Google's removing Pro access · How Cal AI hit 15M downloads
|
The Mobile Takeaway
For people who ship apps
|
Thursday, October 15, 2026
|
Google notified Gemini AI Plus members that their subscriptions will soon downgrade to Flash-only access.
Apple added extra-large widgets in iOS 27 to prepare for the iPhone Duo. Jake Castillo explains how Cal AI reached 15 million downloads by sending thousands of influencer messages daily and pricing the app at $30.
Platform
| • | Apple adds extra-large Home Screen widgets for iPhone Duo in iOS 27 9to5Mac The update introduces a new extra-large widget format and fifteen system defaults, requiring layout adjustments for the upcoming iPhone Duo hardware. |
| • | Samsung Health adds meal scanning and personalized coaching features 9to5Google Samsung is updating its Health app with meal scanning for macro tracking and new coaching tools, increasing first-party competition for independent fitness and nutrition developers. |
| • | Google is removing Pro model access from Gemini AI Plus 9to5Google Google notified Gemini AI Plus members that their subscriptions will soon downgrade to Flash-only access, removing the more capable Pro models from the tier. |
| • | OpenAI voluntarily stops using disputed Apple trade secrets during litigation 9to5Mac OpenAI agreed in a court hearing to stop using Apple’s alleged trade secrets while the two companies continue their legal dispute. |
Monetization
 | 🎙 Podcast · 64 min Mobile Dev Memo Podcast Will consumers actually adopt agentic commerce? |
Key takeaway: Agentic commerce is currently a mirage because the average US consumer only makes 10 discretionary e-commerce purchases a month, which is insufficient volume to overcome the massive fulfillment and identity moats of incumbents.
• The average US consumer makes 47 payments per month, but once bills, food, and gas are removed, only about 10 discretionary e-commerce purchases remain—not enough to drive a mass shift to autonomous agents.
• AI-sourced traffic to major e-commerce websites currently accounts for less than 0.5% of total traffic, despite billions of monthly users on Google AI and ChatGPT.
• Amazon maintains a near-impenetrable moat with deterministic identity for 90% of US households and a fulfillment infrastructure larger than UPS.
• Marketplace take rates of approximately 17% include 10% in unavoidable costs for payments, infrastructure, and demand generation, leaving only a 6-7% profit pool mostly driven by onsite ads.
• AI search platforms like Perplexity are sending traffic to independent Shopify-style "Brand.com" sites five times faster than they are to traditional marketplaces.
• High-consideration purchases like travel or legal paperwork are unlikely to be outsourced to agents because consumers want to verify details personally when the stakes or costs are high.
• Early data shows that while LLMs lead to fewer total clicks, the conversion rates are significantly higher, allowing platforms to increase CPCs and maintain ad revenue.
Why it matters: For app founders, this data indicates that AI's immediate business value is in discovery and conversion optimization rather than trying to build autonomous purchasing tools that lack a fulfillment engine.
Worth it: Yes, it provides a necessary economic reality check on AI commerce using specific payment and marketplace cost data.
themobiletakeaway.com
 | ▶ YouTube · 70 min Sub Club by RevenueCat How Cal AI turned creators into a growth machine — Jake Castillo, Cal AI |
Key takeaway: Cal AI reached 15 million downloads and an acquisition by MyFitnessPal by prioritizing mindshare through a $30 annual price point and a massive volume of influencer-led magic moments.
• Jake Castillo sent thousands of influencer messages daily for 18 months, eventually scaling to 70 creators posting per day.
• The team intentionally priced the app at $30 per year, less than half of MyFitnessPal’s $70, to capture a younger demographic and fuel viral growth over immediate profit.
• Marketing centered on a three-second magic moment consisting of scanning food, AI analysis, and the calorie result appearing instantly on screen.
• Castillo ignored view counts to focus on the truth in the comments, looking for users asking specific questions rather than just posting emojis.
• The app used custom builds for influencers that kept the logo visible during the scan, ensuring brand recognition without requiring a verbal shoutout.
• They preferred flat-rate influencer deals over performance-based pay to capture asymmetric upside when videos hit millions of views.
• The founders claim that copycat apps actually helped them by driving search traffic toward Cal AI, which held the top ranking for AI calorie trackers.
• A proprietary renewal score was used to decide which influencer contracts to extend, factoring in comment sentiment and specific download spikes rather than just reach.
Why it matters: This provides a blueprint for using high-volume influencer marketing and aggressive pricing to disrupt established incumbents in a crowded category.
Worth it: Yes, because it details the specific operational grind and unconventional pricing choices that led to a major exit.
themobiletakeaway.com
Business
| • | The right paywall question @nickgdwn Nick adds his insights to the discussion about the paywall type. He argues that unless free users provide specific value like referrals or network effects, your product should probably just be a paid one.
Ok, this is killing me. People look at the hard vs soft paywall question the wrong way, and it's high time we straighten some things out.
If you're asking if your app should use a hard or soft paywall, you're asking the wrong question.
What you should be thinking about is what business model your app needs.
So many (and I mean SOOOO many) use a soft paywall because of some intrinsic idea that it was "the right thing" without putting ANY thought into building a solid freemium model.
- Is there a clear product-led path and value prop to upgrading?…
|
 | ▶ YouTube · 16 min Starter Story I built this app for my dad. Now it makes $18K/month |
Key takeaway: Kourosh Ghaffari built Voicy into an $18,000 per month business by targeting seniors and users with disabilities in a crowded speech-to-text market.
• The app generates $11,277 in MRR from 1,411 subscribers, with the remaining revenue coming from $260 lifetime deals.
• Ghaffari offers a 20% permanent discount for users with disabilities, who now make up 20% of his total user base.
• The initial MVP was a Chrome extension built in two weeks using ChatGPT; he later used Cursor and Claude to port the app to Windows, Mac, iOS, and Android.
• Chrome Web Store optimization provided immediate traffic after Ghaffari asked five friends for reviews, which helped him reach his first 100 subscribers.
• A YouTube series titled AI for Seniors grew his channel from 1,000 to 12,000 subscribers in one month and became a major acquisition channel.
• The marketing strategy relies on volume over perfection, using 200 blog posts and 30 landing pages localized into five languages to capture SEO.
• Ghaffari argues that visibility matters more than quality, noting that customers usually click the first link they see in search results regardless of competitor funding.
Why it matters: This is a practical blueprint for solo founders to compete against well-funded giants by finding a high-need niche and using AI tools to maintain a multi-platform presence.
Worth it: Yes, because it details specific tactics for SEO, platform expansion, and identifying a customer base that views your app as a necessity rather than a luxury.
themobiletakeaway.com
The next issue, in your inbox.
Apple and Google updates, pricing and paywalls, and the numbers behind app businesses. Get the brief every weekday morning.
Free. Every weekday. Unsubscribe anytime.