David Shapiro’s Substack
David Shapiro
Billions of Agents
7
1
0:00
-22:40

Billions of Agents

AI's future is billions of agents haggling in digital bazaars, not one big brain.
7
1

🥈Heavy Silver

My novel HEAVY SILVER is out! You can read more about it and watch BOTH trailers here: https://www.heavysilver.net/

🧭 New Era Pathfinders

Check out my growth community for navigating the Fourth Industrial Revolution! The New Era Pathfinders is a group of people looking to find meaning and build a career in the next wave. I am now teaching FOUR frameworks for adaptation!

  • TLC - Therapeutic Lifestyle Changes for a balanced, happy, and healthy lifestyle.

  • PBL - Project Based Learning to master any skill or technology

  • Systems Thinking - To approach problems like I do, and other geniuses like Mark Zuckerberg and Elon Musk

  • RUPA - My proprietary framework specifically for pivoting into the 4IR and Meaning Economy. It means “Reduce worry, Understand impact, Prepare for changes, Adapt and align”

🚀 Project Updates

The Raspberry project is progressing well with positive internal team calls and external interest. The team is excited and getting organized as they move forward.

🤖 Future of AI Deployment

The deployment of general intelligence or superintelligence on the software side is expected to involve billions of agents spread across tens of thousands of data centers globally. These agents will be both public and private, working inside companies, between companies, and for individual users. The ability to easily spin up large numbers of agents once the underlying architecture, models, and APIs are in place will drive this proliferation.

💻 AI Model Landscape

There is a strong consensus that while closed-source foundation models like GPT-5 will likely remain proprietary due to their high cost, many cheaper open-source models will become commoditized. These open-source models are expected to rapidly catch up in capability, being only 6-9 months behind frontier closed-source models. This commoditization will make language models akin to a new type of CPU or programming language that is freely available for widespread use.

🏗️ Importance of System Architecture

The software architecture surrounding AI models is becoming just as critical as the models themselves. Multi-agent frameworks have already outperformed frontier models and will continue to do so. The real differentiator will be in how these models are integrated into larger systems and cognitive architectures. Companies focusing solely on model development may find their advantage shrinking as the emphasis shifts to overall system design.

🛒 AI Marketplaces and Economic Activity

AI-powered marketplaces using blockchain and cryptocurrency are expected to emerge. These will enable agents to find and create valuable products or services for each other. Users will be able to task their agents with finding solutions or resources, and the agents will navigate these marketplaces to fulfill requests. This could lead to a new form of decentralized, networked intelligence where innovation is rapidly facilitated through market mechanisms.

🔐 Privacy and Zero-Trust Proofs

Zero-trust proofs will allow agents to work with encrypted information without decrypting it, maintaining privacy while still proving the authenticity of information or the accuracy of processing results. This technology will be crucial for sharing sensitive data in AI marketplaces.

🌐 The Nature of Artificial Superintelligence

Rather than manifesting as a monolithic entity, artificial superintelligence is likely to emerge as a decentralized network of billions of agents using various models and frameworks. This distributed intelligence will be facilitated by marketplaces that match supply with demand across a vast ecosystem of AI services and capabilities.

🏢 Future Business Models

Companies may specialize in hosting libraries of AI models, offering API access for routing requests to the most suitable model or set of models based on the problem and budget constraints. This could lead to a competitive marketplace where AI services bid for tasks, driving down prices and favoring those that offer consistent, high-quality results.

🗣️ AI-to-AI Communication

As this ecosystem develops, AI agents are expected to spend more time communicating with each other than with humans. This shift will fundamentally change the nature of how artificial intelligence operates and integrates into various aspects of the economy and society.

Discussion about this podcast