In-Datacenter Performance Analysis of a Tensor Processing Unit
The seminal paper from Google detailing the design, architecture, and datacenter performance of the first-generation custom ASIC Tensor Processing Unit.
Page 6 of 10.
The seminal paper from Google detailing the design, architecture, and datacenter performance of the first-generation custom ASIC Tensor Processing Unit.
Proposes that traditional software moats are being commoditized by open-source and cloud infrastructure, and the new defensibility lies in proprietary data and AI-driven Systems of Intelligence.
Documents the immediate impacts of the 2017 YouTube advertiser boycott (the 'Adpocalypse') on independent video creators, highlighting their extreme vulnerability to platform-level algorithmic and policy shifts.
Addresses the puzzle of why rapid AI progress hasn't boosted productivity growth yet, proposing a J-curve of intangible asset development.
A comprehensive survey of neuromorphic computing systems, analyzing hardware implementations, designs, and key challenges for neural network deployment.
Balaji Srinivasan provides an analytical perspective on tokens, categorizing them and describing how they enable decentralized applications to bootstrap networks and fund open-source development.
An influential examination of how decentralization and blockchain technology can address security and privacy challenges in Internet of Things (IoT) architectures.
Explores the economic impact of blockchain technology through the reduction of verification and networking costs in decentralized ecosystems.
Investigates how blockchain technology could revolutionize corporate governance by providing real-time, transparent ownership records and auditing.
A framework for building iconic companies by focusing on a three-tier hierarchy of user engagement: growing core daily actions, retaining users through accruing benefits and mounting loss, and establishing virtuous lo...
Distinguishes strong from weak network effects and how marketplace design choices determine which kind a business gets.
Surveys emerging platforms and argues open-sourced deep learning tools create startup opportunities in AI.
An analysis of the technological roadmap and physical limits facing semiconductor dimensional scaling in the post-FinFET era.
Argues traditional linear pipeline businesses must convert to platforms to harness network effects or be displaced by those that do.
A seminal post mapping out the architectural layers of the decentralized web, comparing the components of decentralized apps (storage, computation, identity) to traditional web architecture.
Critiques previous task-based automation predictions by arguing that individual jobs consist of heterogeneous tasks, showing only 9% of OECD jobs are highly automatable.
Explains policy gradient reinforcement learning by teaching a network to play Atari Pong from raw pixels.
A comprehensive, academic textbook and course outline providing a rigorous technical introduction to decentralized cryptocurrencies.
Contends the EU case against Google is the first of many because the end game of aggregation is monopoly.
Applies classic social contract theories to understand governance mechanisms within decentralized blockchain networks.
Taxonomizes types of network effects and explains why not all of them create equally durable marketplace defensibility.
A foundational scholarly analysis mapping the emergence of the web-native 'creator' class and the social media entertainment system.
A key essay on how digital platforms have become the foundational infrastructure for digital labor, creative production, and economic organization.
Proposes regulatory approaches to slow bad actors and speed good actors in developing machine intelligence.