KI in der Industrie

In this episode, I sit down with Steven Yates, CTO and co-founder of Federant, to dive deep into the urgent need for runtime governance in edge AI. Drawing on decades of experience in embedded systems and PLC design, Steven reveals why the shift to the edge demands more than just powerful inference—it requires robust, local authority to keep operations safe when connectivity falters. We unpack real-world incidents where lack of governance led to costly mishaps, and explore how new open-source solutions are bridging the gap between cloud convenience and industrial reliability. If you think cloud SLAs are enough for industrial AI, this conversation will make you rethink the fundamentals. Join me as we explore the future of safe, autonomous operations—and why the old rules of industrial control are more relevant than ever.

In this episode, we dive deep into the evolving landscape of industrial AI, exploring why simply offering an API is no longer enough for true business value. We reconnect with leading minds behind recent billion-euro deals and discuss the journey from foundational models to real-world industrial impact. Our conversation with Davy Demeyer spotlights the creativity gap in industry, the challenges of automation standards, and what it takes to build the next killer application. We share firsthand insights from pioneers, reflect on lessons learned, and debate how agents, humans, and deterministic code generation will shape tomorrow's factories. If you want the latest and greatest in industrial AI—and why it matters for your business—this is the episode you can't miss.

It's Thursday morning and I know you're expecting a fresh episode, but today brings a twist—I have exclusive news that can't go live just yet. We're delaying our usual release by one day to bring you something truly special and timely. In the meantime, I invite you to consider joining us at this year's top AI events or reaching out for more information. Thank you for your patience and understanding as we prepare to share these exciting updates with you. I look forward to reconnecting with you tomorrow for a can't-miss episode.

We have a new podcast partner - please welcome NXAI! In this episode, I dive deep into the evolving world of AI-driven CAD and CAM with Dr. Christian Heining. We unpack why generative AI is capturing the attention of industrial designers and whether the latest integrations—like Anthropic's connector for Autodesk Fusion—are truly game-changing or just another step on a long journey. Dr. Heining shares real-world experiments, candid takes on current limitations, and what it means for both startups and big players in the industrial AI space. We discuss the speed of change, the challenges of adapting teams to new tools, and what's still missing before AI can fully transform the machine-building industry. If you're curious about the intersection of AI, engineering, and the future of design, this episode is a must-listen.

In this episode, we dive straight into the heart of the industrial AI landscape – beyond flashy headlines and Instagram-worthy trade shows. We tackle the uncomfortable truths behind 'white label' solutions, the real value (or lack thereof) in APIs, and why simply following the hype won't cut it for true innovation. We share our candid impressions from Hannover Messe, exploring the tension between substance and showmanship, and what it means for the future of industrial-grade AI. Plus, we sit down with Dr. Ferri Abolhassan from T-Systems to uncover how Germany's new AI Factory in Munich is aiming to rewrite the rules for secure, sovereign, and scalable AI infrastructure. Join us for honest insights, pointed critiques, and a look at where industrial AI must go next if we want to deliver real value.

In this episode, I sit down with Bernard, co-founder of Ethon AI, to explore the seismic shift happening as machines move from simply reacting to truly understanding their own processes. We dive into how knowledge graphs, causal AI, and computer vision are transforming manufacturing—from chocolate factories in Switzerland to global giants like Siemens and Bosch. We discuss the real impact of connecting siloed industrial data, the journey from correlation to causation, and what it means for the people on the shop floor. Bernard shares concrete examples of AI-driven quality control, autonomous process optimization, and the evolving role of operators. Join me for a conversation that peeks into the future of manufacturing intelligence—and what it takes to build the digital quality engineers of tomorrow.

In this episode, we dig deep into the evolving landscape of industrial AI, from April Fool's pranks to real advances in robotics and automation. We break down how the line between hype and reality is blurring, and why it's more challenging than ever to separate fact from fiction in the age of agentic AI. We welcome Jonas Messner from NEURA Robotics to unpack how their 'robot gym' is collecting real-world data, why new forms of memory and multi-modal sensing are critical, and how open platforms are redefining collaboration in physical AI. Join us as we connect industry history, current breakthroughs, and bold visions for the future—where robots learn, adapt, and even monetize their skills in dynamic environments.

In this episode, we dive deep into the challenges facing time series AI model leaderboards, from hidden information leakage to the complexities of benchmarking foundation models. I sit down with Marcel Meyer to unpack why traditional approaches fall short and how our new TS Arena leaderboard is setting a new standard for fair, future-proof evaluation. We explore the pitfalls that plague current benchmarks, the surprising ways data contamination can skew results, and the innovative pre-registration protocol we've developed to keep evaluations honest. If you've ever wondered what it takes to build a truly trustworthy AI leaderboard—or why it matters for industry and research alike—this conversation is packed with insights you won't want to miss.

In this episode, we dive into the cutting edge of industrial AI, from the latest advancements in large language model distillation to the transformative power of AI agents in manufacturing. I sit down with Tanmay Aggarwal, CEO of Lambda Function, for a deep conversation about autonomous CNC programming, real-time shop floor integration, and why memory and causal inference matter for the next wave of physical AI. We also unpack industry trends – from SD card AI accelerators to Europe's strategic push for data sovereignty – and share candid thoughts on what it takes to move from AI-assisted workflows to true autonomy. If you're curious about how AI is reshaping industrial strategy, or just want to hear how real-world manufacturing is meeting the future, this one's for you. Join us as we connect the dots between research breakthroughs, practical deployments, and the evolving competitive landscape of industrial AI.

In this episode, I sit down with Tobias, a robotics and automation expert, to unpack what AI agents and generative AI mean for the future of industrial automation. We explore how combining human expertise with advanced agents can optimize production cycles, break down technological silos, and open new possibilities for greenfield automation projects. From hands-on case studies to the hurdles of legacy systems, we dig into what works today, what doesn't, and where the next breakthroughs might come from. I share insights from our own experiments—what happens when you let an AI agent optimize a real automation cell, and why, for now, human know-how still matters. If you're curious about the practical realities of AI in production lines and what's hype versus what's already delivering value, this conversation is for you.

In this episode, we dive into the real-world impact of AI-powered knowledge platforms for industry. We sit down with Jakub from Zeta Alpha to uncover how companies like Festo are leveraging advanced search, RAG, and agentic AI to unlock decades of expertise buried in documents and data silos. We share practical examples, discuss integration challenges, and debate the future of specialized versus generic AI in the industrial landscape. You'll hear why trust, customization, and domain focus are key for adoption – and how European innovation is shaping sovereign AI solutions. Join us as we reveal what it really takes to turn industrial knowledge into a competitive advantage.

In this episode, we dive into the resurgence of world models in AI and explore why the industry is revisiting neurosymbolic approaches. We challenge the dominance of data-centric models and discuss how combining symbolic knowledge with machine learning could unlock new breakthroughs. Our conversation highlights Europe's unique strengths and potential as a leader in this evolving landscape. We also address the importance of domain expertise and mathematics, and question whether the current AI race is stifling real innovation. Join us for a candid discussion on what the next wave of AI could—and should—look like. We thank our partner [**SIEMENS** ](https://www.siemens.de/de/)

In this episode, we dive into the evolving world of industrial AI, exploring everything from CNC innovation in Tel Aviv to the surprising rise of digital cousins over traditional digital twins. We unpack how startups and established players alike are reimagining automation, robotics, and the very nature of intelligent agents on the factory floor. Along the way, we discuss the parallels between LLMs and the invention of refrigeration, ask who will become the 'Coca-Cola' of AI, and reflect on the real impact of democratizing advanced technology. Join us as we debate the future of humanoid robots, the shifting boundaries between tool and actor, and the practical challenges of bringing AI-driven agents into real-world industry. If you're curious about where industrial automation is headed—and what innovations might define the next era—this conversation is not to be missed.

In this episode, I sit down with Andre Sagodi, a manufacturing and innovation expert at Audi, to explore how artificial intelligence is transforming the car production process. We dive into Audi's journey from piloting AI-based crack detection in sheet metal parts to scaling these solutions across multiple global sites. Andre shares practical insights on overcoming challenges like data consistency, model scalability, and workforce collaboration, all while aligning with Audi's strategic vision. Together, we discuss what it really takes to turn promising AI pilots into production-ready systems that deliver real business value. If you're curious about the future of industrial AI and want a behind-the-scenes look at innovation inside a world-class manufacturer, this episode is for you.

In this episode, we dive into the seismic shifts happening in industrial AI, from the so-called 'SaaS-mageddon' disrupting established software models to the rise of AI-powered agents that promise to redefine productivity. We share first-hand insights on how large language models and open-source projects are reshaping the landscape, and what this means for the future of industry solutions. We then turn to the frontier of pharmaceutical innovation with Stanisław Jastrzębski, CTO & Co-Founder from molecule.one, who reveals how deep learning, automation, and custom data are accelerating drug discovery. Join us as we connect the dots between industrial AI trends, the evolution of personal assistants, and the chemistry breakthroughs powering tomorrow's medicines. Tune in for a global perspective and a dash of inspiration straight from the heart of Istanbul.

In this episode, I sit down with Norbert Jung, CEO of Bosch Connected Industry, to explore the evolving landscape of manufacturing co-intelligence. We dive into how AI agents and data are reshaping factories, why human expertise remains irreplaceable, and what 'industrial grade' really means in this new era. Norbert shares practical examples from Bosch's own production lines, debunks common myths, and explains how semantic data and agentic AI are driving real-world results. We also tackle the challenges of scaling innovation, the future of lights-out factories, and why Europe's workforce transformation is both a necessity and an opportunity. If you want an inside look at the practical side of AI in industry—and what it means for the people behind the machines—this conversation delivers fresh perspective and actionable insights.

In this episode, we journey from weather woes in Franconia straight into the heart of Amsterdam's AI innovation, exploring what happens when robotics, world models, and agentic AI collide. We share our firsthand impressions from deep-dive research sessions at Lab42, where startups and researchers are joining forces to push industrial AI forward. You'll hear my conversation with John Harrington of HighByte, who breaks down the role of MCP services and why they're critical for bridging the gap between factory data and intelligent agents. We don't just talk tech—we talk about the people, the practical challenges, and the evolving landscape that's making industrial data accessible to everyone, not just engineers. If you're curious about how symbolic AI, digital cousins, and scalable architectures are transforming manufacturing, or you want to know what's next for AI in the Alps, this episode is your front-row seat. Tune in for insights, laughs, and a clear-eyed look at the road ahead for industrial-grade AI.

In this episode, I dive deep into the future of industrial AI with Andreas de Jonge, exploring how inline AI is reshaping real-world manufacturing. We discuss why tailored, ultra-efficient AI models can outperform bulky, generic solutions—especially in Europe's highly specialized industries. Andreas shares how his company's EdgeBrain device brings intelligence directly to the data stream, enabling plug-and-play integration and unprecedented robustness for even the oldest production lines. Together, we unpack the challenges of brownfield integration, the power of synthetic data, and the path to true European digital sovereignty. If you're curious about the next wave of industrial automation and what it means for productivity, this conversation is not to be missed.

In this episode, we dive into the evolution of AI robotics, exploring how lessons from cucumber-picking robots are now revolutionizing industrial inspection. We sit down with Wouter from VBTI to unpack the journey from complex greenhouse challenges to the development of Adam One—an inspection robot that doesn't just automate, but truly understands the task at hand. We reflect on the unique technical hurdles, the fusion of multiple AI models, and the potential for scalable, domain-specific foundation models. As always, we bring our authentic curiosity and skepticism to the latest industry trends, from federated learning revivals to the shifting boundaries of edge and cloud AI. Join us for a conversation that connects practical innovation with the big questions shaping industrial AI.

In this episode, Peter sits down with James Zhang, co-founder and CPO of OpsMate AI, to explore the game-changing role of AI agents in factory operations. We discuss why traditional human-centric manufacturing models are hitting their limits and how a hybrid workforce—combining humans, robots, and digital agents—can address the skilled labor shortage and boost productivity. James shares firsthand insights into the unique challenges of deploying AI in diverse factory environments, the importance of knowledge graphs, and why empowering frontline teams is key to successful adoption. If you're curious about the practical steps to bring AI agents onto your shop floor, or just want to understand what the future of industrial automation looks like, this conversation is for you.

In this episode, we dive into the ambitious European initiative offering up to one billion euros to fuel the next wave of AI innovation. I sit down with Johannes Otterbach, the so-called "billion euro man," to discuss how SPRIND is bridging the gap between research and industry, and what it really takes to launch a world-class AI lab in Europe. We explore the challenges of scaling, the importance of industrial-grade AI, and why Europe's unique landscape demands new strategies for success. Along the way, you'll hear our candid thoughts on talent, technology, and the future of frontier AI. If you're curious about the intersection of ambitious funding, cutting-edge research, and industrial application, this conversation is for you.

In this episode, we dive into the world of responsible data annotation with Andreas Schachl, co-founder of Responsible Annotation Services. We explore how neurodivergent individuals—especially those on the autism spectrum—are bringing unparalleled focus and precision to the task of data labeling, raising the bar for AI model training quality. Andreas shares the origins of their company, the unique strengths of their team, and why European sovereignty and data protection matter more than ever. We discuss the practical steps involved in their annotation process, how they measure quality, and the real impact their work has on customers and the broader tech ecosystem. Join us for an insightful conversation that challenges assumptions and highlights how inclusive innovation is driving the next frontier in industrial AI.

In this episode, Robert sit down with Nick Tebeau from BizLink to explore how AI and machine learning are revolutionizing something as fundamental as cable management in industrial robotics. We dive into why predictive maintenance matters, how edge devices can deliver real-time insights, and what it takes to keep production lines running smoothly without ever touching the robot controller. You'll hear about the challenges of creating adaptable AI models for complex manufacturing environments and how customer needs are shaping the future of smart factories. If you're curious about the ROI of industrial AI, the realities of edge computing, or how even the simplest hardware can become a game-changer, this conversation is packed with practical takeaways and forward-looking ideas.

In this episode, we dive into why sound is the next big leap for AI in industry. I share insights from recent events, including the SPS exhibition and Europe's AI scene, and discuss why edge AI is outpacing the cloud on the shop floor. Then, I sit down with Dr. Finnur Pind from Treble Technologies to explore how synthetic sound data is powering smarter machines—from voice recognition in cars to giving robots a sense of hearing. You'll hear how simulating sound opens up new possibilities for robotics and spatial awareness, and why industries are racing to make AI models that can truly listen. We also touch on the unique AI startup scene in Reykjavik, the challenges of accurate sound modeling, and what the future holds for audio-driven innovation. Join us to discover why the future of industrial AI might just depend on what machines can hear.

In this episode, I sit down once again with Professor Oliver Niggemann to unravel the world of numerical machine learning—where traditional engineering meets cutting-edge AI. We explore real-world projects from diagnosing the International Space Station to designing safer bridges, smarter batteries, and even optimizing biodiversity. Oliver breaks down how fusing symbolic knowledge and neural networks is revolutionizing simulation, design, and problem-solving across industries. If you've ever wondered how AI can speed up material discovery or what the future holds for interdisciplinary engineers, this conversation is for you. We dive into the power of surrogate models, the evolution of engineering education, and why tomorrow's innovations demand both deep technical expertise and creative collaboration. Join us for a look at the next frontier in industrial AI.

In this episode, we dig deep into the evolving landscape of industrial AI, from billion-dollar cloud partnerships in Europe to the fierce debate over digital sovereignty. We break down what sets 'sovereign by contract' apart from 'sovereign by origin,' and why this matters for companies navigating global AI strategies. We're joined by Professor Stratis Gavves from the University of Amsterdam to unpack the world of robotics models and the critical role of physics-informed AI. Alongside the latest industry news—like SAP's move into open-source models and the rise of edge AI—we explore what it takes to stay competitive in a rapidly shifting market. Join us as we question, challenge, and celebrate the innovations set to define the next era of industrial intelligence.

In this episode, I sit down with Richard Rovner, Vice President of Marketing at MathWorks, to explore the fascinating evolution of AI in engineering. We trace the technology's journey from its early days in the 1980s, through transformative milestones like deep learning and code generation, all the way to today's AI-driven design tools and copilots. Richard shares insider perspectives on how AI is changing the engineer's workflow, what generative AI means for the future, and why the human touch remains essential in a world of accelerating automation. If you're curious about where engineering and AI are headed—or how to stay ahead in this rapidly shifting landscape—this conversation is packed with insights you won't want to miss.

In this episode, we dive deep into the real-world state of AI adoption across industries, unpacking the findings of a major new survey with Gabriel Krummenacher. We explore which AI use cases actually move the needle—hint: operational efficiency tops the list—and why simply chasing the latest GenAI trends won't deliver results without a solid data science foundation. You'll hear our honest takes on what separates hype from impact, why ethical frameworks and governance drive innovation (not stifle it), and how US and European companies are taking very different paths with AI. If you're wondering where AI is truly transforming business—and what it really takes to succeed—this episode is for you.

In this episode, I sit down with Schunk CTO Timo Gessmann and Google DeepMind's Phillip Lippe to explore how AI is transforming the world of robotics and industrial automation. We kick things off with Timo, who shares his vision for merging mechanical expertise with AI, the challenges of attracting top talent, and why Europe needs its own world-class robotics models. Then, I'm joined by Phillip Lippe from the Gemini team, who dives into the latest advances in generative AI, the power of foundation models, and what the future holds for scalable, domain-specific intelligence. Whether you're fascinated by embodied AI, curious about industry trends, or want a peek into the minds shaping tomorrow's robots, this episode is packed with insights you won't want to miss.

We spent time in the Black Forest, where innovation met inspiration at our latest industrial AI gathering. In this episode, we share stories—from empowering autistic talent in data annotation to the technical leaps in foundation models for time series. You'll hear from leading minds like Jakub Tomczak, who flew in from San Francisco to challenge what's possible in generative AI, and explore how the landscape is shifting from traditional machine learning to adaptive, context-rich automation. Join us as we question what makes AI truly intelligent, debate the role of human factors, and look ahead to the next wave of industrial transformation. If you're curious about how AI is evolving in real factories, how new job profiles are emerging, or why continuous learning and time series matter more than ever, this conversation is for you.

In this episode, I dive into the fascinating world of non-destructive material testing with Christian Wagner from Vogt Ultrasonics and Jan Koutnik from Evoptima. Together, we explore how AI is revolutionizing spot weld inspection in the automotive industry, making quality control smarter and more precise than ever before. We discuss the challenges of processing massive ultrasonic datasets, adapting AI models to real-world manufacturing, and the journey from legacy systems to cutting-edge solutions. If you're curious about how collaborative robots, vision systems, and robust AI are reshaping factory floors—and what it takes for mid-sized companies to lead the charge—this is an episode you won't want to miss.

In this episode, we dive into the realities behind industrial AI adoption—both the successes and the setbacks. We share candid conversations with engineers frustrated by unreliable AI and explore why real-world robustness remains the central challenge. Then, I sit down with Andy from Path Robotics, who reveals how their AI-driven welding robots are redefining adaptability and reliability on the shop floor. We unpack what makes their approach unique, from multi-modal sensors to reinforcement learning and massive real-world datasets. If you care about the future of manufacturing, data-driven automation, and the evolution of industrial roles, you won't want to miss this discussion. Join us as we look ahead to what's possible when AI finally delivers on its industrial promise.

In this episode, I sit down with Armin Hadzalic to tackle the burning question: will the Model Context Protocol (MCP) truly transform industrial AI? We dive deep into how MCP connects large language models and agents to real-world data, what makes it different from protocols like OPC UA and MQTT, and why the biggest players in tech are rallying behind it. Armin shares practical examples, addresses skeptics, and explains what MCP means for engineers, technicians, and anyone working in industrial automation. If you're curious about the rise of agents, the future of dashboards, and what it takes to stay ahead in the evolving world of industrial AI, you won't want to miss this conversation. Join me as we explore where MCP is headed and what it could mean for your job and your factory.

In this episode, we dive into the evolving landscape of industrial AI, starting with a lively Oktoberfest recap before shifting gears to the latest breakthroughs in physics-informed neural networks and user interfaces. We discuss the real-world impact of Europe's AI Act, featuring insights from industry leaders and an in-depth interview with Sampo Leino of MinnaLearn on building AI literacy for enterprises. As we unpack strategic investments, robotics trends, and the challenges of compliance, we question what it means to use AI safely and competitively. Throughout the conversation, we keep it grounded in everyday experience—how regulation, technology, and practical learning are shaping the factories and workplaces of tomorrow. Tune in to hear how we're navigating this complex, fast-moving frontier and what it means for anyone working with AI today.

In this episode, I sit down with Jens Stapelfeldt, Global AI Lead at AMD's University Program, to explore how AMD is challenging the status quo in AI hardware. We dive into the evolution from proprietary ecosystems to AMD's open, enterprise-ready platforms, discussing what this means for industrial AI applications and decision makers. Jens shares insights on AMD's broad AI portfolio, from GPUs to edge solutions, and how strategic collaborations are accelerating innovation. We also examine the significance of open-source principles and the impact of diverse hardware options on industrial use cases. Join me as we uncover why AMD positions itself as the most versatile player in the AI space, from edge devices all the way to the cloud.

In this episode, we dive into the rapidly evolving world of AI-driven material discovery with Victor Schmidt, tracing his journey from research at Mila in Montreal to launching a cutting-edge startup in Paris, Entalpic. We explore how generative models and graph neural networks are reshaping the way we discover new materials, with a focus on environmental impact and sustainable industry. I discuss the challenges of bridging domain expertise with machine learning and how 'research as a service' is opening new doors for industrial innovation. Paris emerges as a vibrant AI hub, and we uncover what makes its ecosystem unique for startups and researchers alike. If you're curious about the intersection of AI, chemistry, and entrepreneurship, this conversation is packed with insights and forward-looking ideas.

Peter Seeberg talks to Ram Seetharaman, Head of AI & Product Manager at Synera about Building your first digital engineer.

Pravin Waghmare is the CEO of Prescient from Pune. Magnus from Sandvik gave us the tip to talk to him about AI and CAD.

What is happening in the field of industrial AI in India? An entire generation of young people is striving to enter this new field. We wanted to know: What does the training look like?

With the help of artificial intelligence, Audi analyzes around 1.5 million weld points on 300 vehicles per shift at its Neckarsulm site. By way of comparison, employees in production previously monitored the quality of the processes involved in resistance spot welding (WPS) manually on a random basis using ultrasound, checking just over 5,000 weld spots per vehicle. Thanks to the use of AI, employees can now concentrate on potential anomalies and thus check quality even more efficiently and in a more targeted manner.

Peter Seeberg spoke with Vlad Larichev Industrial AI Lead at Accenture about Agents, LLMs and AI for Engineering Tasks.

Rico Knapper is the CEO of Pailot and loves PCBs shopfloors. Why? Because his AI based solution outperforms other approaches. We met him at Siemens.

We celebrate... with WinCC, Agents and a special guets

Richard Socher is our guest. We discuss with the German wunderkind the role of LLMs in business and how architectures are changing. We met Richard Socher at the AI with Purpose Summit hosted by Siemens.

Peter Seeberg talked to Praveen Kapse. He is the Vice President - Corporate Strategy & Data Analytics Head of Tridiagonal Solutions in Pune, India.

In this episode of the Industrial AI Podcast we talk about MAN TruckScenes: the world's first freely available sensor and driving data for autonomous trucks. In the newspart Peter and Robert discuss their plans and their news.

In the interview with Peter Seeberg, Sepp Hochreiter explains how he did it, what state tracking is, how he came up with the devices and what the next steps are.

In this episode, Robert Weber talks to Tom Zehle from Airbus. He explains how he stays up to date, how he analyzes papers, and what he wants from research. The Industrial AI Podcast reports weekly on the latest developments in AI and Machine Learning for the engineering, robotics, automotive, process and automation industries. The podcast features industrial users, scientists, vendors and startups in the field of Industrial AI and Machine Learning. The podcast is hosted by Peter Seeberg, Industrial AI consultant and Robert Weber, tech journalist.

Peter Seeberg talks to Thomas Dhollander, Co-founder & CPO at Timeseer.AI about Trusted IIoT Data as the Key to Proactive Operations.

In this episode, we talk to Lucas from MathWorks about their Safe AI Approach and how it can be used to make money, think about the Pope, and introduce our new US correspondent.

MAGS harness the creative potential of generative models while enabling intelligent agents to collaborate, adapt, and solve complex problems in real time. This new synergy elevates AI systems beyond traditional automation, pushing them into the realm of autonomous problem-solving, creative thinking, and adaptive decision-making. The development of MAGS marks a significant shift in the way organizations approach challenges in rapidly evolving environments. By integrating Large Language Models (LLMs) and other advanced generative models with agents capable of learning and interacting independently, MAGS enable organizations to address problems that are not only multifaceted but also constantly changing. This guide explores the foundational elements of MAGS, their architecture, and why they matter for industries that need to navigate complexity, unpredictability, and scale.