ‘deeper’ and more ‘meaningful’ than the standard social media options.

So far, the app has been getting a lot of love from the niche bubbles that are using it (largely based in NYC), with one user saying: “It’s so perfect and brilliant because recommendations are the primary thing I want from the internet. Like I’m always Googling for restaurant recs, or the best barefoot shoes or whatever, but have such a low level of trust in the results because of how gamified SEO is and all the sponcon and affiliate bullshit. In contrast, I have an extremly [sic] high level of trust in the user base on here intrinsically because I love PI and trust that other folks who follow already share a lot of my interests. I’m much more likely to try something I see on here even if it’s outside of my normal taste.”

And it’s taste that is the key point of interest here. In a world drowning in content, where our attention is constantly being pulled in every direction, what you choose to give your attention to – what you consume, what you engage with, what you amplify – has become more important than ever before. “We used to associate intelligence with accumulation,” says Stephanie Tyler on Substack. “The smartest people were the ones who knew the most. But that model doesn’t hold anymore. AI knows more than anyone. Wikipedia is free. The internet has flattened information access so thoroughly that hoarding knowledge is no longer impressive. What matters now is what you do with it. How you filter it. How you recognise signal in the noise. Curation is the new IQ test.”

It is from this same realm of thinking that PI has sprung. The idea being that, when you are stuck in these data-ridden echo chambers, it’s hard to know if our tastes are actually that: our tastes. Or, if they are being shaped by what we are pumped on the internet. Of course, taste has always been shaped by an amalgamation of sources and inspirations, but these have previously been human-centred, dictated by what a critic or musician likes, and not by a set of data points.

PI is reversing that process by putting human discovery and human curation front and centre again. Here, there are no CEOs trying to make money off what you get to see and what you don’t, no data being collected on how much time you spend looking at something. Instead, recommendation networks are being built between strangers.

But, unlike the old days when curation was so often placed in the hands of a few, PI intends to ‘democratise taste-making’, as its co-founder, Tyler Bainbridge, told Dazed, allowing anyone to become the decider of what is cool and what is not. This is essential in a climate where elitism and exclusivity value certain opinions over others, and by consequence prevent those others from being heard.

There is, however, a question as to whether “taste-making” should be “democratised”. Whilst elitism and exclusivity are definitely wrong, not everyone has the cultural authority to speak on everything. In a recent article on Substack, the writer Dani Offline poses some interesting questions about our collective desire to be the maker, the creator, and by consequence our devaluation of the person who participates in, or consumes this content – the community. Perhaps not everyone needs to be a ‘taste-maker’, for if everyone is a ‘taste-maker’ then whose tastes are they making?

For the time being, PI remains a sub-cultural anomaly with a small cult-like following. The challenge will come when the app widens its audience, which it inevitably will and hopefully does (I would personally love to see some other recommendations other than ‘tahini’, ‘green scents’ and indie bands). Then, are we not opening the door for people with vastly different opinions and tastes from your own, some of them surely unsavoury? If we don’t want to see these, are we not just re-entering the same echo chambers we so desperately sought to escape on the other apps?

Perhaps I’m getting too far ahead of myself. The path of PI is yet to be decided. It is still very much in its early days. For the time being, the app makes money via a tiered system, by which some users can pay for the ability to customise their profile, as well as access extra newsletter content and archival interviews. It also participates in brand partnerships, which are limited to the newsletter. How long it can and will last like that is yet to be seen. Substack felt different until it incorporated a space for short-form content. Either way, PI feels like a positive and necessary step in the right direction, and certainly reveals something important about people’s desire for an AI-free, community-driven form of social media – or social media as it used to be.