German industry missed out on the initial wave of smart connected products, allowing companies like Tesla to gain an advantage with their software-centric offerings. However, with the emergence of artificial intelligence (AI), German industry now has a second chance to reclaim leadership. While the public's attention has been focused on the creative and visual capabilities of AI tools like ChatGPT, Midjourney, and Stable Diffusion, the true transformative power lies in the unnoticed impact on the entire software development industry. As Jenson Huang said in 2017: "AI is eating software".
Over the next few years, the IT industry, employing ten million software engineers worldwide, including a significant number in India, is set to be disrupted and completely transformed. A new model known as robo-development is on the horizon, where human developers are replaced by infinitely scalable robotic coders guided by small teams of software architects and user experience designers. This new model offers German companies an opportunity to leapfrog the first generation of software-defined products—from the likes of Tesla and Nest in the US, and BYD and Midea in China—and create software-first, AI-first products.
The iPhone Moment: GPT-4 and the Age of AI
The recent announcement of the launch of GPT-4 on March 14th by Sam Altman marked a significant shift in information technology, akin to the impact of the iPhone's introduction by Steve Jobs in 2007. Such ground-breaking moments are rare occurrences that bring about transformative change. While previous technologies like Web 3.0 and blockchain and concepts like the Metaverse have sparked interest, the advent of AI, which Bill Gates refers to as the "Age of AI," represents a profound shift, potentially even more significant than the internet itself.
GPT-4 has already altered the established rules of the game, causing some companies to quickly pivot and embrace the new reality. These early adopters are likely to benefit from the change. However, other companies are clinging to the old ecosystem and downplaying the impact of this transformation. They may experience a "Kodak moment," becoming obsolete as their products are overtaken by the new paradigm, and they fail to adapt.
Legacy Challenges and the Software-First Approach
Legacy factors, including legacy technology, talent, and business models, pose challenges for companies attempting to pivot. Even earlier generations of AI can also be considered legacy, as the introduction of generative AI and large language models has ushered in a new paradigm. These models outperform humans in various cognitive tasks, possess deep understanding, and can be easily utilized through prompt engineering and simple API calls. Embracing this technology as the core of their business is crucial for companies aiming to succeed in the changing landscape.
The Right Attitude: How Mercedes Saved Tesla
In 2008, Elon Musk made a pivotal decision to go all-in on Tesla, which was facing imminent bankruptcy. Daimler unexpectedly stepped in, not only injecting $50 million but also providing expertise and access to its automotive network including software capabilities for the first Tesla OS. This partnership proved crucial in stabilizing Tesla's product, the Roadster, and advancing subsequent models like the Model S. Despite this involvement, Daimler ultimately sold its stake in Tesla. Had it stayed fully invested, the company would be sitting on a tidy profit of around $75 billion today. This episode and others, like the acquisition of KUKA by Chinese electronical appliances manufacturer Midea, serve as a reminder that German software capability was never the issue; it was a matter of belief and a willingness to take calculated risks.
The Right Approach: How Software-First Differs from Hardware-First
The agile approach to software development, characterized by quickly releasing a minimum viable product and iterating based on user feedback, is well-known. However, in the realm of industrial software, a traditional "hardware-first" approach has prevailed. Legacy cars, for example, consist of mechanical and electronic components from various suppliers, making it challenging to deliver software updates seamlessly. Tesla disrupted this approach by prioritizing software and implementing a streamlined electronic control unit ("ECU") architecture that supported it. This allowed it to introduce over-the-air updates in 2014 and commence a cycle of iterative improvements. The same principles can be applied to any engineered product, recognizing that the value increasingly lies in software and the services it enables. The key is this: “How can my physical product support innovative and adaptable software-driven experiences?”
Robo-Development: Entering the Third Epoch of the Software Industry
The software development industry is on the cusp of entering a third epoch. The first epoch, from 1950 to 1999, involved on-site experts, while the second epoch, from 2000 to 2022, saw the rise of off-site industrial development, notably in far-shore software factories in India. The forthcoming epoch, starting in 2023, will be characterized by the adoption of robocoders and robodesigners. Not only will artificial intelligence transform the product itself by integrating hitherto undreamed-of intelligence into product services, but the very way of developing software products will also be changed. Robo-development means that smaller teams of designers and experts can work with generative coding and design agents to deliver products which previously would have required large numbers of specialists in a factory-like approach. The new model can reduce time-to-market and teams are likely to be more regional, highly integrated and closely linked to users.
All that is needed is the courage to seize the moment and gain leadership in AI driven software. And this time the profits and kudos need to go to the German companies with the capabilities, rather than bolder US or Chinese rivals.
Conclusion: German Industry's Opportunity in Software-Defined Products powered by AI
Software-defined products powered by AI, offer German industry an opportunity to regain a competitive edge. OpenAI and Google DeepMind's recent product launches in early 2023 have disrupted business models and fundamentally changed software development itself. Software-first and AI-first approaches are no longer optional but imperative for German industry to lead. The key to making this happen lies in:
- embracing simplicity and flexibility from the earliest stage of product conception and empowering talented teams to think differently,
- designing for software experiences rather than using software as an add-on,
- empowering smaller, highly talented teams which are powered by AI tools to take on the factory model ("class is better than mass"),
- redefining software delivery models and ecosystems with regional teams and nearshoring within the EU preferable to far-shore models,
- regaining the spirit of the Gründerzeit and of visionaries like Carl Benz, Werner von Siemens and Robert Bosch.
Leading this disruption necessitates a mindset shift from incremental to exponential change and disruption. Designing product experiences with a focus on user-centricity, embracing simplicity and flexibility, empowering small teams, and leveraging AI to increase productivity are crucial steps. German industry must accept that maintaining the status quo is not an option. The choice is clear: either lead this disruption or be disrupted.