Khan Academification Pipeline as a prolific neo-media company
You get a team of killer video editors and producers and you collectively put in work to hone them in, stylistically, on a video presentation style that you’ve determined to be effect (based on existing standout educational videomakers). At the center of this team must be someone with excellent taste, self-confidence, and a deeply unhumble appreciation for their own convictions. This is a Steve Jobs company.
So you get the team, right? Now you run through the process a bunch of times internally, and you watch it like gametape until you’re all game-ready. Do not just publish it.
Meanwhile, someone on the team is on a mission to track down high-touch domains (of which there are infinite) — particularly ones ripe for Video-Educational Disruption. Of those domains, you find people particularly qualified to write really high-quality educational scripts (not just accounting for educational value, but crucially also storytelling ability, pacing, presentation, effective use of metaphor, effective use of state-prompting language before key points, … the list goes on — there is a huge amount of variables that contribute to the effectiveness of a message, and with this in mind, the bar is quite low for a team that understands and executes on this.)
So you track these people down — maybe they write bangin’ blog posts, maybe they are already a youtuber with a subpar video production pipeline, maybe they are a podcaster who just doesn’t yet have an Educational Video arm built into their operating system. You reach out, you explain the deal, you work out proper partnership agreements, and you collaborate. They write the script. They work closely with the video production team; if it made sense for them, they could optionally become a co-producer of that video/series.
The video team serves as the transformation function ( i.e. the f() ), the bloggers who can write highly effective scripts for the videos serve as the input, and the best-in-class videos you produce are incredibly valuable, in both their own rights and in their contributions to the company’s overarching brand value.
It’s kind of like a podcast, where the guest is the star and they’re here to drop some knowledge on you, except there’s no podcast host and you’re able to tastefully enhance the quality of their message.
Eventually, you become a prolific and educational media production company. Neomedia.
Radar
World of Warcraft Group Finder x Yik Yak
RingMe (domain: ring.me)
It’s a rainbow-colored Oura ring with a little bluetooth button on the inside and the ability to vibrate.
- You wear it to signal you’re single and open to serendipitous meet cutes.
- If you both tap your buttons, the ring vibrates! Physical stimulus; a shared wave of adrenaline.
- In the onboarding, somewhere, clearly: “if you want to chat right there, put your hand up! like, “yo, this thing is vibrating! yours too? crazy! so, …” – doesn’t have to be like this, but should encourage some “default” “protocol” to reduce the initial friction.
- As a backup option, i.e. if one person doesn’t put their hand up right there, we connect the two of them and open up their contact info to each other on the phone app
- The phone app is completely free. Do not monetize on the app. This is a tech-enabled fashion/culture company, not a subscription service. You sell an identity: when you wear this ring, you are a quirky, fun-loving person embracing serendipitous connection.
- It’s not a dating app, it’s an enhancement for the real-world dating experience. A catalyst for serendipity.
- Enjoy Dating — banger tagline.
- Tell people to wear it on their right middle-finger? Maybe biased, that’s my favorite ring finger, but suggesting a specific finger is probably good.
- Real-world dating, so hot right now. Given the choice, 99% of people would prefer to meet in real life. This product rekindles real-life meet cute stories. So hard and so uncommon to recreate that early excitement when you begin on a dating app. The real-life adrenaline sets the tone for your first interaction; you are more authentically able to communicate your vibes. We are so complex, and text is so lossy.
tech
- Bluetooth ring connects to your phone; your phone has an app. If you tap, and they tap: you match, the ring vibrates, the phone app registers them in your contacts.
- Vicinity-based; small possibility of a mix-up situation. The possibility is good for marketing; skits about mix-ups would make for funny TikTokky 2024 marketing content.
AR/VR trend is a medium-term wave that a well-positioned company (this) could ride after getting a foothold on branding
Make 100 or 2,000 of them and give them away for free in a health-focused daytime setting frequented by a disproportionate number of single people not explicitly for the purpose of dating: yoga classes?, gyms?, coffee shops?
like this, but better
Love.ly
The dating app market is currently controlled by one giant company that owns Hinge, Tinder, OkCupid, … the list goes on. Bumble is the exception, but the app is basically the same as Tinder; there is no differentiation in the incentive structure or business model.
philosophy:
- Stop overriding the higher-level desires of your users by trapping them in a cycle of primal satisfaction by way of social validation. Disconnect the relationship between expressed preferences and the validation of a match. People will be more honest about who they are and they will be more honest about their true desires if they are confident that their inputs will remain private. From your perspective as the matchmaker, this would be a direct improvement to the quality of your Matching function’s input data. The matches will be better and realer.
- Respect your user’s intelligence: she is your wife. Tell the user exactly that this is what you’re doing. Brand the app around ‘this is the better way to find your person.‘
- e.g. “Type 2 Dating” (reference to ‘Type 2’ thinking, the slow type, from thinking fast and slow which I have not actually read)
- Keep the number of matches constant and low. Probably capped at 1 per day (0 is not out of the question, we do not surface matches unless the threshold is of sufficient).
- company-level metrics:
- ”# of dates went on”
- +1 Company Did Good point for 1st date
- +2 Company Did Good points for 2nd date
- +3 Company Did Good points for 3rd date
- …
- people would probably be happy to privately confirm # of dates if asked of them at a reasonable cadence and with messaging ~= “you’re helping us improve this for you and for everybody else"
- "# of couples still together after X months/years”, where there are multiple instantiations of this general format and the X is varied, e.g. 1 month, 3 months, 6 months, 1 year, …;
- all of these would be trailing indicators of effectiveness
- max-bound would be “amount of time it’s been since we onboarded our first cohort, minus one month”
- ”# of dates went on”
app design:
- public profile: name, pictures, that’s it. we decided you were a match and you decided to put your trust in us as an incentive-aligned matchmaking service so get to chattin fuckers
- profile metadata provided and maintain private: objectively measured dealbreakers, bio, desires (broken down into “for sure”, “pretty sure”, “nice to have”, and “definitely not necessary but wouldn’t that be something”) (maybe just the first 3)
- Start appealing to higher-level desires of your users as opposed to trapping them in a cycle of
- Stop incentivizing people to fulfill their primal needs for social validation by connecting their expressed preferences with the number of matches they receive.
- the Match-Surfacing System is the CoffeeMeetsBagel model (X# of matches revealed per day, I say 3, maybe even 1)
- Match-Surfacing System technical design document outline:
- backend cronjob-type system is generating connections on a regular cadence (daily would be ideal, bandwith supporting)
- connections are generated
- backend embedding-connection system with some built-in form of day-to-day variance
- the variance:
- could ask the user “what are you not seeing that you’d like to see
- the variance gets applied to the base user Love Preferences metadata, via LLM summary compression (“reconcile these two things, put just a little weight on the fact that, today, the user responded to our question of ’…’ with: ’…’“) -> embed the summarization -> this is the query used for the embeddings search, which excludes all past matches.
- final step: continuously filter for quality by making new users pay a lump sum to enter
- go to market plan:
- at first, do not charge anything, and do things that don’t scale to recruit the first 1,000 in a location[^2]
- by this app’s nature, and especially if you go with the constant of one match per day, a user ‘stays fresh’ for longer; they are far less likely, by comparison to other apps, to burn themselves out or conclude ‘there are no good people here’ because they only see one piece of the puzzle each day
- ’expensive things must be good’ effect
- at first, do not charge anything, and do things that don’t scale to recruit the first 1,000 in a location[^2]
”so, basically…“
a traditional matchmaking service’s incentive structure, scaled up as an app, where LLMs facilitate the whole thing by offering infinite data processing throughput across a variety of customized + intertwined RAG architectures
A digital company pinboard
It’s a board for ‘move fast and break things’, for ‘be relentlessly resourceful’, for a screenshot of an early rejection letter, for growth rate side-by-side comparisons to competitors, company principles, aphorisms, mantras, words, mugshots, etc.
- It can’t just be text; definitely needs to be room for creative spunkiness. The founders would need to feel like they were able to customize it exactly the way they wanted. somebody call the dingboard man
- you could do this today in something like Figma, export, and then host it as an image on wherever your team manages their stuff, but this would be a centralized purpose-built solution
- it could be retro-themed. if you search “messy brutalism” or “neobrutalism” on Dribbble, and this could fit well into that mold (i.e. lean in to the fact that you’re trying to emulate this basic physical structure in an overly-complicated but elegant digital reenactment) -> examples: 1, 2
- I will be your first user
subscribe.to/YourFavoriteCreator
It’s a personal brand cross-platform subscription aggregator.
- if I subscribe to A Creator on Substack and she also has a premium X and she also has a paid curation space feed, it feels a bit strange and fragmented to subscribe to each individually
- what if we could go to A Creator’s aggregated subscription page
- 60-second domain name spitball brainstorm for inspiration: sub.com, sub.net, sub.scribe, subscri.be, its.me, itsme.sub, itsme.wanttosubscribe, allsubs.com (lol), oneplace.com, thesubscriptionaggregator.com (shock them with your length. who’s doin that? nobody!)
- requirements: B2B partnerships with all the individually-paid subscription services (e.g. Substack, Twitch, X, etc.)
- somehow, everybody needs to make more money (partnered sites, creator, )
- probably: the personal brand subscription aggregator needs to create some real value for the subscriber, enough to justify $5/$10 on top of the cost of the other subscriptions, of which you can distribute a portion to the person and any participating services
- I do not currently have good ideas about how that real value could be presented, so maybe this is like drawing the rest of the owl
- probably: the personal brand subscription aggregator needs to create some real value for the subscriber, enough to justify $5/$10 on top of the cost of the other subscriptions, of which you can distribute a portion to the person and any participating services
A viking themed gym
- viking war music is playing at all times
- rowboat-themed rowing machines??? rowboats on a water treadmill???
- workout classes: everybody get in the boat and row. one instructor at the helm, one instructor navigating while playing viking drums
- the entrance hall to the gym is filled with a bunch of viking-style Great Hall statues of fitness icons
- each deadlift rack is a shrine to Odin
- there is one larger deadlift rack; this one is not a shrine, but a throne. it is reserved for the person lifting the most. king of the hill style — you can walk up to the bar, add weight to it, and make your claim to the throne
If you want to manifest any of these, I release any (are there any?) rights to you and wish you luck. If you’d like to chat, contact me!