Lovable HIts $500M in ARR, Apple Announces AI Siri

Summary of Lovable HIts $500M in ARR, Apple Announces AI Siri

by Lex Fridman Podcast Fan

21m•June 9, 2026
Listen to Episode

Overview of Lovable Hits $500M in ARR, Apple Announces AI Siri

This episode covers a fast-moving snapshot of the AI and startup landscape, centered on Lovable’s explosive growth, Apple’s new AI-powered Siri and automation features, Perplexity’s IPO plans, OpenAI/Anthropic’s rumored public-market moves, Tools for Humanity’s layoffs, and Google DeepMind’s robotics accelerator. The second half shifts into a personal, practical case study on vibe coding, where the host explains how he and his wife rebuilt an old iOS app from scratch using AI tools.

Major AI and Startup News

Lovable’s surge

  • Lovable, a vibe-coding platform, reportedly hit:
    • $500 million in annual recurring revenue
    • $100 million growth since February
    • 1 million new projects launched per week
  • The host says Lovable is especially strong for:
    • Database integrations
    • AI integrations
    • Security and SEO audits
  • He argues it has become one of the best options for non-technical users building:
    • Internal CRMs
    • HR tools
    • Inventory systems
    • Other production-grade SaaS replacements

Apple’s AI Siri and natural-language automation

  • Apple introduced new AI features at WWDC 2026, including:
    • A revamped Siri
    • Natural-language creation of Shortcuts
    • Natural-language control of Safari extensions
  • The host highlights the idea that users can now “vibe code” automations in plain English, such as:
    • Sending recurring texts
    • Triggering work or calendar-related actions
    • Creating custom phone behaviors without writing code
  • He sees this as a genuinely novel direction for Apple because it moves beyond chatbots into practical device programming

Siri gets more flexible

  • The new Siri is described as getting:
    • A standalone app
    • Chat history
    • The ability to choose between different AI models
  • The host likes the model-choice approach, comparing it to his own product that lets users access many models in one place.
  • He predicts Apple may eventually push toward its own in-house model, similar to how it replaced Google Maps with Apple Maps as the default experience.

Perplexity, OpenAI, Anthropic, and IPOs

  • Perplexity says it is targeting an IPO in 2028, regardless of what happens with other AI companies.
  • The host also references:
    • OpenAI confidentially filing for an IPO
    • Anthropic reportedly pursuing a major public-market valuation
    • SpaceX also filing
  • Key takeaway: major frontier AI companies may all enter the public markets in a similar time window.

Perplexity’s model-routing strategy

  • Perplexity reportedly routes some queries to open-source models if they’re reliable enough.
  • For harder tasks, it sends requests to models like Claude.
  • The company claims this approach can reduce costs by 10–20% versus relying on a single premium model.

Tools for Humanity / World faces pressure

  • Sam Altman’s Tools for Humanity is reportedly cutting staff as regulatory issues hit its expansion.
  • The main pressure points:
    • Kenya banned the project over privacy concerns
    • South Korea fined it for privacy violations
  • The host notes that the company’s growth depended heavily on cheap enrollment in emerging markets.
  • He also points out the company’s incentives:
    • Users in some countries were reportedly paid around $50 in WorldCoin to scan their iris

Google DeepMind’s robotics push

  • DeepMind launched a three-month accelerator in London
  • It will support 15 European robotics startups across 10 countries
  • Areas mentioned include:
    • Surgical micro-robots
    • Ocean autonomy
    • Construction automation
    • Humanoid robots
    • Industrial automation
    • Waste-sorting robots
  • The host interprets this as a sign that the next AI frontier is increasingly embodied AI: systems that operate in the physical world, not just in chat interfaces.

Vibe Coding in Practice: Rebuilding an App

Rebuilding Self Pause

  • The host shares a detailed story about rebuilding Self Pause, an app he originally built years ago with his wife.
  • The original app:
    • Focused on meditation and affirmations
    • Let users record their own voice
    • Included layered background sounds and custom audio experiences
  • Originally, it cost about $200,000 to build.

What changed with AI tools

  • In just two weeks, his wife—who is pregnant—recreated the app as a native iOS Swift app.
  • They used AI tools to:
    • Clean up and improve old hand-drawn graphics
    • Regenerate assets in a new visual style
    • Rebuild buggy features
    • Add missing capabilities quickly through Claude and Xcode

Practical benefits of the rebuild

  • The native rebuild fixed a major old issue:
    • Meditation playback now loops reliably overnight without crashing
  • The host says the app’s previous version was limited by cost, maintenance burden, and technical debt.
  • With AI-assisted development, they were able to:
    • Rebuild the iOS app
    • Launch a web version
    • Improve payments and account management
    • Avoid some of Apple’s app-store revenue cuts by routing users through the website first

Broader lesson

  • The host’s main point: people can now build real apps with AI, even if they are not professional developers.
  • He strongly encourages viewers to try building if they have an idea.
  • He frames vibe coding as a major shift in how software gets made and maintained.

Key Takeaways

  • Vibe coding is moving from novelty to mainstream utility.
  • Lovable’s growth shows real demand for AI-assisted app creation among non-technical users.
  • Apple is entering the AI automation space in a more practical way than many expected.
  • The AI market is fragmenting, with different models and companies specializing in different strengths.
  • Robotics looks like the next major AI battleground, especially in Europe.
  • AI is now good enough to rebuild legacy products quickly, making old software dramatically easier to modernize.

Action Items / Mentions

  • Check out Lovable if you want to build apps without coding.
  • Watch Apple’s rollout of:
    • Siri’s standalone app
    • Natural-language shortcuts
    • Model selection in Siri
  • Keep an eye on:
    • Perplexity’s 2028 IPO target
    • OpenAI / Anthropic / SpaceX public-market developments
    • DeepMind’s robotics accelerator
  • The host also mentions:
    • His School community for vibe-coding tutorials
    • The rebuilt app Self Pause as a live example of what AI-assisted development can do