inherent

Overview

Inherent is a London-based AI research lab developing general-purpose AI systems for scientific discovery. Its stated approach combines AI agents, coding tools, reinforcement learning, and human oversight to investigate unfamiliar research problems and improve the process by which AI systems are developed.

The company is positioned as a frontier-AI and AI-for-science startup rather than a consumer software brand. It is notable for treating the research organization itself as part of the experiment, with an emphasis on recursive self-improvement, machine-assisted experimentation, and collaboration between researchers and AI systems.

History

Inherent was founded in London by Tantum Collins, Edward Hughes, Louis Kirsch, and Kaloyan Aleksiev. The founders have backgrounds in frontier AI research, including work associated with Google DeepMind. The company was established as an AI research lab focused on building systems that can recursively improve research processes and support scientific discovery. Its public materials describe a human-machine research model rather than a conventional software-product company.

Product Lines

Manufacturing

Not applicable in the conventional manufacturing sense. Inherent develops AI models, agents, and research infrastructure in-house from its London base; no evidence indicates physical-product manufacturing or outsourced production.

Notable Products

• Faraday - A 27-billion-parameter AI Scientist agent designed to replicate scientific research using long-horizon reinforcement learning and coding tools.
• AI Scientist agents - An internal or emerging AI research-agent platform intended to help automate and accelerate scientific investigation.
• Recursive research platform - Research infrastructure and organizational systems designed to enable recursive improvement of AI research workflows.

Reputation

Inherent is perceived within frontier-AI circles as an early-stage, research-led lab focused on AI-enabled scientific discovery rather than commercial applications. Its founders’ backgrounds and emphasis on recursive improvement have attracted attention from specialist investors and AI researchers. Public evidence remains limited because the company is young and has not established a broad commercial product portfolio. The approach also raises familiar concerns about evaluation, reliability, scientific validity, and the governance of increasingly autonomous research systems.