Ex Twitter CEO Parag Agrawal’s Parallel: Redesigning the Internet for Artificial Intelligence

In a neoteric move, former Twitter CEO Parag Agrawal, ousted in 2022 amid Elon Musk's takeover, floats one more feather in his entrepreneurial hat- Parallel Web Systems Inc. Founded in 2023 and located in Palo Alto, the company intends to change how artificial intelligence interacts with the web at scale.

Supported by a competent team of 25, Agrawal's venture has already landed $30 million in investment from renowned places such as Khosla Ventures, Index Ventures and First Round Capital.

Building Infrastructure for AI Research

Parallel's vision is to build a cloud platform that enables large-scale web research for AI models. Agrawal says the platform adorns millions of research queries per day, for the inchoate startups and the most established companies.

ALSO READ: Poland Seeks EU Probe into Elon Musk’s Grok Over Alleged Hate Speech

The newly launched Deep Research API claims to be superior to humans and state-of-the-art models, such as GPT-5, on key research benchmarks. This tech may change how companies use data in their decision-making and build automation.

Why Does the Web Need to Evolve?

Agrawal argues that the original architecture of the internet for human uses has become incompatible with artificial intelligence. Whereas humans enter the internet through browsing a handful of pages or taking a couple of short searches, AI systems ought to browse through vast data sources in a dynamic fashion, cracking insights, and performing large-scale reasoning exercises all at the same time.

ALSO READ: ‘Eat Microsoft Alive’, Elon Musk warns Satya Nadella After GPT-5 Launch

The traditional web business models, the ones based on ads, clicks, and paywalls, do not fit into a machine-first paradigm, constituting impediments towards open-scaled knowledge access.

Principles of a Programmatic Web

Parallel introduces the concept of a "Programmatic Web," a version of the internet structured to serve AI as its second major user. The core principles are:

• Unified infrastructure integrating data, computation, and reasoning to provide actionable insights instead of mere information.
• Declarative interfaces enabling an AI to ask for its needs while leaving the actual work of determining how to satisfy it to the system.
• Fair attribution making sure that original sources receive appropriate credit for their contributions.
• Economic incentives through open markets rewarding those making contributions to knowledge and keeping that knowledge from becoming siloed away.

Towards an AI-First Future

By transforming the way AIs interact with online information, Parallel envisions making the web more accessible, trustworthy and sustainable-for both humans and machines. Agrawal has framed that mission as building for abundance, where creativity is no longer curtailed by old systems but rather unleashed via openness and collaboration. 

Essentially, Parallel envisions the internet not only as a memory for humanity but also as a memory space shared between mankind and excitingly prosperous AI that may probably power a whole new chapter of digital evolution.

ALSO READ: Elon Musk Says xAI Attracts Meta Talent Without Huge Pay: ‘Do something great’