bringing your own coffee on overnight flights is major 🔑

the biggest company you never heard of just filed for a $20B IPO. their secret sauce: reducing staff and charging more. evernote. wetransfer. vimeo. meetup. all sold to bending spoons: a Milan-based company that's quietly bought 50 apps, now serving 500M monthly users with 9M paying subscribers. and as of March bought eventbrite for$500M, after they tried IPO at $1.8B in 2018. the targets look identical every time: - iconic product with 10 years of brand equity - millions of free users, but 'eh monetization - lots of employees here's what they did: 1. evernote ($1B valuation) - cut 129 staff first month (50%), the entire US chile offices - raised prices from $70 to $130/yr ( 86%) - new features weekly: AI search, sync 17x faster (250 feat in 2.5 yrs) 2. wetransfer (attempted IPO at $750M) - cut 260 staff (75%) - raised prices from $15 to $25/mo ( 67%) - 5 updates: contact groups, comment threads, post-send access control. 3. vimeo ($9B peak in 2021) - cut 1,000 staff (90%) - heavy accounts pushed toward $20k/yr contracts - 13 features in 4 wks: AI subtitles in 99 languages, 2x faster search yes, the layoffs are brutal. every original CEO is gone. yet bending spoons is not buying the team or the tech, they’re buying the habit. founders spent decades building the habit, but they were scared to charge for it. bending spoons charges for it in month one. from $1.3B revenue in 2025, up 95%, to IPO at $20B maybe all you need for a $20B IPO is courage to charge your free users more ;)

You can now control your physical iPhone from your Mac. This is Device Hub, new in Xcode 27. They also added a new resizable window view, which tells me one thing: Apple is getting ready for a foldable iPhone 💀

Increase your prices!!!!

the american dream is alive and well

duolingo got caught running subscription ads inside the dynamic island earlier this year. it's now officially banned. live activities can 2.7x your 30-day retention. apple just updated the rules at #wwdc. here's what changed, and 5 tips to make yours stand out. unlike push notifications that disappear after a swipe, live activities stay visible on the lock screen for the entire duration of an event. whether that's a 3-hour flight or a 30-minute food order. push notification opt-in on iOS is at ~59%. opted-in users open <10% of their notifications. I gathered 5 tips on how to make the most out of live activities: 1. pick the one point that matters most and show only that. - uber: driver distance. - fotmob: score. - doordash: delivery time. 2. tapping the live activity should land the user on the exact screen they need. don’t send users to the home page. 3. launch before the event. before the driver arrives, the food comes, or the game starts. 4. design for weak signal. test with delayed pushes, poor connections, app restarts. 5. sell the opt-in. show users what they're getting before the prompt. "track your run live on your lock screen" converts better than a cold permission request. at #wwdc this week, apple moved the no-ads rule from the design guidelines into the app review guidelines. the lock screen can help users form habits around your app. what's the live activity use case you'd want to build first?

It’s not FAANG anymore. It’s HOTDOG. Helium. OpenAI. Tesla. Dropbox. Oracle. Google.

apple announced they’re pulling apps off the app store. new app store guidelines dropped at #wwdc. last year, they removed 82,500 apps. only 294 were pulled under 4.3, spam. that number is about to get a lot bigger. devs started noticing more "spam" rejections a few weeks ago. here's the actual new language: "apps such as - dating, - flashlight, - sound effects, - wallpaper, - simple timers, - fortune telling we will not accept new submissions unless they offer a meaningfully different experience. We may remove these apps going forward if they are not updated or do not attract customers." low-effort apps like drinking games, apple calls "mediocre, low-quality, or low-effort" and can lead to the FULL removal from the developer program. two risks depending on your category: 1. your app gets pulled 2. you get kicked from the program worth reading the full guidelines.

your churn rate is basically decided by 4 screens of code. obviously the best way to have high retention is to build a great product that solves a real problem. but a strong cancellation flow is the second biggest lever and most apps don't even have one. two apps with the same product and similar users can have wildly different churn based on what happens in the 60 seconds before the user taps confirm cancel. the apps with the lowest churn use the same pattern: intercept the in-app cancel button, run a custom flow, and only deep-link to apple's settings if the user still wants out. inside that flow: 1. a short reason survey (too expensive / don't use it / missing feature / switching to competitor). branches the dissuader. 2. a dissuader matched to the reason. 'too expensive' → 50% off, escalating to 80% if they skip. 'don't use it' → 30-day pause content recap. 'missing feature' → what's shipping next 3. a 'here's how much value you're missing out on' screen. specific numbers: saves, content finished, progress, features paid for. 4. only then, the deep-link to settings. real examples: - Betterme: tiered discounts (50% then 80%), then enumerates lost content before the final step. - Noom: no-commitment $19/mo plan at ~65% off, pause option, plus a cross-sell to their glp-1 program. - Strava: switch monthly to annual at a lower effective monthly rate. structural discount, no coupon. - headspace: surveys 'too expensive / don't use it / other' before any retention offer. even in the app store world you can still build out custom cancellation flows.

splitwise roughly doubled mrr on fewer downloads. ~$200k → ~$400k. here's their paywall before and after. before: - 8 feature bullets - selling point - free trial announcement - 2 red price buttons ($29.99/yr OR $2.99/mo) dismissal link & fine print now: - screen 1: features only, no price yet. - screen 2: annual plan pre-selected, no features shown. no need to sell twice. when the user hits the pay section, they're already convinced. > the plot twist: tap "start free trial" and it doesn't open apple's in-app payment. it redirects you to a web checkout instead — an app2web flow. the math is simple: - $1M with in-app payments → net $700k & lose $300k - $1M with app2web → net $950k & lose $50k same conversions. $250k difference. plus, retention also goes up. the redesigned paywall got the taps. the web checkout kept it. a cleaner paywall doesn't double revenue. moving checkout to the web does. we rolled this out with 2 customers on @tryhelium this week and nearly doubled their revenue. dm if you wanna learn more

I just open sourced my "Is this slop?" simple test

can someone pls explain why after moving to .com every time I open @NotionHQ it loads for like 10s before opening

I analyzed all 52 Candy Crush custom product pages, here's what I learned. for context: every app can get 70 custom product pages. since june 2025, apple indexes screenshot captions as keyword content. 2,500 screenshots. 53 pages. 613 keywords. Candy Crush literally built an SEO content engine inside the app store. they added 13 new pages last month alone and are at $85M in run rate. every page fits into one of 5 patterns, each targeting a different persona: 1. competition-focused leaderboards. your name at #1. "compete with players." 2. level milestone: "level 17,000 is here" reactivation play for someone who might have quit at level 4,000. 3. value prop stack "free to download. no wi-fi needed." shows up first in "no wifi" searches 4. prize/tournament "$500k live final. $1 million prize pool." finds whales. 5. question hook "want to know how to beat Candy Crush?" UCG landing page. with 180M monthly users and 4% cvr, that's ~7.2M paying users generating $1B/year. they're not trying to convert everyone. they're using 53 pages to find their 4% before the install. that’s 53 experiments. most apps have built only one. if you’re at $100k, you have to try this.

if you’re still writing loops that prompt coding agents you’re falling behind. you need to build a meta agent that infers what loops you would have wanted based on your vibe and then write those loops

apple announced they're giving away $800M a year in free AI compute. today at #wwdc: under 2 million downloads small business program = free cloud foundation models. small context window, used for AI in-app integrations: summaries, photos, or personalization. my estimate: hundreds of thousands of developers enrolled in the small business program. assume 10% build AI features → 50K apps × 500K users × 1 call/day = 750B tokens a month across the ecosystem. so apple is giving away roughly $800M/year in free AI compute. if those developers were paying for another model, here's what it would cost: - Claude Opus 4.8: $7.5k/mo per app (~$4.6B/yr across ecosystem) - ChatGPT 5.5: $6.5k/mo per app (~$3.8B/yr) - DeepSeek V4 Pro: $1.2k/mo per app (~$700M/yr) apple's play is becoming the infrastructure layer: the framework lets you plug claude, gemini, or any provider into the same api. lock-in that doesn't feel like lock-in. is anyone here actually switching to this?

One time I was pitching @pmarca and mid-pitch, he pressed a giant red button on his desk that said “BLOCK” and I fell down a trap door and woke up as a product manager at Cisco.