AI Adoption Challenges, Waymo’s Big Expansion, Amazon hits $2 Trillion

Episode Summary

Executive Summary: The episode argues that enterprise AI is entering a reality check: expensive tools often need heavy handholding, while consultants are profiting by helping companies implement them. The hosts also cover Waymo’s San Francisco expansion as a major autonomous driving milestone, Amazon’s $2 trillion valuation and China-linked discount strategy, and copyright battles facing Perplexity and Suno, suggesting AI’s biggest near-term winners and losers are becoming clearer.

Main Topics: Enterprise AI is harder to implement than expected (Priority: 5/5): The hosts discuss Wall Street Journal examples showing AI assistants frequently give outdated, incorrect, or incomplete answers inside large companies, especially when data is messy or use cases are poorly scoped. They frame this as an emerging 'trough of disillusionment' for enterprise AI. Consultants as early winners of the AI boom (Priority: 5/5): The New York Times reporting on IBM, Accenture, McKinsey, and KPMG is used to show how consulting firms are monetizing AI by selling implementation, change management, and strategy services. The hosts note this may be a warning sign that AI adoption is difficult for enterprises to manage alone. Waymo’s San Francisco expansion (Priority: 4/5): Waymo opening its robotaxi service to everyone in San Francisco is presented as a major commercial and technical milestone, especially because the city’s traffic and driving conditions are more challenging than Phoenix. The segment emphasizes that autonomous driving is already real, even if scaling remains gradual. Amazon’s $2 trillion valuation and e-commerce strategy shift (Priority: 4/5): Amazon’s market cap milestone is tied to AWS strength and AI demand, but the hosts also question Amazon’s move to launch a Temu-like China-direct discount section. They worry it signals strategic drift and dependence on cheap imported goods rather than a clear brand identity. Copyright and scraping controversies around Perplexity (Priority: 4/5): Wired’s reporting on Perplexity is discussed as evidence that AI search tools may ignore robots.txt and crawl content anyway, raising serious ethical and legal concerns. The hosts argue that respecting web norms is essential if AI companies want trust. Music AI lawsuits against Suno and Udio (Priority: 3/5): The AP report on record companies suing AI music generators is used to explore the difference between inspired creation and derivative copying. The hosts note that music feels more emotionally and legally fraught than text summarization because outputs can sound directly similar to existing songs. DoorDash, delivery economics, and the lingering viral pizza arbitrage story (Priority: 3/5): The conversation closes by revisiting a past viral story about DoorDash and pizza arbitrage, using it to argue that food delivery remains economically fragile despite becoming more normalized. The hosts cite ongoing losses as evidence the business model is still not solved.

Key Arguments: Enterprise AI often fails not because companies are stupid, but because implementation is structurally difficult: data quality, prompt design, and organizational change management matter more than hype. The current failure mode of AI is not total collapse but diminished expectations: companies may renew less, slow spending, or cut tools if they don’t see reliable ROI. Consulting firms are benefiting because they sell the exact translation layer enterprises need, even though the same AI tools may eventually threaten parts of consulting work. Waymo’s SF rollout matters more than many realize because it proves robotaxis can operate safely in a dense, complex urban environment. Amazon’s move toward a Temu-style model may be strategically reactive and could dilute its long-term positioning around speed, quality, and convenience. AI companies’ willingness to scrape and summarize content while ignoring robots.txt undermines claims of ethical conduct and may trigger more lawsuits and publisher pushback. AI music generation raises a more emotionally charged copyright issue than text because listeners can directly perceive resemblance in style and composition. Food delivery is now ubiquitous, but the economics still do not look healthy enough to justify the scale of investment without further business model innovation.

Data Points: Amazon market cap: $2 trillion - Amazon joined the small group of companies to reach this valuation AWS/AI contribution: About $1 billion - Microsoft attributed roughly this amount to AI services in a recent quarter; cited as comparison for AI revenue scale McKinsey business tied to generative AI: 40% - Share of McKinsey’s business this year expected to be generative AI related KPMG generative AI opportunities: More than $650 million - Targeted US business opportunities over the past six months IBM generative AI commitments: More than $1 billion - Sales commitments related to generative AI consulting work and Watson X Accenture generative AI sales: $300 million - Sales booked last year related to consulting/technology services US management consulting industry sales: $392 billion - Expected sales this year, up 2% from a year ago Amazon closing price: $193.61 - Closing price on Wednesday when the company hit $2 trillion market cap DoorDash Q1 2024 net loss: $25 million - Used to argue food delivery profitability remains elusive DoorDash 2023 net loss: $558 million - Discussed in the context of persistent delivery industry losses DoorDash 2022 net loss: $1.4 billion - Referenced to show historical losses in the food delivery sector Chinese firms’ revenue into Meta: $7 billion - 2023 revenue from Chinese firms, much of it attributed to Temu advertising Waymo expansion cities mentioned: San Francisco, Phoenix, Los Angeles - Examples of Waymo’s gradual geographic rollout Taiwan temperature: 97 degrees - Ranjan notes the heat while joining remotely from southern Taiwan

Pivotal Quotes: "I remain an AI optimist and I'm confident that we'll get there. It's just taking a little longer than perhaps we thought." — Wall Street Journal source quoted in the episode: Used to capture the enterprise AI disappointment without abandoning the long-term bullish view "I want that trough. Give me that trough of disillusionment." — Ranjan Roy: A tongue-in-cheek endorsement of the market moving from hype to realistic expectations "This is a terrible sign." — Ranjan Roy: His reaction to consultants being among the early winners of generative AI

Implications: Expect slower enterprise AI spending growth, more consultant-led implementations, and sharper scrutiny of AI vendors. Meanwhile, robotaxis are reaching real-world scale, and copyright, scraping, and AI-generated media will face more legal and ethical pressure.