> Markdown version of [/videos/1452-robots-2-0-when-artificial-intelligence-meets-steel?t=669](https://www.wearedevelopers.com/videos/1452-robots-2-0-when-artificial-intelligence-meets-steel?t=669). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Robots 2.0: When artificial intelligence meets steel

Embodied AI is moving from simulation to reality. Master ROS2, sensor fusion, and hardware abstraction to build the next generation of autonomous robots.

- **Speakers:** [Thomas Tomow](https://www.wearedevelopers.com/@thomas-tomow)
- **Event:** World Congress 2025
- **Published:** August 20, 2025
- **Duration:** 28:19
- **URL:** https://www.wearedevelopers.com/videos/1452-robots-2-0-when-artificial-intelligence-meets-steel

## Summary

Artificial intelligence is shifting from software confined to screens into embodied agents that navigate and interact with the physical world. Similar to historic technological tipping points in electricity and transportation, robotics is experiencing a massive leap forward driven by a convergence of technological forces: the exponential growth of Large Language Models (LLMs) and Vision Language Models (VLMs), accelerated edge AI hardware, and advanced simulation capabilities. These elements enable robots to actively comprehend their environments, moving machines far beyond the rigid, 24/7 pre-programmed industrial arms of the past into adaptive, autonomous behavior.

Modern robotics relies heavily on Vision Language Models acting as a cognitive core, allowing machines to ingest raw visual data and process it using language comprehension to execute real-time decision-making. This technology is already deploying across global supply chains, powering autonomous defect detection in manufacturing facilities, optimizing food logistics, and enabling independent medicinal delivery in healthcare. However, bringing these systems to life introduces immediate developer challenges categorized by edge computing efficiency, critical power consumption constraints, and ensuring the absolute security of interactions within dynamic human environments.

Creating an autonomous system requires developers to build complex abstraction layers that bridge mechanical hardware limitations with high-level software commands. Utilizing frameworks like the Robot Operating System (ROS2) alongside Small Language Models (SLMs) and near real-time simulations, engineers must creatively map physical components—like raw motor telemetry—to intelligent algorithmic controls. A critical requirement for this spatial awareness is executing robust sensor fusion, which marries 2D YOLO-based object detection with precise 3D Lidar measurements to grant machines true environmental mapping. Mastering these hardware and software abstraction layers remains the defining hurdle for developers aiming to resolve edge cases and bring functional, embodied AI out of simulation.

**Keywords:** embodied ai, vision language models, sensor fusion, lidar spatial mapping, yolo object detection, robot operating system, ros2, hardware constraints, autonomous machine navigation, real-time robotics simulation, small language models, hardware abstraction layers, edge ai computing efficiency, robotic power consumption

## Chapters

1. **AI transcending software to enter the physical world** (00:00) — Artificial intelligence is moving beyond screen interfaces to interact directly with physical environments through robotics.
1. **Historical parallels in technological paradigm shifts** (00:49) — Past innovations like electricity and automobiles demonstrate how skeptical audiences eventually adapt to transformative technological changes.
1. **Transitioning artificial intelligence into physical manifestations** (02:58) — Large language models are evolving from digital chatbots into physical entities that interact with complex real-world environments.
1. **The evolution of artificial helpers and industrial machines** (03:56) — The pursuit of robotic labor spans from visionary historical concepts to highly efficient modern logistical warehouses.
1. **Capabilities of state-of-the-art modern humanoid robots** (05:26) — Leading platforms from companies like Tesla and Boston Dynamics showcase unprecedented robotic dexterity and environmental interaction.
1. **Core drivers fueling modern robotic capabilities** (08:46) — Exponential advances in language models, accelerated hardware chips, and extensive simulation data converge to power cognitive machines.
1. **Processing real-world environments with vision language models** (10:18) — Vision language models act as central intelligence hubs by fusing visual input with semantic understanding for autonomous decisions.
1. **Automating complex industry operations with embodied intelligence** (11:09) — Autonomous robots execute precision tasks across automotive manufacturing, food safety logistics, and sensitive hospital environments.
1. **Navigating hardware constraints and edge computing challenges** (12:23) — Engineering interactive robots requires overcoming significant battery power limits and latency constraints to maintain operational safety.
1. **Leveraging open source frameworks and environmental simulation** (14:08) — Developers construct rapid autonomous systems utilizing specialized operating layers alongside localized small language models and virtual testing environments.
1. **Bridging foundational coding with autonomous robotic navigation** (15:06) — Developing autonomous robotic behavior forces engineers to account for numerous edge cases inherent in uncontrollable physical spaces.
1. **Structuring mechanical commands through software abstraction layers** (17:25) — Standardized operating sets translate high-level software instructions into granular mechanical actions for precise physical execution.
1. **Mapping physical environments via extensive sensor fusion** (20:50) — Synthesizing two-dimensional camera feeds with point cloud data enables machines to construct accurate three-dimensional operational maps.
1. **Processing real-time spatial data for structural navigation** (22:42) — Live ingestion of integrated camera and depth statistics provides immediate feedback loops for accurate trajectory planning.
1. **Executing selective object identification through edge processing** (24:17) — Coupling visual streams with localized language models enables complex filtration commands despite unstable hardware dependencies.

## Related Moments

- [Shifting focus from large language models to physical AI](https://www.wearedevelopers.com/videos/100039-physical-ai-for-the-next-wave-of-industrial-digitalisation) (from "Physical AI for the Next Wave of Industrial Digitalisation")
- [Transitioning from rigid machines to autonomous AI robots](https://www.wearedevelopers.com/videos/472-enhancing-ai-based-robotics-with-simulation-workflows) (from "Enhancing AI-based Robotics with Simulation Workflows")
- [Moving from AI software buzzwords to building actual robots](https://www.wearedevelopers.com/videos/100097-how-i-built-my-own-intelligent-robot-arm-from-scratch) (from "How I built my own intelligent Robot Arm from Scratch")
- [Introduction to physical AI and the physical world](https://www.wearedevelopers.com/videos/100295-from-perception-to-autonomy-building-agentic-edge-ai-robots-with-ros-2) (from "From Perception to Autonomy: Building Agentic Edge AI Robots with ROS 2")
- [Progressing from standard robotics to cognitive learning systems](https://www.wearedevelopers.com/videos/86-rpa-in-the-public-sector) (from "RPA in the Public Sector")
- [Technical catalysts driving real-world artificial intelligence](https://www.wearedevelopers.com/videos/100141-physical-ai-the-era-of-intelligent-machines) (from "Physical AI: The Era of Intelligent Machines")

## Related Articles

- [MLOps And AI Driven Development](https://www.wearedevelopers.com/magazine/82-mlops-and-ai-driven-development)
- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)
- [WWC24 Talk - Scott Hanselman - AI: Superhero or Supervillain?](https://www.wearedevelopers.com/magazine/469-wwc24-talk-scott-hanselman-ai-superhero-or-supervillain)
- [How we Build The Software of Tomorrow](https://www.wearedevelopers.com/magazine/120-how-we-build-the-software-of-tomorrow)

## Related Jobs

- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/1355348-machine-learning-engineer) at **TWILIO**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [Machine Learning Engineer](https://www.wearedevelopers.com/jobs/ext/588393-machine-learning-engineer) at **Twilio**
- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**
- [Senior AI Agent Software Engineer (Go, Python) (m/f/x)](https://www.wearedevelopers.com/jobs/48277-senior-ai-agent-software-engineer-go-python-m-f-x) at **Dynatrace**