The $30M AI app no one saw coming

Happy Thursday 👋 and welcome to our 80 new subscribers! Today’s issue is loaded with lessons for makers moving fast. From Apple’s AI talent crisis to $30M consumer apps built without code, to audit frameworks clients will actually pay for—we’re unpacking the wins, missteps, and ideas shaping the next wave of builders like you.

TL;DR: Apple just lost Ruoming Pang, the guy behind Apple Intelligence, to Meta’s new Superintelligence Labs. It’s a big swing in the AI talent war, and a gut punch to Apple’s already-shaky momentum. While Meta goes full throttle on AGI, Apple’s stuck outsourcing its AI chops to OpenAI… and bleeding talent along the way.

🧠 Meta’s not playing around – They’re recruiting like it’s transfer season, throwing massive signing bonuses at Apple, OpenAI, and Google engineers. Pang is just the latest scalp in Zuck’s AGI arms race.

📉 Apple’s direction looks... fuzzy – The privacy-first approach is noble, but slow. Some insiders say Apple teams feel lost. When your best AI feature is built by someone else (hi, ChatGPT), it’s not a great look.

💼 Leadership vacuum alert – Pang’s second-in-command already left last month. Replacing top AI talent isn’t easy, and doing it while also figuring out your long-term AI strategy? Rough combo.

🔥 Superintelligence Labs = serious intent – Meta isn’t just buying talent; they’re declaring war on Google and OpenAI. And when $14B acquisitions are the opening move, it’s clear they’re here to dominate.

TL;DR: Frey Chu didn’t use a fancy AI tool or keyword planner to find his next business idea—he just went to the beach. His latest video breaks down how casual, real-life moments (like noticing broken water refill stations) can reveal powerful directory opportunities. From marine mechanics to crane rentals, Frey shows how to turn everyday annoyances into niche, money-making websites.

🌊 Business ideas from a beach day – By simply observing his surroundings, Frey spotted multiple underserved markets like mobile marine services and water refill stations, with low competition and high demand.

🔍 Directories that don’t suck – He noticed the Primo water refill station directory was trash. His idea to fix this was a crowdsourced version that lets people flag broken machines in real time (kinda like the Citizens app).

⚓ Hidden gold in local services – Mobile marine mechanics and crane rentals are perfect for directories: high-ticket services with low online visibility. Just a few leads = serious revenue.

📱 TikTok as an engagement engine – Forget SEO at first. Use short-form content to crowdsource reviews and validate demand before scaling.

You don’t need to guess what people want. Just pay attention to the friction around you and then build the thing that solves it.

TL;DR: Cursor was on fire, until it quietly changed its pricing. Suddenly, users were getting surprise rate limits, weird error messages, and bills they didn’t expect. Turns out the “unlimited” $20 plan wasn’t so unlimited after all. Their founder apologized, but the damage was already done.

💸 Surprise charges leave a bad taste – The old $20/month Pro plan got quietly nerfed with fuzzy “extended limits,” leaving users confused and frustrated when they hit invisible caps.

🧠 That “unlimited” plan? Not exactly – Turns out it only covers usage in “auto mode,” which doesn’t always use the best models. If you wanted power, you hit limits fast.

🔧 The fix? A new $60 Pro Plus plan – A solid option for power users… but for small devs or casual builders, it felt like a “pay more or leave” moment.

⚡ People are looking elsewhere – Devs started canceling and airing receipts. It became clear: pricing changes without transparency = instant trust issues. Open-source tools like 1MCP are starting to gain traction. Builders want something simple, flexible, and upfront. No surprise bills. No vague rules.

TL;DR: Blake Anderson helped grow Cali: a viral AI-powered app to $30M ARR with no dev background. He used first-mover ideas, crazy-effective influencer marketing, and a product that solved emotional, human problems. His playbook proves that timing, psychology, and scrappy execution beat credentials every time.

📱 Viral products start with emotional triggers – Blake didn’t build generic AI utilities. He built apps that spoke directly to insecurity (attractiveness), curiosity (dating advice), and control (calorie tracking). It’s a masterclass in making AI feel personal, and that’s why they spread.

💬 He cracked the influencer growth loop – Instead of overthinking virality, Blake focused on creators whose comments were alive. That signal engaged reactive audiences, outperformed follower counts and drove downloads. Pair that with tight briefs and clear CTAs, and you’ve got a scalable growth engine.

👥 Co-founders aren’t optional—they’re multipliers – Equal equity. Shared values. Defined swim lanes. Blake credits strong partnerships for their velocity.

🛠️ Hardware might be the next frontier – Blake’s now eyeing local AI devices: pocket recorders, offline LLMs, smart assistants that don’t phone home. Latency + privacy will matter more as AI becomes infrastructure. And whoever builds the interface wins.

⚠️ Advice is context-dependent—don’t just copy success – Blake warns against copying top-level strategies. What worked in 2023 won’t work in 2025. The best founders test fast, ignore most advice, and stay close to what feels inevitable but underbuilt.

TL;DR: Liam Ottley breaks down how his agency pivoted from AI dev shop to full-on AI strategy partner and why that shift landed them a $250K pipeline. They didn’t worry about building flashy tools. They focused on helping businesses actually use AI in a way that works (and doesn’t fail 85% of the time).

🧠 AI isn’t the product, transformation is – Tools are getting cheaper. Advice isn’t. Companies want someone to guide the why, where, and how behind their AI plans.

💰 $60K audits, $5K entry points – Liam walks through the 4-week process that gets businesses to open budgets fast: interviews, process maps, opportunity scores, then a roadmap they can run with.

📉 Most AI projects flop – 80-85% of AI projects fail. That means companies are scared and confused. Good consultants ease the fear and turn AI from “maybe” to “mission critical.”

🚀 Free audits aren’t wasted time – Their first one lost them $30K. It also unlocked a $250K dev contract. Do the work → earn the trust → get the bigger deal.