Machine, Platform, Crowd with Erik Brynjolfsson

The Evolution of Factories: From Steam to Electricity

Transition from Steam Engines to Electric Motors

• Over a century ago, factories transitioned from steam engines to electricity, initially replacing the steam engine with large electric motors while maintaining existing machinery setups.

• The early adoption of electricity in factories mirrored previous designs, simply substituting the power source without rethinking organizational structures.

• It took approximately 30 years for factories to evolve into designs where each piece of equipment had its own electric motor, allowing for greater flexibility and efficiency.

• This shift led to significant productivity increases as companies began distributing power through wires rather than relying on a single central source.

Rethinking Business Models in the Digital Age

• Today’s companies are making similar mistakes by adopting new technologies without reevaluating their business models or organizational structures.

• The discussion focuses on three principles: machine vs. mind, product vs. platform, and core vs. crowd—highlighting a rebalancing in decision-making processes.

Data-Driven Decision Making

The Impact of Digitization

• Digitization has become crucial for better decision-making; data is now considered the lifeblood that feeds machine learning and enhances productivity.

• Research indicates that companies embracing data-driven decision-making (DDD) have seen a threefold increase over five years, with DDD firms being about 4% more productive and 6% more profitable.

Artificial Intelligence in Decision-Making

• AI is increasingly taking over not just assisting but also making decisions autonomously; an example includes DeepMind's experiments with Atari games where machines learned strategies independently.

Reinforcement Learning Applications

Efficiency Gains Through Machine Learning

• In practical applications like data centers, reinforcement learning systems have improved energy efficiency significantly—up to 15% overall and 40% specifically on air conditioning usage when compared to human management.

Expanding Machine Capabilities

• Machines are now excelling in areas traditionally dominated by humans such as sensory perception (vision), physical interaction, and problem-solving capabilities.

Advancements in Vision and Recognition Technologies

Improvements in Visual Recognition

• Significant advancements have been made in visual recognition technology; error rates dropped from around 30% in 2010 to less than 5% today due to deep learning algorithms.

Voice Recognition Progress

• Similar improvements occurred in voice recognition technology, reducing error rates from 8.5% down to approximately 4.9%, showcasing rapid advancements within just one year.

Machine Learning Beyond Traditional Domains

Medical Imaging Innovations

• AI has begun identifying medical markers beyond human capabilities; machines can recognize additional markers that pathologists were previously unaware of during image analysis.

Broader Implications for Employment

The Shift from Core Competence to Crowdsourcing

The Role of Crowdsourcing in Innovation

• Professors and consultants have traditionally advised CEOs to focus on core competencies, but there is a shift towards leveraging crowdsourced solutions for innovation.

• A team at the National Institute of Health developed a system that could sequence white blood cells with 70% accuracy in over 10,000 seconds; collaboration with Harvard Medical School improved both speed and accuracy.

• The Korean community initiated a contest to invite external participants, leading to remarkable improvements in sequencing accuracy and time by utilizing diverse perspectives.

• Participants included students and professionals from various fields like mechanical engineering and crystallography, showcasing the value of diversity in problem-solving approaches.

Transitioning from Products to Platforms

• Steve Jobs initially resisted opening the iPhone platform for third-party applications, emphasizing product control before eventually allowing developers access after realizing its potential.

• Once opened, the App Store saw rapid growth with 500 applications released within days, demonstrating the power of an open ecosystem for innovation.

• Apple shifted its strategy to become an orchestrator of its platform, managing an ecosystem that encourages app development beyond their initial vision.

Advantages of Platform-Based Business Models

• Economists refer to network effects as a key advantage; larger networks create more value through demand-side economies of scale.

• Successful companies maintain strict control over user experience while also managing their ecosystems to ensure quality and trust among users.

Conclusion: Evolving Value Creation

• The division between minds (humans) and machines is rapidly changing towards greater machine involvement in decision-making processes.

• There are vast opportunities within crowdsourcing that can lead to groundbreaking innovations across industries.