The world of technology is in a constant state of flux, with artificial intelligence (AI) advancing at an unprecedented pace. Every week brings new chatbots, AI image generators, and software breakthroughs, as companies strive to create ever more capable systems. However, beyond our screens, some experts argue that physical innovation has not kept pace with digital advancements, leading to a divide between 'bits' and 'atoms'.
The 'bits' refer to digital technologies such as software, smartphones, and AI, which have accelerated rapidly. In contrast, the 'atoms' represent the physical world, including robotics, manufacturing, transport, energy, and infrastructure, where progress has often been slower, more expensive, and harder to bring into everyday life.
This divide has sparked an uncomfortable question: Has physical innovation slowed down while digital technology races ahead? While it's true that physical technology has not advanced as dramatically as AI, it's not accurate to say that physical innovation has stopped. Reusable rockets have reduced the cost of space travel, gene-editing tools have transformed biomedical research, battery technology has improved, and materials science continues to emerge.
However, for many consumers, the world doesn't feel as futuristic as expected. Predictions made in the 1950s, 1960s, and 1970s envisioned cities filled with flying cars, widespread household robots, and routine commercial space travel by the early 21st century. Instead, the biggest technological revolution of the past three decades has largely taken place in software.
Robotics expert Dr. Sue Keay explains that the comparison isn't surprising because software and physical engineering operate on fundamentally different timelines. Hardware is hard, and generative AI now means someone can write code, launch an e-commerce platform, and be in business online within minutes. Building anything that has to move through and act on the physical world is a different order of problem.
Keay highlights the challenges of physical technologies, such as design, manufacturing, testing for safety, certification, and reliable operation in unpredictable real-world environments. Advances in AI are helping robots better understand language and surroundings, but the physical world still presents obstacles that software alone can't overcome.
The reliability of robots remains a significant hurdle, as machines that perform well in laboratories can struggle when faced with different lighting, changing environments, or unexpected obstacles. While the cost of humanoid robots has fallen dramatically, making them more accessible for research and development, expectations still need to match reality.
Economist Tyler Cowen popularized a related idea in his book 'The Great Stagnation', arguing that many transformative inventions that reshaped society, such as electricity, cars, antibiotics, and aviation, had already been discovered, leaving today's innovators to tackle increasingly difficult problems.
Other researchers have also argued that new ideas are becoming harder to find, requiring larger research teams and greater investment to produce the same productivity gains. Centre for Future Work chair Jim Stanford notes a gap between the capabilities of new software and its actual use in the real world of employment and production.
Stanford emphasizes the importance of investment in machinery, equipment, and infrastructure for technological innovation. Business investment in machinery and equipment has been sluggish for much of the past decade, although spending has picked up more recently. Australia, for instance, has shown it can develop new technologies but needs greater investment in putting those technologies to work across the economy.
Australia's struggle to translate research into large-scale manufacturing or globally dominant technology companies is a concern. Keay points out that many successful Australian robotics companies are acquired by overseas firms before they can grow into large domestic businesses, resulting in offshore profits and intellectual property.
Stanford argues that Australia's broader challenge is less about a shortage of ideas than decades of underinvestment in industries capable of commercializing them. The economy's focus on resource extraction, property development, and finance has come at the expense of manufacturing, advanced engineering, and physical infrastructure.
Some economists have also pointed to Australia's long-running productivity slowdown and relatively weak investment in research and development as signs that the country risks falling behind in the physical technologies that underpin future industries. Whether AI can reverse this trend remains an open question.
Stanford believes that excitement surrounding AI has outpaced its demonstrated economic impact. He argues that there is little evidence of economy-wide productivity improvements from AI and believes the current AI boom will collapse one day. Keay adds that people often underestimate the difficulty of building machines capable of operating safely and reliably in the real world.
In conclusion, while AI continues to advance rapidly, the question isn't whether innovation has stopped but whether today's digital breakthroughs can deliver the same sweeping physical transformation as electricity, automobiles, and aviation once brought to everyday life. The divide between 'bits' and 'atoms' highlights the need for a balanced approach to technological development, where both digital and physical innovations advance together to shape a more advanced and connected future.
