For people in a tiny slice of the world and internet, there’s no greater canon event than the prospect of a new social app taking off. Sometimes there are multi-year droughts. In a plentiful year, there are one or two prospects that excite us. (And I wrote a stream of consciousness musing about the usual psychological journey.)
Arguably the first meaningful new prospect of this year is Airchat (version 2.0). It launched to the public 5 days ago, and it’s taken over at least the tech twitter hype train. I joined it a few hours after launch with a few hundred other people.
And since then, my mind has formed thoughts, some that I’ve decided to type out here. But this isn’t a comprehensive analysis or review or profile or whatever else will be written in spades in coming weeks. (And disclaimer, I’m not about to make comparisons to Clubhouse here.)
But if you’re another one of the few people amused by this little corner of the world, you might enjoy my off-the-cuff notes — on product, format, creators vs. consumers, use cases, comparables, culture, and what may come next:
Version 1 of Airchat was too complicated. Now that there’s a generally-agreed better version out, people seem to be comfortable saying this. Last fall, in my essay The Brunch Test, I wrote out my concerns about the product:
It’s feature rich — async and sync, text, audio, visual, stories format, feed, profiles — but possibly too rich. What is the core action that a new user should take? What is the main use case? What time, place, setting is this app for? I haven't quite figured out the answers yet (though I appreciate a lot about the implementation). With quasi super-apps like this, I find myself thinking of how you could chip away at the big block of marble to reveal a simpler core.
Version 1 was a super-format-app treating every medium as primary. Version 2 is a lot simpler. But, I think it’ll need to be even simpler still.
Version 2 is strong on product fidelity and integration of novel tech. These are the top 3 things I found impressive from a features / UX POV:
speech-to-text transcription / cleaning up voice / sped up / ?moderation
feed auto-plays voice and auto-advances (and you can optionally pause it)
floating mini-video option along with voice (and good that it stays small)
(I’m going to leave the UI / UX bugs, nitpicks, and requests for users to opine on.)
If the “medium” is the message, the message is still undefined. One of the first things that strikes you is hearing real voices, sometimes of people you’ve known online for a while — so there’s the power of voice. But it’s hard to define the medium, maybe because it’s novel or because it’s still a bit too complicated. Is it a “text” app or an “audio” app? (And video is still tucked in there too.) I’m inclined to see it as text+ (and I’ll explain why).
The format feels different for creation versus consumption. We’re historically used to the medium being the same for both sides. In Airchat 2.0, you have to create by recording either voice alone or voice + video, which is then transcribed to text and posted (no edits). This is a differentiating constraint and helps maintain a feeling of unscripted authenticity. But the primary consumption method still seems to be text, i.e. reading. Read on for why …
It’s nice that the consumer has the option to read or listen (or watch in some cases). This makes it more versatile than just being text-only or audio-only. E.g. read when you’re in the subway but listen when you’re alone. That said, it still feels like text-first consumption with voice narration — because people read a lot faster than they talk (and they can scan and skip ahead too). Playback seems fixed at 2x speed, which helps but is still slower than reading.
There is a good amount of friction for creation. If we’re comparing to text-only apps (i.e. Twitter), it’s still higher friction to create despite the lower burden of not having to type. I felt all 3 of the following hesitations when I thought about what to post or when to engage:
most people are less comfortable speaking “extemporaneously” (using this word cus it’s the right word, not to spite Paul Graham) and with no editing
people kinda hate hearing the sound of their own voice back, and
thinking of what to say is hard and there’s no precedent or template for it. Twitter and IG and TikTok have a lot of mimetic behavior embedded, but it’ll be harder to figure out here.
Who will “do well” on Airchat? Well, the higher barrier to “creating,” means that it’ll have to be people who are more comfortable putting themselves out there and owning their statements. If Twitter has a 90/9/1 ratio, then this will be even more skewed (i.e. even less than “1” who will want to initiate chats). The people who will will fall into two categories:
people who do well on Twitter and have more personality than the average tweeter and aren’t afraid to show it (e.g. people who are more socially comfortable and confident and post stream of consciousness thoughts, feelings, personal stories, and/or post video or photos of themselves)
people who are generally comfortable speaking in soliloquy in public even in real life, tend to be opinionated, not afraid to be associated with their own voice and strong opinions (e.g. loud political pundits or academics)
Who will want to do well on Airchat, and why?
It’s people who haven’t already built big audiences on other platforms or are looking for their next 10X unlock who benefit the most from engaging early on a new platform. And the perennial other category: tech people looking to just “network” or invest in the product itself. Spending time on Airchat right now likely will increase your surface area of luck. For selfish reasons, Naval (and others) will be a lot more responsive to you on Airchat than on Twitter.
I recently wrote about this very real phenomenon that I call early-arriver arbitrage.
The potential to find and seize opportunities in new arenas is what I call “early-arriver arbitrage.” And every new arena has unique arbitrage opportunities.
There are many kinds of potential arbitrage: knowledge arbitrage, social network arbitrage, technological arbitrage, cultural or geographical arbitrage, to name a few. And overlaying all of this is simply time arbitrage.
It’s “arbitrage” because there are temporary inefficiencies in new arenas (or systems, situations, contexts — whatever you call them). Most people haven’t spotted the value in the arena yet, so those who have get an advantage.
And since the arena is new and sparsely populated, it’s easier to scale up to be the “expert.” The classic advice that you need to put 10,000 hours into something to master it can all but be thrown out the window in new arenas.
The good thing for both of these groups is that there are plenty of “high power” people on the app at this stage largely because a) people are curious about the potential next big thing b) the founders are very well-connected, with high-powered networks and distribution.
Still, only a few of these people will contribute to content this early. We’re in a cultural phase of people feeling overcommitted to the apps they’re already time-invested in, and most people want to wait and see.
How would this format, if successful, fit into the broader media landscape? It won’t replace podcasts or text-based messaging or platforms. I’ve written about my belief in companionship content, which is certain kinds of long-form passive content. In contrast, short-form video requires more energy:
What is the cause of this excess energy consumption? Short-form video requires more attention & action in a few ways:
Context switching, i.e. wrapping your head around a new piece of context every 30 seconds, especially if they’re on unrelated topics with different styles
Judgment & decision-making, i.e. contemplating whether to keep watching or swipe to the next video effectively the entire time you’re watching a video
Multi-sensory attention, i.e. default full-screen and requires visual and audio focus, especially since videos are so short that you can easily lose context
Interactive components, e.g. liking, saving, bookmarking, commenting
Airchat is public, short-form, async audio, which I think is hard to make truly passive (even though the auto-play feature does help a lot). It still has the first two challenges of context-switching and decision-making, which I believe places comparatively more energy and attention-demand on people. That is, as is, I don’t believe it can’t be something used hours a day indefinitely. But it can provide a higher level of authenticity for conversations that do take place.
Let’s talk about the network graph (nerd topics ftw). It’s an asymmetric follow graph (i.e. I can follow you and you don’t have to follow me back). The way I see it, a follow graph is first and foremost an interest graph and influencer graph until proven otherwise. By influencer graph, I mean that it encourages creators. As long as the primary use case is public and I have “followers” listed on my profile, I’m likely to try to maximize that number. The tradeoff for limited privacy is the opportunity for public exposure and growth.
According to the founders, Airchat is a social messenger or a conversational network (versus the usual “social media” label). I see this aspiration for authentic conversations and think it’s noble, though it’ll be hard to maintain for the reasons I mentioned re: the network graph. The Twitter comparisons are warranted imo. This is a public conversation app that’s small enough right now to remain communal and conversational — this will become harder as it scales.
It’s very hard to build an open network that doesn’t default to broadcast behavior — the only hope is impeccable execution in creating semi-permeable communities. You try to do this with a) the algorithm b) walls within the app. The asynchronous nature of the content will definitely help with the algorithm’s efficacy. And re: walls, a good example is subreddits or, as they’ve alluded to, “channels” that can be set to private or public). (I joined the “WAR” channel that was recently created and it mostly acts as a topic filter for the main feed so far, so this is very nascent, but an interesting direction.)
There’s a high density of smart and connected people on the app right now. This is a style of seeding done well (and differentially available to more experienced, insider founders). On a conversational platform, the quality of the early users is the major factor in every user’s experience. It’s like getting access to the VIP section of a nightclub but for tech people. It’s hard to separate the product from the people who are on it right now — this is good for Airchat the company, but they’ll need to prove that the product can maintain the quality of experience as the affinity of the network dilutes.
It’s giving “friendly stranger” vibes and people are feeling the momentum. My overall visceral feeling: it feels a bit like well-meaning conversational theater. I say theater because there’s still something a little unnatural about a start-and-stop, asynchronous conversation in public. Even private voice notes are a little disjointed from the texts that tend to come before and after them, but given they’re private, they at least tend to have a more casual and ranting quality (in the best way). But when most people are asked to do voice notes in public, they’re giving guarded mini performances. But referencing note #7 above, there are a few people who sound authentic.
As of today, Sunday the 14th, the founders decided to halt open invites given the fast growth of the app. And this makes sense. It’s a good move because I don’t think the app with its current network structure will scale well. I also don’t think there’s a good understanding yet of what use cases this is really best for and how it’s meaningfully different from Twitter for some of those use cases (despite the aspirational intent to be different, again, which I get). And hype is dangerous in that it can make it hard to see what’s real; the best counter is time and honesty.
Like everyone else who cares about this niche world of social apps, I think it’s exciting to watch a new product get some traction and try to evolve in real time. We tend to see commentary about high-profile apps fall into two camps — either obsequiously positive for the sake of social gain or hyper critical for the sake of sensationalist reporting or rage-bait. I don’t have any agenda here except to share some authentic reflections before the discourse gets even more noisy.
As I said on Twitter, it’s unbelievably hard to simulate real life conversation in any digital medium, yet perhaps one of the most noble pursuits. If you know me, you know I’ve spent some time around this space before. And in part because of that, I don’t want to make this about predictions or recommendations (though I’ve woven some in between the lines) — because there are so many paths still open for the team to take. What I know for sure is that some paths hide traps and others treasure. And the fun part will be seeing which they happen to choose.
#6 in full swing :) but curious what you’ll think after a day or two
All right, you convinced me to start testing it. I'm on it already but haven't recorded anything yet... 👀