GPT-5 comes alive, Apple is struggling, and Ripple bought Rail | E2162

Episode Summary

Executive Summary: The episode centers on a broad critique of Apple under Tim Cook, arguing that the company has prioritized operational excellence and stock performance over breakthrough innovation, while AI is reshaping interfaces and product categories. The hosts also cover GPT-5’s mixed reception, Ripple’s acquisition spree amid new stablecoin regulation, concerns about AI therapy and state-level AI rules, Anthropic’s rapid revenue growth, and earnings from DoorDash, Airbnb, and Bumble.

Main Topics: Apple’s innovation slowdown under Tim Cook (Priority: 5/5): The hosts argue Apple has become overly focused on supply chain efficiency, stock buybacks, and political positioning while failing to deliver the kind of bold product leaps seen under Steve Jobs. They cite Siri’s weakness, the underwhelming Vision Pro, and missed opportunities in cars, TV, health, and AI. ChatGPT-5 launch and the commodification of LLMs (Priority: 5/5): The conversation evaluates GPT-5 as incrementally better, especially for correctness and coding, but not a dramatic leap. The hosts suggest AI models are converging, while third-party products like Cursor, Lovable, and Bolt may still matter because they add workflow and product-layer value. AI governance, therapy, and regulation (Priority: 4/5): The episode debates whether AI therapy should be regulated, whether states should set AI rules, and whether platform-level age verification should be required for minors. Illinois’ ban on AI therapy is discussed as a preemptive safeguard, while broader concerns focus on misuse, echo chambers, and consumer misunderstanding of LLMs. Ripple, stablecoins, and crypto’s regulatory turning point (Priority: 4/5): Ripple’s acquisition of Rail is framed as a sign that stablecoins are becoming a legitimate payments business after the Genius Act. The hosts argue regulation can enable new markets when updated intelligently, and they contrast viable crypto models with speculative or underused tokens like XRP. Anthropic’s explosive growth and AI infrastructure spending (Priority: 4/5): Dario Amodei’s comments about Anthropic reaching massive revenue growth are used to illustrate how quickly AI businesses can scale. The hosts discuss the broader AI capex wave, the high cost of data centers and chips, and uncertainty about whether the market can eventually justify the spend. Earnings and product strategy across consumer internet companies (Priority: 3/5): DoorDash is praised for strong growth and expansion into acquisitions and logistics innovation; Airbnb is described as growing but facing valuation pressure and North American softness; Bumble is seen as struggling with youth dating behavior and experimenting with a friendship-focused BFF product.

Key Arguments: Tim Cook has been excellent at operations and navigating politics, but Apple lacks the product-manic, visionary leadership that drove the iPhone, iPad, and MacBook Air era. Apple’s strongest post-Jobs contribution may be silicon, not consumer software or new product categories. GPT-5 appears meaningfully better in correctness and coding, but AI progress is becoming more incremental and competitive rather than singularly transformative. Third-party AI application layers will likely survive model upgrades because they offer specialized workflows and product polish beyond raw model access. LLMs are useful for summarization, search, and organization, but unreliable as a source of truth; users often misunderstand that they are optimized to please, not to verify. AI therapy poses special risks because it can mimic empathetic conversation without medical judgment, licensing, or safety guardrails. Stablecoin regulation can unlock legitimate crypto use cases by replacing vague legal uncertainty with a workable framework. Apple, and perhaps other major tech firms, missed opportunities to own obvious AI-first product categories because management incentives favored efficiency over experimentation. Consumer apps like DoorDash and Airbnb are still strong businesses, but travel and delivery are maturing and face valuation and demand normalization. AI-driven memory and personalization may deepen echo chambers if users begin treating chatbots as trusted, bias-free authority figures.

Data Points: Tim Cook’s CEO track record: 2011–present referenced indirectly; clip from 2019 says he should step down - Used to contrast Apple’s current leadership with Steve Jobs-era innovation Apple stock performance under Tim Cook: Up 250%+ (speaker-estimated) - Mentioned as evidence of strong financial performance despite weak innovation Apple dividend and buybacks: Huge dividend; large buyback program - Cited as part of Apple’s financial-engineering success Vision Pro product cadence: Only one major release, no follow-up model discussed - Used to argue Apple failed to iterate on AR/VR GPT-5 launch: Live for subscribers on chatgpt.com - Hosts discussed initial impressions shortly after launch GPT-5 performance on coding: Fewer false positives; about 80–90% of answers felt right vs. GPT-4.x roughly 50/50 - Based on Lon Harris’s live testing for research questions Humanity’s Last Exam: 2,500 questions - Referenced as an objective benchmark for AI model performance Anthropic revenue: $4 billion+ in 2025; reporting already pegs it above $5 billion - Dario Amodei interview and market reaction Anthropic 2023 revenue expectation: $100 million in first year from zero starting revenue - Illustrates how quickly expectations scaled and were later validated DoorDash orders: Up 19% year over year to 761 million - Earnings update cited as evidence of continued platform growth DoorDash revenue: Up 24% year over year to $3.3 billion - Strong quarterly performance Airbnb revenue: $3.1 billion - Quarterly revenue mentioned alongside softer growth outlook Airbnb net income: $642 million - Reported in the earnings discussion Airbnb growth in nights/seats booked: 7.4% - North America softness affected overall growth Ripple acquisition of Rail: $200 million - Stablecoin payments startup acquisition after Genius Act passage Rail funding prior to acquisition: $10.4 million Series A - Galaxy Ventures led the round last July Rail scale: 10% of world B2B stablecoin payments; about $3.6 billion in value - Cited as evidence of meaningful adoption Ripple XRP holdings: $122 billion worth of XRP, about 41% of total supply - Used to describe Ripple’s asset treasury potential 401(k) alternative assets executive order: Private equity and crypto permitted in retirement accounts - Discussed as expanding access but increasing retail risk Diversification benchmark for angel investing: 30–40 seed bets - Used to explain why ordinary 401(k) investors may struggle with private assets Illinois AI therapy ban: Signed Monday - Governor Pritzker’s law prohibits AI therapy in the state Australia and UK social media age verification: Under 16 restricted; age verification required - Example of platform regulation for minors Korean social media registration: Requires Korean phone number and often resident ID/alien registration card - Used as a model of strong local identity verification

Pivotal Quotes: "I think there's a case for Tim Cook, you know, leaving." — Jason Calacanis: Early criticism of Apple’s current leadership and innovation direction "It's disappointing. That's how I feel. I feel disappointed." — Jason Calacanis: Summarizing the emotional reaction to Apple’s post-Jobs product trajectory "We're building these things not for truth, we're building them to please people." — Jason Calacanis: Core warning about LLMs, memory, and the risks of trusting AI outputs too literally

Implications: The episode suggests Apple may be entering a long plateau unless it accepts riskier, founder-led innovation. AI will keep improving, but trust, regulation, and product-layer differentiation will decide winners. Crypto’s legitimization and consumer internet earnings show strong markets still reward execution.