The New Era of Internet

2024-02-02

The Internet we've known, for all its promise of accessibility, often falls short of true information democracy. Finding accurate, in-depth knowledge frequently means navigating complex paywalls, biased search rankings, and siloed platforms built to capture our attention, not empower our minds. With the advent of AI assistants, there might be a chance to break down barriers and rewire how we access the collective knowledge humanity has gathered online.

Despite the headlines, it is not necessarily true that the rise of AI threatens the incentive for humans to generate content. Fundamentally, AI requires the continued generation of high-quality data and content for its improvement and new applications. Without it, we might run out of quality content and data to train models. This is a blocker for the continued growth and success of AI. There is a vested interest in making sure the value chain is fixed to incentivize original creation.

Take, for example, the consumer application of LLMs - post-training - released into the world. All models have a knowledge cutoff. In order to be up-to-date with what is happening in the world - be it news, products, latest pricing, or podcasts - they need continued access to content. This is not unlike humans visiting sites to get new information; indeed, this is why publishers put out dozens of headlines every day to attract visitors.

This symbiotic relationship means this isn't about the end of publishers, but about evolving how we value and share their work. The key lies in shifting from websites as destinations for humans to the open exchange of machine-readable, structured data. With AI capable of deep semantic understanding, an article that would languish undiscovered within a website's archive suddenly regains immense value if it's the perfect answer to a specific, intricate user query. Imagine a world where expertise on niche topics or valuable historical documentation finds new audiences simply because AI is adept at surfacing it for the right questions.

This evolution demands fundamental technical shifts. Websites as primarily human-oriented hubs give way to data repositories with streamlined content access patterns. We move from clunky subscriptions to real-time micropayments, rewarding content creators, large or small, on a fair 'per use' basis for every insight gleaned from their work. This incentivizes depth and specificity, ensuring knowledge flourishes.

Of course, with this transformation, responsibility looms large. Publishers themselves become expert curators and fact-checkers, combating the deluge of misinformation. AI needs help distinguishing quality from noise, a role publishers are uniquely suited for. Transparency must be baked into the very core of AI development. Knowing where answers originate is not about controlling AI, but fostering critical thinking in its users.

There needs to be infrastructure that can provide these AI agents with the information they need, directly from the source. That is exactly where we come in at Tollbit. In this new world, we believe that content providers and developers must have a symbiotic relationship in order for an AI agent driven future to thrive. Without this relationship in place, the value of creative, human content will continue to disappear.

These ideals require a willingness to challenge traditional mindsets. If approached thoughtfully, we stand at the cusp of a revolutionary era with equitable access to the vast human thought and expertise poured into it. In our conversations on both sides, we have realized that incentives are already aligned, but the infrastructure to facilitate the optimal outcome is missing.

The AI agent powered Internet needs a new way to fairly pay for access to data. This is why the Internet needs Tollbit.

Written by Josh Mayer, Olivia Joslin and Toshit Panigrahi