Uber's AI Spending, the Pope on AI, TikTok and Universal's AI Music Deal

Summary of Uber's AI Spending, the Pope on AI, TikTok and Universal's AI Music Deal

by Lex Fridman Podcast Fan

18m•May 26, 2026
Listen to Episode

Overview of Uber's AI Spending, the Pope on AI, TikTok and Universal's AI Music Deal

This episode is an AI industry roundup covering four major themes: enterprise AI spending and ROI concerns, the Vatican’s nuanced new AI guidance, the gap between enterprise AI ambition and infrastructure readiness, Google’s push to reshape search with AI agents, and major changes in AI music licensing across TikTok, Universal Music Group, and Spotify. The host argues that the biggest winners will be companies that redesign workflows around AI rather than simply layering tools on top of old systems.

Uber’s AI Spending and the ROI Problem

What’s happening

  • Uber’s president said AI spending is becoming “harder to justify.”
  • The company reportedly burned through its annual AI budget in just four months.
  • Uber said it is struggling to prove that the spending is translating into measurable results.

Key context

  • Uber reportedly spent $3.4 billion on R&D in 2025, up 9% year over year.
  • The company is framing AI investment as a tradeoff against hiring fewer humans.
  • Executives described the link between token usage and business outcomes as difficult to measure, even “impossible” in some cases.

Host’s take

  • The real problem may be bad incentives and poor measurement, not AI itself.
  • At large companies, teams may optimize for metrics like “AI usage” instead of actual productivity or shipped outcomes.
  • Benefits from AI can also be delayed, especially for projects still in progress.

Pope Leo’s AI Document: Ethics, Power, and Human Dignity

Main themes

  • The Pope’s roughly 200-page AI document is presented as not anti-tech and not anti-AI.
  • It frames AI as a milestone of human ingenuity that can uplift humanity if guided by ethics and shared responsibility.

Important points highlighted

  • AI is described as more than a technical issue — it is an anthropological question about how humans respond to it.
  • The document is strongly pro-worker:
    • warns against surveillance and deskilling
    • calls mass unemployment from AI a “true social calamity”
  • It also treats AI as a potential equalizer, especially for:
    • developing nations
    • education
    • economic development
    • preserving local cultures and languages

Notable detail

  • The Vatican reportedly presented the document alongside Anthropic co-founder Christopher Olah (the transcript’s “Chris Ola” appears to refer to an Anthropic leader), signaling engagement between religious institutions and frontier AI companies.

Enterprise AI Adoption: Demand Is High, Infrastructure Is Missing

What the report says

  • 85% of enterprises plan to adopt agentic AI by 2029.
  • 76% lack the infrastructure needed to deploy it effectively.

Core issue

  • Many companies are bolting AI onto legacy systems instead of redesigning workflows and systems around it.
  • The result is shallow adoption: chatbot overlays rather than real operational change.

Reported impact of agentic AI

  • Processes can accelerate by 30–50%
  • Low-value work time can drop by 25–40%
  • McKinsey projects that three-quarters of current jobs may need redesign, upskilling, or redeployment by 2030

Host’s takeaway

  • The real winners will be companies that rebuild from scratch around AI.
  • Organizations that simply “add AI on top” of old systems will likely miss the ROI.

Google’s AI Push: Search, Web, and YouTube

What Sundar Pichai is saying

  • Google is actively reshaping search and the web with AI agents.
  • Google’s major products now run on unified Gemini infrastructure.
  • The company is standardizing its AI stack across products like:
    • Search
    • Maps
    • YouTube
    • Chrome

Why it matters

  • Google has unmatched distribution, so changes to its AI experience can shift the entire market.
  • The transcript notes that Google and Gemini products have reached a billion monthly active users across key surfaces.

Product implications

  • Google is training AI models on YouTube content and using AI to:
    • summarize videos
    • jump to relevant timestamps
    • reduce the need to watch long intros or filler sections
  • The host sees this as solving a problem the platform itself helped create: overly long, SEO-driven, or monetization-driven video structures.

Broader trend

  • Google appears to be moving like a startup:
    • launching lots of products
    • seeing what sticks
    • cutting what doesn’t
  • The host mentions NotebookLM as a surprise breakout success.

Search shift

  • The episode notes that Google Search is increasingly dominated by AI-generated answers.
  • A Condé Nast executive reportedly said the company is planning for zero search traffic from Google, and Sundar Pichai acknowledged that as a possible future strategy.

TikTok, Universal Music Group, and AI-Generated Music

What changed

  • Universal Music Group and TikTok renewed a licensing deal.
  • The deal includes efforts to:
    • remove unauthorized AI-generated music
    • improve artist attribution
    • address content moderation issues around AI music

Why this matters

  • AI-generated tracks have already gone viral, including fake songs attributed to real artists.
  • The transcript references AI-generated tracks imitating artists like Drake and The Weeknd, which previously accumulated millions of streams before being removed.

Industry context

  • UMG had earlier pulled its catalog from TikTok over disputes related to AI content moderation and licensing.
  • The new deal appears to be part of a broader move toward opt-in licensing and royalty sharing for AI-generated covers or voice-based remixes.

Spotify tie-in

  • The host notes similar moves with Spotify and Universal, where artists may be able to opt in and earn royalties when their voice or work is used in AI-created content.

Main Takeaways

  • Enterprise AI is not failing because the technology is weak; it’s failing when companies don’t redesign around it.
  • ROI measurement is becoming the central debate in AI spending, especially for big firms like Uber.
  • The Vatican’s AI stance is more nuanced than headline coverage suggests: it’s cautionary but not anti-AI.
  • Google is using its distribution advantage to reshape search, YouTube, and the broader web with AI.
  • AI music is moving toward a licensing-based future, where platforms, labels, and artists negotiate opt-in rules and royalties rather than trying to fully ban synthetic media.

Closing Notes

The episode closes with a promo for the host’s AI newsletter and a listener review request. The overarching message is that AI is rapidly moving from experimentation into institutional policy, enterprise infrastructure, and platform economics — and the next phase will be defined by who adapts their systems, incentives, and rights management most effectively.