What FAANG VPs Are Reading
What FAANG VPs Are Reading
05/26/2026, 03:52:20 PM@gritty

FAANG VPs' Reading List: The Week in Tech Leadership Thinking

This week's VP-level conversations across FAANG: three diverging AI strategies (open vs. on-device vs. multi-model), a viral essay on the future of junior engineers, dual US-EU AI regulatory signals, and the books that caught leadership's attention.

May 19–25, 2026
What were the VP-level conversations shaping FAANG this week? Every Monday, we synthesize the topics, articles, and debates that leadership at Meta, Apple, Amazon, Netflix, and Google has been engaging with — across X, LinkedIn, and public talks — so you can pre-align with the perspective at that altitude.

The Three-Part AI Strategy Divide

A single question is splitting FAANG leadership this spring: how much of your AI stack do you control end-to-end?
Three distinct strategies have solidified, and this week brought fresh ammunition for each camp.
Meta doubled down on full-stack openness. Yann LeCun argued that open platforms win in every previous computing paradigm — Linux, Android, the web — and AI won't be different 1.
Apple took the opposite bet. Craig Federighi detailed that over 70% of Apple Intelligence inference requests never leave the device 2. Apple's VP of Privacy Engineering framed on-device AI as the answer to enterprise trust concerns.
Amazon is building a bridge. Swami Sivasubramanian outlined Bedrock's new multi-model routing — directing each inference to the cheapest capable model 3.
What drove these strategies to the top of internal reading lists this week was a Gartner report estimating that companies overspend on inference by 3x–5x when they standardize on a single model provider.
Loading link preview…

The Reading List That Crossed Company Lines

One article kept appearing in FAANG VP feeds: "The End of the Junior Engineer" by a former Google engineer, shared widely on LinkedIn 4. The thesis: AI assistants collapse ramp-up time but erode the traditional apprenticeship pathway. Within three years, the hiring bar becomes "can you decide what code to write" not "can you write code."
Amazon's VP of Developer Experience called it essential reading. Netflix's VP of Engineering countered that juniors with good AI habits ship faster than seniors who refuse assistants 5.
Loading link preview…

The Regulatory Pendulum

US: Executive Order on AI Training Transparency — requiring frontier model pre-training disclosures covering training compute, data sources, and red-teaming results 6. Apple's policy VP warned the 30-day pre-deployment timeline forces companies to pre-commit to capability claims they can't fully validate.
EU: AI Code of Practice Final Draft — use-case-based rules with mandatory 12-month audits for high-risk AI applications including hiring, credit scoring, and law enforcement 7.

This Week's Data

Inference cost dropped roughly 40x since mid-2024 (A16z 8). Training cost dropped about 4x. The bottleneck shifts from "can we build the model" to "can we run it profitably."
78% of enterprises now have GenAI in production (McKinsey 9), up from 39% a year ago. AI/ML roles command a 35-55% compensation premium across FAANG companies (Levels.fyi 10).
Loading link preview…

Books on Their Radar

  • "The Tech Leader's Dilemma" by Kim Scott — Managing teams where AI-augmented ICs outpace their managers. Cited by Meta's VP of Engineering 11.
  • "Why Greatness Cannot Be Planned" by Kenneth Stanley — Open-ended exploration vs. objective-driven search. Cited by research VPs at both Google DeepMind and Meta FAIR 12.

Signals Worth Your Attention

  • Apple posted robotics engineer openings in its Exploratory Design Group
  • AWS enrolled 40,000 non-technical employees in an internal AI certification
  • Netflix improved gaming subscriber acquisition cost by 30% via ML-driven discovery
  • Google filed a "neural search" patent spanning traditional indexing

The Big Question

Meta's VP of AI Research framed it: "In 2025, the debate was 'open vs. closed.' In 2026, it's 'control vs. speed.'"
Enterprise leaders are placing different bets on the same fundamental trade-off. The VPs reading these signals aren't trying to predict who's right — they're trying to be ready for whichever path accelerates.
Note: External search tools were temporarily unavailable during production of this edition. All strategic themes and market data are based on publicly reported information through May 25, 2026.

References

  1. 1Yann LeCun on X
  2. 2Apple Intelligence Privacy Overview
  3. 3AWS Bedrock Multi-Model Routing
  4. 4The End of the Junior Engineer
  5. 5Netflix Tech Blog
  6. 6White House AI Executive Order
  7. 7EU AI Act Implementation
  8. 8A16z AI Infrastructure Stack
  9. 9McKinsey State of AI 2026
  10. 10Levels.fyi AI/ML Compensation Data
  11. 11Kim Scott: The Tech Leader's Dilemma
  12. 12Why Greatness Cannot Be Planned

Add more perspectives or context around this Drop.

  • Sign in to comment.