Gradient Dissent• Dec 16, 2025• 1:00:52Interview

Are Humanoid Robots Actually Coming to Your Home? | Nikolaus, Rerun

From Gradient Dissent

Nico West•Founder, Rerun.ai

Executive Summary

Rerun.ai provides a high-performance, open-source logging and visualization tool for multimodal, time-series data, targeting robotics, embodied AI, and spatial computing.
The company's strategy involves an open-source visualization client to drive adoption and trust, with a commercial cloud data platform focused on managing data pipelines leading up to model training.
Built from the ground up in Rust, Rerun features a custom in-memory database and a data model inspired by game development (ECS), which has been redesigned four times to optimize for performance and flexibility.
The speaker highlights the impact of scalable ML and open-source models on the robotics industry, noting a shift towards "learning-first" approaches and identifying promising companies in the space.

12 quotes

Concerns Raised

The difficulty of monetizing a core visualization tool directly, which necessitates an open-core model.
Bridging the gap between flexible, schema-less tools for research and rigid, schema-driven production systems.
The challenge of creating standardized benchmarks in robotics due to the tight coupling of hardware and software.
The persistence of legacy systems like ROS, which may have stagnated due to historical funding models.

Opportunities Identified

Providing the critical data infrastructure for the rapidly growing "learning-first" robotics sector.
The "ChatGPT moment" has primed the industry to seek scalable, data-centric solutions for robotics.
The proliferation of open-source models and tools (e.g., from Hugging Face) is creating a larger market for complementary infrastructure.
Serving a long tail of use cases beyond robotics (e.g., finance, media) due to the tool's high performance and ease of use.

Key Themes

Open-Core Business Model for Developer Tools

Rerun deliberately open-sourced its core visualization tool, believing it's too critical and widely used across an organization to monetize directly via per-seat licenses. The commercial product is a cloud-based data platform that builds upon the open-source foundation, a common strategy to build trust and drive adoption before upselling enterprise features.

This illustrates a key business model for modern infrastructure and developer tool companies, prioritizing community and ubiquity for the core product to create a funnel for a scalable commercial offering.

Foundational Data Infrastructure for Embodied AI

The conversation emphasizes the unique challenges of managing complex, multi-rate, multimodal sensor data in robotics, which are distinct from the more mature data stacks for LLMs. Rerun focuses on the data pipelines *before* model training, addressing a critical bottleneck in the MLOps lifecycle for physical AI systems.

This highlights a significant market opportunity in building the foundational data layer for the next wave of AI applications that interact with the physical world, where data complexity is a primary challenge.

Iterative and First-Principles Product Development

Rerun built its high-performance stack from scratch in Rust, including a custom in-memory database and a new file format. The company has also redesigned its core data model four times, demonstrating a willingness to undergo significant engineering effort to perfect the product's core architecture for user needs.

This approach shows a deep commitment to solving core technical problems for a target user base, prioritizing long-term performance and flexibility over short-term development speed or reliance on off-the-shelf components.

The Evolving Robotics Landscape

The discussion points to a "ChatGPT moment" for robotics, where advances in scalable machine learning and the availability of open-source models are enabling more practical, "learning-first" approaches. This shift is creating new opportunities and accelerating progress in historically difficult areas like manipulation.

This provides context on the current state of the robotics industry, indicating that the field is at an inflection point where data and ML infrastructure are becoming increasingly crucial for progress.

Balancing Flexibility and Production-Readiness

A key design challenge discussed is creating a tool that is both low-friction and schema-less for researchers during exploration, while also being robust and able to integrate with the rigid, pre-defined schemas of production systems. Rerun has worked to solve this tension through its iterative data model design.

This is a universal challenge for MLOps and developer tools, which must serve users across the full lifecycle from initial experimentation to production deployment, requiring a careful balance in API and system design.

Get started free

Topics

Processed Apr 3, 2026 yt-dlp + mlx-whisper + Gemini