xAI launches Grok for iOS · RevenueCat automates EU reports
|
The Mobile Takeaway
For people who ship apps
|
Friday, October 2, 2026
|
xAI released a standalone Grok app for enterprise users on iOS to manage team access and workspace requirements.
RevenueCat now handles the reporting Apple requires for external Stripe transactions in the EU, which cuts platform commissions to 10% for small businesses.
Platform
| • | xAI launches standalone Grok enterprise app for iOS 9to5Mac Following OpenAI’s enterprise push, xAI released a dedicated Grok app for corporate users to manage team access and workspace requirements on iOS. |
| • | Magnet Forensics reportedly bypasses Apple's new iPhone Inactivity Reboot feature 9to5Mac Magnet Forensics claims it can now bypass the 72-hour inactivity timer that triggers an automatic reboot, according to internal documents obtained by 404 Media. |
| • | Apple spends €9 million annually lobbying EU officials in Brussels 9to5Mac Financial Times data shows Apple met with EU officials 257 times since June 2024, spending €9 million annually to influence regional platform policy. |
Business
 | ▶ YouTube · 62 min David Senra How Long Lake is using AI to scale legacy service businesses to $4B in revenue |
Key takeaway: Long Lake is scaling to $4 billion in annual revenue by acquiring legacy service businesses and deploying a proprietary, model-agnostic AI platform called Nexus to automate high-friction tasks like email management and vendor sourcing.
• Long Lake recently acquired American Express Global Business Travel for $6.3 billion, aiming to integrate AI into a 111-year-old institution that manages travel for half of the Fortune 500.
• The firm’s "Nexus" platform uses a shared infrastructure where 80% of the code—covering model routing, security, and agentic orchestration—is common across all business verticals.
• To handle the "last mile" of integration, 70% of the engineering team is deployed in the field to sit directly with employees and map real-world workflows rather than working from a central office.
• In the HOA management vertical, Nexus drafts complex vendor-comparison emails with 99% accuracy, reducing a 60-minute manual task to a 5-minute edit.
• Founder Alex Taubman rejects the "gut and cut" private equity model, arguing that making employees 50% more productive allows a firm to handle 50% more clients and increase revenue without layoffs.
• The company maintains a 3% hiring acceptance rate, focusing on AI engineers and deployment specialists who can bridge the gap between frontier models and messy legacy data systems.
• Taubman targets "iconic franchises" with logo retention in the high 90s and dollar retention over 100%, viewing AI as a tool to remove the "threshold resistance" that prevents technology adoption.
Why it matters: It demonstrates a viable strategy where the value isn't in selling SaaS to others, but in owning the underlying business and using custom AI agents to capture the margin of increased productivity.
Worth it: Yes, because it provides a blueprint for how to deploy agentic AI into "unsexy" industries with heavy legacy tech debt.
themobiletakeaway.com
 | ▶ YouTube · 86 min My First Million Sam Altman's coach destroys your limiting beliefs for 86 minutes straight |
Key takeaway: High-performance leadership is driven by emotional regulation and the speed of resolving conflict rather than raw effort or "grinding."
• Maximalists like Sam Altman refuse standard trade-offs, operating on the belief that they can figure out how to have both options simultaneously.
• The "Golden Algorithm" states that the emotion you try to avoid is the one you invite; an entrepreneur who avoids helplessness by refusing to delegate eventually limits their ability to scale, causing the failure they feared.
• Effort is rarely the deciding factor in startup success; Joe Hudson notes he has never seen a venture-backed company fail due to a lack of effort, but many fail due to poor relationships and decisions.
• Company health is measured by two metrics: how quickly a team has a hard conversation and how quickly they resolve it.
• When having hard conversations, leaders should state the performance gap and ask for a plan by a deadline rather than solving the problem for the employee.
• Neurological research shows that all decision-making is emotionally based; logic alone is insufficient for even simple choices like which pen to use.
• Visualizing total failure—including layoffs and going broke—removes the fear of those outcomes, allowing founders to build the company they actually want instead of acting out of self-preservation.
• The "self-reliance pattern" often prevents founders from asking for help or clearly stating what they want, usually stemming from childhood environments where they had to fend for themselves.
Why it matters: Small app founders often stall growth by falling into a self-reliance trap that prevents them from delegating or having the "unkinked" difficult conversations necessary to scale.
Worth it: Yes, because it provides the specific psychological framework used by the founders of OpenAI and Twitter to navigate high-stakes growth.
themobiletakeaway.com
 | ▶ YouTube · 6 min Fireship OpenAI DevDay 2026: Autonomous Agents and the Sign in with ChatGPT Model |
Key takeaway: OpenAI is shifting from chat boxes to autonomous Dots agents and a new Sign in with ChatGPT authentication that lets users pay for their own API usage inside third-party apps.
• OpenAI hired Peter Steinberger, creator of the viral OpenClaw agent, to lead development of Dots, which are always-on agents powered by the GPT-6 Astra model.
• A keynote demo showed an agent named Alfred migrating an app from a legacy API by reading documentation, writing code, and submitting three GitHub pull requests autonomously.
• The new Sign in with ChatGPT feature allows Plus and Pro subscribers to use their own token allowances in external apps like Vercel and Notion, shifting API costs from the developer to the user.
• New Pro tiers range from $100 to $500 per month, with the Pro 500 plan offering 25x the usage of ChatGPT Plus and Ultrafast 300 tokens-per-second access.
• The Decisions API uses a specialized Luna model to pick from a fixed list of answers in roughly 150ms, eliminating the need for expensive JSON retry loops.
• GPT-6 Astra API pricing is set at $60 per 1M input tokens and $300 per 1M output tokens, though a Sol model offers near-Astra intelligence at 20% of the cost.
• Current code quality data indicates AI-generated pull requests produce 1.4x more critical issues and 1.7x more major issues than human-authored code.
• OpenAI continues to clone successful third-party tools, recently launching Pages and Space as competitors to Google Docs and Notion.
Why it matters: For small teams, the Sign in with ChatGPT model could eliminate the financial risk of scaling high-usage AI features by making the user's subscription the primary payment method.
Worth it: Yes, because the shift to user-funded API usage and the new Decisions API directly change how you architect and price AI-integrated apps.
themobiletakeaway.com
| • | EU link-out reporting @RevenueCat RevenueCat now files every report Apple requires for Stripe Billing transactions in the EU. This lets you link out to Stripe and pay Apple 15% instead of 26% without building your own reporting pipeline. If you're in the Small Business Program, the rate for alternative payments drops to 10%. RevenueCat is also opening early access for an upcoming in-app Stripe checkout.
Link out to @stripe and pay Apple 15%, not 26%. Apple's new EU rules go live today.
RevenueCat now files every report Apple requires for your Stripe Billing transactions. No reporting pipeline to build: https://t.co/cPwQguPqBH
|
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.